WO2022042655A1

WO2022042655A1 - Method for constructing fingerprint map of multi-floor building, positioning method, and apparatus

Info

Publication number: WO2022042655A1
Application number: PCT/CN2021/114822
Authority: WO
Inventors: 张云; 曾丹丹; 王永亮
Original assignee: 华为技术有限公司
Priority date: 2020-08-31
Filing date: 2021-08-26
Publication date: 2022-03-03
Also published as: CN114125696A

Abstract

Provided in embodiments of the present invention are a method for constructing a fingerprint map of a multi-floor building and an apparatus. The method is applied to a server, and comprises: receiving crowdsourced data from multiple mobile terminals, the crowdsourced data comprising motion data and location fingerprint data acquired when the mobile terminals move between floors in a multi-floor building; determining, according to the crowdsourced data, floor structures of at least two floors in the multi-floor building, locations of passages connecting the respective floor structures on the floor structures, and structures of the passages; and generating a fingerprint map of the multi-floor building according to the floor structures of the at least two floors, the locations of the passages, and the structures of the passages. By means of the embodiments of the present invention, a fingerprint map of a multi-floor building can be automatically generated, thereby eliminating the need for manual editing, and reducing required human and material resources and costs.

Description

Construction method, positioning method and device of fingerprint map of multi-story building

This application claims the priority of the Chinese patent application filed on August 31, 2020 with the application number 202010899650.X and the application title "Construction method, positioning method and device for fingerprint map of multi-story buildings", all of which The contents are incorporated herein by reference.

technical field

The present application relates to the technical field of indoor positioning, and in particular, to a method for constructing a fingerprint map of a multi-storey building, a positioning method and a device.

Background technique

With the rapid development of smartphones and mobile Internet, location-based services have attracted more and more attention. Real-time positioning has become a basic technology for many high-level applications such as transportation, commerce, logistics, and personalized services. In outdoor situations, GNSS provides a good positioning service such as Global Positioning System (GPS). However, in indoor environments, GPS cannot achieve suitable accuracy due to signal fading and multipath effects. Therefore, indoor positioning technology has become a hot topic in recent years. According to relevant statistics, 80% of people's mobile communication services, business transactions and public entertainment take place indoors, and 80% of information acquisition is related to time and location. With the large-scale popularization of intelligent terminals and 5G network communications, indoor scenarios Such as shopping malls, office parks, airport stations, underground garages, hospitals and other environments, the demand for location services and applications is becoming more and more urgent, such as indoor map navigation, store-level advertising push and value-added services, fire rescue, etc. Driven by comprehensive factors such as big data, intelligence of everything and policy support, the global indoor positioning market will reach tens of billions of dollars, so indoor positioning has high market potential and market prospects.

WiFi location fingerprinting method is a commonly used indoor positioning method. This method can utilize the existing wireless local area network infrastructure, and can achieve positioning through mobile terminals (such as smart phones), without the need for users to add additional equipment, so its application is the most widely. The existing WiFi location fingerprinting method includes two stages: offline fingerprint collection and online positioning. The purpose of offline fingerprint collection is to build a WiFi location fingerprint database in the indoor area to form a WiFi fingerprint map; in the online positioning stage, according to the current WiFi signal strength (Received Signal Strength, RSS) information, the positioning algorithm is used to match it with the WiFi location fingerprint database. The WiFi fingerprint map is matched and compared to estimate the user's location.

Nowadays, the need for indoor positioning of multi-storey buildings (such as multi-storey shopping malls, office buildings, etc.) is increasing day by day. The existing solutions to achieve multi-floor indoor positioning usually rely on the maps of each floor, and manually edit different floor maps and The corresponding relationship between WiFi fingerprint maps has low work efficiency and cannot realize the positioning between floors.

SUMMARY OF THE INVENTION

The present application provides a method for constructing a fingerprint map of a multi-story building, a positioning method and a device, which can realize the automatic generation of the fingerprint map of the multi-story building, avoid the intervention of manual editing, and save the cost of manpower and material resources.

In a first aspect, an embodiment of the present application provides a method for constructing a fingerprint map of a multi-story building. The method is applied to a server and includes: receiving crowdsourced data from multiple mobile terminals, where the crowdsourced data includes the mobile terminal Motion data and position fingerprint data collected in the process of moving across floors in a multi-story building; according to the crowdsourcing data, determine the leveling skeletons of at least two floors in the multi-story building, and connect between the leveling skeletons The position point of the channel on the leveling frame and the configuration of the channel; according to the leveling frame of the at least two floors, the position point of the channel and the configuration of the channel, generate the Fingerprint maps of multi-story buildings.

Among them, the passage is a passage facility used to connect different floors of a multi-storey building, the passage is connected between different floors, and people and goods can be transferred between different floors through the passage. There are many configurations of passages, such as stairs, elevators, escalators, ladders, steps, and ramps. The location point of the passage refers to the connection position of the passage and the level skeleton, that is, the endpoints at both ends of the passage. It is understandable that each Channels all have pairs of location points.

It can be seen that in the embodiment of the present application, through the interaction of multiple mobile terminals and servers, the locations of landmark points of the channel on the leveling skeleton (such as the endpoints at both ends of the channel and the uplink and downlink connections of pairs of landmark points), the channel configuration , automatic creation of multi-floor 3D topology map, the whole process does not need manual editing and annotation, no map dependence, can automatically and efficiently output the fingerprint map of multi-floor buildings in the whole process, improve the efficiency and accuracy of fingerprint map generation, solve the problem The problem of relying on manual editing for map construction in the prior art saves manpower and material costs.

The crowdsourced data is the data obtained by the mobile terminal through sensors, radio frequency signals, or network positioning without the user's perception. In the embodiment of this application, the crowdsourced data includes the mobile terminal moving across floors in a multi-story building. motion data and position fingerprint data collected during the process, the motion data is specifically data collected by at least one sensor, and the at least one sensor includes at least one of an accelerometer, a gyroscope, a magnetometer, and an inertial measurement unit; the Location fingerprint data includes the number and signal strength of wireless access points.

In an existing technical solution, a large number of collection point samples are required in the offline collection stage, and the workload of data collection is relatively large, especially for a large-scale indoor area, it is necessary to deploy specific infrastructure and perform manual offline collection, which consumes huge amounts of data. The cost of manpower and material resources seriously limits the large-scale popularization and application of WiFi location fingerprinting.

In the embodiment of the present application, the server can make full use of crowdsourced data collected by various mobile terminals to construct a fingerprint map of a multi-story building, which replaces manual data collection and manual annotation, and can realize 3D topology without map dependence. The construction of the map does not require the deployment of special data acquisition devices indoors, which greatly reduces the cost of labor and material resources.

Based on the first aspect, in a possible embodiment, the configuration of the channel includes a direct connection type or an oblique connection type; the direct connection type means that the position points at both ends of the channel are collinear in the vertical direction, and the diagonal connection type Type means that the location points at both ends of the channel are not collinear in the vertical direction

Based on the first aspect, in a possible embodiment, the generating a three-dimensional topological map according to the leveling skeleton, the location points of the channel and the configuration of the channel includes: according to the relationship between the leveling skeletons The connection relationship of the passages is to sort the respective leveling skeletons so that the order of the respective leveling skeletons conforms to the upper and lower relationship of the floors in the multi-storey building; according to the location points of the passages and the structure of the passages type, and aligning the sorted flat skeletons in the three-dimensional coordinate space to obtain the fingerprint map.

It can be seen that this application makes full use of the characteristics of the landmark points of the passage, and automatically sorts and aligns multiple leveling skeletons. The whole process does not require manual editing and labeling, and has no map dependence. The entire process can be automated and efficiently output multi-story buildings. It improves the efficiency and accuracy of fingerprint map generation.

Based on the first aspect, in a possible embodiment, before aligning the sorted leveling skeletons in the three-dimensional coordinate space according to the location points of the channel and the configuration of the channel, the method further includes: : Determine the mapping relationship between the sorted leveling skeletons and the floors in the multi-storey building.

It can be seen that, in addition to automatically sorting and aligning multiple leveling skeletons, the solution of the present application can also identify the absolute floors corresponding to the leveling skeletons, which is helpful for automatically and efficiently outputting fingerprint maps of multi-story buildings. In this way, the indoor positioning of the building can be accurately positioned to the absolute floor, which improves the efficiency of fingerprint map generation and the positioning accuracy.

Based on the first aspect, in a possible embodiment, sorting the leveling skeletons according to the connection relationship of the channels between the leveling skeletons includes: according to the channels between the leveling skeletons to construct a directed acyclic graph of the respective leveling skeletons; the respective leveling skeletons are sorted according to the directed acyclic graph.

For example, when defining the up-down relationship of the leveling skeleton, the up-down direction of the vast majority of connection pairs can be taken as the up-down direction of the leveling skeleton among all the connection landmarks. Convert the direction to a consistent direction (both up or down), the directed acyclic graph segmentation adopts a depth-first search algorithm, starting from a certain node, the set is divided into a searched set and an unsearched set, the current search When there is a connection relationship between the node and the node other than the direct parent node in the combination of the searched node, the connection relationship is disconnected, and the search can continue until there is no connection relationship. After searching, a directed acyclic graph including each leveling skeleton is finally obtained. After the directed acyclic graph segmentation is completed, the ascending and descending ordering relationship can be sorted out according to the connection relationship between the leveling skeletons. By constructing a directed acyclic graph, it is beneficial to realize fast and accurate leveling skeleton sorting, which helps to ensure the efficient generation of fingerprint maps.

Based on the first aspect, in a possible embodiment, the determining the mapping relationship between the sorted leveling skeletons and the floors in the multi-story building includes: determining each of the levelings according to the crowdsourcing data. The number of entrances and exits of the floor corresponding to the floor frame; the floor frame with the largest number of entrances and exits corresponds to the first floor of the multi-storey building; according to the sorting result of each floor frame, the floors corresponding to the remaining floor frames are determined.

Generally speaking, most buildings will set the most entrances and exits on the first floor for the convenience of passage, transportation, fire protection, etc., and use the detected number of valid entrances and exits to select the floor with the most entrances and exits as the first floor, and sort according to the floor. relationship to get the floor corresponding to the real physical building. Therefore, the absolute floor recognition first uses the entrance and exit identification results to perform meanshiff clustering on the entrance and exit to obtain an effective entrance and exit clustering. According to the number of entrance and exit clusters, identify the reference floor - the first floor, and finally use the floor sorting relationship to obtain the absolute floor.

Based on the first aspect, in a possible embodiment, according to the position point of the channel and the configuration of the channel, aligning the sorted leveling skeletons in the three-dimensional coordinate space includes: when When there is a direct connection type channel between two leveling skeletons, the two leveling skeletons are aligned by aligning the position points at both ends of the channel in the vertical direction; when there is an oblique connection type between the two leveling skeletons When the channel is set, the two leveling skeletons are aligned according to the horizontal offset between the position points at both ends of the channel.

For example, for directly connected passages, such as elevators and stairs of the direct connection type, the RANSAC algorithm can be used to calculate the similarity transformation between points with the same name, and then the skeleton coordinates to be aligned are transformed to the aligned leveling skeleton coordinate system using the similarity transformation.

After the straight alignment is performed, the alignment results can be further corrected based on the identified inclined stairs. Taking an inclined staircase as an example, first, the PDR distance between the connected landmark points is used as the hypotenuse distance, and the estimated height of the connected floor is used as a right-angle side, so that the horizontal offset between the diagonally connected landmark points can be estimated. , by compensating the calculated horizontal offset to the obliquely connected landmarks, and incorporating the compensated obliquely connected landmarks into the directly-aligned stairs and elevators for further direct-alignment calculations to obtain further corrected alignment results.

In addition, optionally, after the direct connection alignment and oblique connection alignment are completed, all landmark points can be used for ICP fine alignment, giving higher weights to the directly connected elevators and stairs, and using the escalators with implicit direct connection relationships for further precision Align floors to get a 3D topology map.

It can be seen that, in the embodiment of the present application, firstly, according to the configuration of the detection channel, it is determined whether the landmark point of the channel belongs to a direct connection type (such as a direct connection staircase and a direct connection elevator), an oblique connection type (such as an inclined connection staircase) or an implicit direct connection type. connection type (such as escalator); then, the direct connection type channel can be preferentially used for the initial direct connection alignment of the leveling frame; then, the horizontal offset is calculated using the PDR track distance and floor height difference in the inclined connection type channel, so that Compensate the horizontal offset of the leveling skeleton to further optimize the calculation of the alignment relationship; finally, optionally, the implicit direct connection relationship of all landmark points can be used to perform ICP fine alignment. A 3D topology map can be obtained through the above process. In this way, by adopting the coarse-to-fine floor coordinate system alignment strategy in stages, the fine alignment of floors can be achieved without resorting to indoor floor plans or manual editing, which is conducive to the realization of automatic multi-floor fingerprint map construction, and improves the efficiency of map construction. Accuracy, and the construction process of the 3D topological map does not need to rely on the existing indoor floor plan.

Based on the first aspect, in a possible embodiment, the process of determining the leveling skeletons of at least two floors in the multi-story building according to crowdsourcing data includes: obtaining the plurality of movements according to the motion data the movement trajectories of the terminal in the at least two floors; the movement trajectories of the multiple mobile terminals are aggregated to obtain the leveling framework of the at least two floors; wherein, the leveling framework and the position Fingerprint data association.

It can be seen that the embodiment of the present application can grow the leveling skeleton in a distributed and parallel manner based on the seed track (single movement track), realize the automation of the whole process and the high-quality leveling skeleton growth, and greatly improve the fingerprint map construction. Efficiency and Precision.

Based on the first aspect, in a possible embodiment, the process of determining, according to the crowdsourced data, the location points of the channels connected between the leveling skeletons includes: obtaining the multiple The movement trajectories of the mobile terminals in the at least two floors; identifying the position points belonging to the passage distributed on each of the movement trajectories; combining each of the movement trajectories with the leveling skeleton of the at least two floors Matching is performed to obtain the location point of the channel on the flat skeleton.

Specifically, the matching each of the movement trajectories with the leveling skeletons of the at least two floors to obtain the position points of the passage includes: based on the position point distribution on the movement trajectories, The moving track is divided into a leveling segment track and a cross-layer segment track; according to the leveling segment track and the leveling skeleton, a plurality of candidate position points of the channel on the leveling skeleton are obtained; according to For the plurality of candidate position points, the position points of the channel are obtained. For example, the trajectory of the leveling segment and the leveling skeleton are matched by the K-nearest neighbor algorithm to obtain multiple candidate position points of the channel on the leveling skeleton.

For example, each cross-layer movement trajectory (eg, a PDR trajectory) can be divided into two levels of flat-layer trajectory and a cross-layer trajectory. The track of the leveling segment firstly calculates the effective AP coincidence degree with each leveling skeleton. The maximum AP coincidence degree is the matching leveling skeleton, and then according to the KNN registration algorithm in the matching leveling skeleton, the leveling segment is calculated. The track is registered to the leveling skeleton, and the location of the landmark point on the leveling skeleton can be obtained after the coordinate transformation is performed by using the marks of the landmark points carried by the track of the leveling segment. Cross-segment trajectories are used for subsequent fingerprint clustering to obtain more precise landmark locations. Implementing the embodiments of the present application can achieve fast and accurate acquisition of channel location points.

Based on the first aspect, in a possible embodiment, the obtaining the position points of the channel according to the plurality of candidate position points includes: according to the cross-layer segment clustering centers, from the plurality of candidate positions The candidate position points of the selected part are clustered to obtain the position points of the channel; wherein, the cross-layer clustering center is based on the fingerprint similarity of the cross-layer trajectories of the multiple movement trajectories. The cluster centers obtained by clustering.

Taking the signal fingerprint as the WiFi fingerprint as an example, each cross-segment trajectories can be clustered by WiFi similarity to obtain the cross-segment clustering center. Initial classification, and then combined with the spatial coordinate information of each category of landmark points to perform meanshift clustering, remove outliers, and obtain an accurate cluster center, which is used as the precise location of the landmark points of the levelable skeleton.

Based on the first aspect, in a possible embodiment, the identifying the location point distribution of the channel on the moving track includes: for each mobile terminal in the plurality of mobile terminals, according to the mobile terminal The movement data is used to identify the position where the user behavior change occurs on the moving track; the user behavior change refers to the transition between the user behavior on the floor and the user behavior on the passage. According to the location where the user behavior change occurs, the location point distribution of the location points at both ends of the channel on the movement track is determined.

For example, the server can output various user behaviors of pedestrians walking with mobile phones through the behavior recognition program, including user behaviors such as going up and down stairs, elevators, and escalators. It should be understood that, in the definition of user behavior change, the transition between the user behavior on the floor and the user behavior on the passage can be a gradual process rather than a sudden process. By distinguishing between the two behaviors The junction point of can get the position point of the channel.

Based on the first aspect, in a possible embodiment, the crowdsourced data further includes altitude data collected by a barometer; and the identifying the location distribution of the channel on the moving track includes: for the multiple Each of the mobile terminals identifies, according to the motion data of the mobile terminal, the position where the user behavior change occurs on the moving track; the user behavior change refers to the user behavior on the floor and the user behavior on the floor. The transition between user behaviors on the channel; according to the height data of the mobile terminal, identify the position where the cross-layer event occurs on the moving track; according to the position where the user behavior change occurs and the cross-layer occurs The location of the event determines the location point distribution of the channel on the moving track.

It can be seen that since the current proportion of mobile terminals containing barometers is relatively small, it is necessary to use a large amount of collected data without barometers to increase the amount of data in the public package, and motion sensors (accelerometers, gyroscopes, magnetometers) can reflect The user's walking direction, and according to the existing mobile phone behavior identification APP can output various behavior states of the user walking with the mobile phone, including the user behavior state of the user on stairs, elevators, escalators, etc., therefore, the embodiment of the present application On the one hand, the state information output by the behavior recognition of the mobile terminal can be used to detect cross-layers. On the other hand, the mobile terminal with the barometer can detect the landmark point of the channel by detecting the significant height change switching point according to the height change of the air pressure response. Further, the behavior recognition results and the barometer cross-layer event results can be combined to determine the landmark points of the channel, so as to realize the prediction of the landmark points of the channel robustly, and improve the prediction quality and accuracy.

Based on the first aspect, in a possible embodiment, the fingerprint map of the multi-storey building is generated according to the leveling framework of the at least two floors, the connection position of the passage and the configuration of the passage, The method includes: generating a three-dimensional topological map according to the leveling skeleton, the location points of the passage and the configuration of the passage; and mapping the three-dimensional topological map to a world coordinate system to obtain a fingerprint map of the multi-storey building.

For example, the information obtained by the mobile terminal based on the satellite navigation module that can be included in the crowdsourced data, for example, the mobile terminal can detect the GPS information of the entrance and exit of the reference floor (referred to as entrance GPS) through the GPS module, or the GPS information obtained occasionally in the indoor environment (referred to as Opportunistic GPS). The server obtains the GPS of the entrance or the indoor opportunity GPS by identifying the position of the entrance and exit of the reference floor. Then, the 3D topology map can be mapped to the world coordinate system, such as the WGS84 coordinate system, by using the entrance and exit GPS or indoor opportunistic GPS, so as to realize the absolute coordinate mapping of the whole floor, that is, to obtain the fingerprint map of the multi-story building in the world coordinate system.

By implementing the embodiments of the present application, a fingerprint map of a multi-story building under the WGS84 coordinate system can be obtained without the aid of an indoor floor plan, and a full-floor map coordinate mapping can be realized, so as to achieve a seamless indoor and outdoor positioning effect of multi-floor and full-scene.

Based on the first aspect, in a possible embodiment, the process of determining the configuration of the channel according to the crowdsourcing data includes: performing pedestrian dead reckoning (PDR) according to the motion data, and obtaining the location where the mobile terminal is located. PDR information when moving in the channel; the PDR information includes one or more of changes in speed, number of steps, distance, and height; and the configuration of the channel is determined according to the PDR information.

For example, after the pedestrian dead reckoning (PDR) is performed to obtain the PDR trajectory, according to the detected start and end time stamps, corresponding to the PDR trajectory time stamps, the position of the start and end points in the trajectory is determined to realize single-track landmark point detection. At the same time, according to the behavior between the starting and ending points detected by the PDR, the status, speed/step number, distance, height change, etc. are identified, so as to distinguish the specific configuration of the passage, such as stairs, elevators or escalators, etc., and assign the passage corresponding up and down row direction property.

In a second aspect, an embodiment of the present application proposes an indoor positioning method, which is applied to a first mobile terminal, and includes: downloading a fingerprint map from a server, where the fingerprint map is a The leveling frame of the multi-storey building determined by the package data, the position points on the leveling frame and the configuration of the passage are generated; the real-time position of the first mobile terminal in the multi-storey building is collected. fingerprint data; matching the location fingerprint data with the fingerprint map to obtain the real-time location of the first mobile terminal in the multi-story building, where the real-time location is a floor in the multi-story building or in the passageway between two floors.

It can be seen that the embodiment of the present application actually provides an indoor offline positioning method. Through the offline positioning method, offline positioning can be performed by using the fingerprint map cached in advance in the case of no network signal or poor network signal. In addition, it can also be beneficial to combine the data of sensors, GPS, etc. for fusion positioning, so as to further improve the accuracy of indoor positioning.

In a third aspect, an embodiment of the present application provides an indoor positioning method, wherein the method is applied to a server, including:

receiving a location request and the location fingerprint data from the first mobile terminal;

The location fingerprint data is matched with the fingerprint map to obtain a positioning result; the fingerprint map is the flat skeleton of the multi-story building determined by the server according to the crowdsourcing data of a plurality of second mobile terminals, in the generated from the position points on the flat skeleton and the configuration of the channel;

The first mobile terminal sends the positioning result, and the positioning result is used to indicate the real-time position of the first mobile terminal in the multi-storey building, and the real-time position is the location in the multi-storey building. A position in a floor or in a passageway between two floors.

In addition, for the first mobile terminal, the corresponding method includes:

collecting real-time location fingerprint data of the first mobile terminal in a multi-storey building;

Send the location request and the location fingerprint data to the server, so that the server can match the location fingerprint data with the fingerprint map to obtain the location result; Generated by the leveling frame of the multi-storey building determined by the package data, the position point on the leveling frame and the configuration of the passage;

receiving the positioning result from the server, where the positioning result is used to indicate the real-time position of the first mobile terminal in the multi-storey building, where the real-time position is in a floor in the multi-storey building location or in the passageway between two floors.

It can be seen that the embodiment of the present application actually provides an indoor online positioning method. Through the online positioning method, the cloud server can provide the mobile terminal positioning service interface, and according to the received positioning request of the mobile terminal and the position fingerprint data. , through the positioning algorithm, use the latest fingerprint map to locate the mobile terminal, and return the positioning result to the mobile terminal, which improves the convenience of indoor positioning of the mobile terminal and saves data storage overhead.

In a fourth aspect, an embodiment of the present application proposes a device for constructing a fingerprint map of a multi-story building, the device comprising:

a receiving module, configured to receive crowdsourcing data from a plurality of mobile terminals, where the crowdsourcing data includes motion data and position fingerprint data collected by the mobile terminal during a process of moving across floors in a multi-storey building;

A fingerprint map building module for determining, according to the crowdsourcing data, the leveling skeletons of at least two floors in the multi-storey building and the positions of the passages connected between the leveling skeletons on the leveling skeletons point, and the configuration of the passage; generating a fingerprint map of the multi-storey building according to the leveling framework of the at least two floors, the location points of the passage and the configuration of the passage.

In a specific embodiment, the functional modules of the apparatus are used to implement the method described in any embodiment of the first aspect.

In a fifth aspect, an embodiment of the present application provides an apparatus for indoor positioning, where the apparatus is applied to a first mobile terminal, including:

The receiving module is used for downloading a fingerprint map from the server, where the fingerprint map is the leveling skeleton of the multi-story building and the position on the leveling skeleton determined by the server according to the crowdsourcing data of a plurality of second mobile terminals generated by the configuration of the points and channels;

a data collection module, configured to collect real-time location fingerprint data of the first mobile terminal in the multi-storey building;

A local positioning module for matching the location fingerprint data with the fingerprint map to obtain the real-time location of the first mobile terminal in the multi-story building, where the real-time location is in the multi-story building A position in a floor in a , or a position in a passageway between two floors.

In a specific embodiment, the functional modules of the apparatus are used to implement the method described in the second aspect.

In a sixth aspect, an embodiment of the present application provides an apparatus for indoor positioning, where the apparatus is applied to a server, including:

a receiving module for receiving a positioning request and the position fingerprint data from the first mobile terminal;

The cloud positioning service module is used for matching the position fingerprint data with a fingerprint map to obtain a positioning result; the fingerprint map is the data of the multi-story building determined by the server according to crowdsourcing data of a plurality of second mobile terminals. Generated by the leveling framework, the position points on the leveling framework and the configuration of the channel;

A sending module, configured to send the positioning result to the first mobile terminal, where the positioning result is used to indicate the real-time position of the first mobile terminal in the multi-storey building, where the real-time position is in the A location in a floor in a multi-story building or in a passageway between two floors.

In a specific embodiment, the functional modules of the apparatus are used to implement the method described in the third aspect.

In a seventh aspect, an embodiment of the present application provides a chip, the chip includes a processor and a data interface, and the processor reads an instruction stored in a memory through the data interface to execute the first aspect or the second aspect or the method described in any embodiment of the third aspect.

In an eighth aspect, an embodiment of the present invention provides another non-volatile computer-readable storage medium; the computer-readable storage medium is used to store the data described in any embodiment of the first aspect or the second aspect or the third aspect. method's implementation code. The program code, when executed by a computing device, can implement the method described in any embodiment of the first aspect or the second aspect or the third aspect.

In a ninth aspect, an embodiment of the present invention provides a computer program product; the computer program product includes program instructions, and when the computer program product is executed by a computing device, executes any of the foregoing first aspect or second aspect or third aspect methods described in the examples. The computer program product may be a software installation package, and the computer program product may be downloaded and executed on the controller to implement the method described in any embodiment of the first aspect or the second aspect or the third aspect.

Description of drawings

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only These are some embodiments of the present application. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings without any creative effort.

FIG. 1 is a schematic diagram of a scenario for explaining terms provided by an embodiment of the present application;

2 is a schematic diagram of a system architecture provided by an embodiment of the present application;

3 is a schematic structural diagram of a device in a system architecture provided by an embodiment of the present application;

4 is a schematic diagram of a system architecture including functional modules provided by an embodiment of the present application;

5 is a schematic flowchart of a map construction method provided by an embodiment of the present application;

6A is a schematic flowchart of a positioning method provided by an embodiment of the present application;

6B is a schematic flowchart of another positioning method provided by an embodiment of the present application;

7 is a schematic flowchart of a mobile phone operation scenario provided by an embodiment of the present application;

8 is a schematic diagram of a fingerprint map construction module provided by an embodiment of the present application;

9 is a schematic diagram of a channel self-learning module provided by an embodiment of the present application;

FIG. 10 is a schematic diagram of a multi-sensor fusion scenario provided by an embodiment of the present application;

11 is a schematic diagram of a scenario of trajectory segmentation provided by an embodiment of the present application;

12 is a schematic diagram of a scene in which a trajectory and a skeleton are matched according to an embodiment of the present application;

13 is a schematic diagram of a scene of KNN registration provided by an embodiment of the present application;

14 is a schematic diagram of a scenario based on fingerprint similarity registration provided by an embodiment of the present application;

FIG. 15 is a schematic diagram of a skeleton sequence, floor identification, and alignment provided by an embodiment of the present application;

16 is an example diagram of a trajectory and skeleton matching scene provided by an embodiment of the present application;

17 is a schematic diagram of floor alignment provided by an embodiment of the present application;

18 is a schematic diagram of a directed acyclic graph construction process provided by an embodiment of the present application;

Fig. 19 is a scene example diagram of absolute floor recognition provided by an embodiment of the present application;

Fig. 20 is another scene example diagram of absolute floor recognition provided by an embodiment of the present application;

21 is an example diagram of several channel configurations provided by an embodiment of the present application;

FIG. 22 is an example diagram of a skeleton alignment scenario provided by an embodiment of the present application;

23 is an example diagram of some skeleton alignment scenarios provided by an embodiment of the present application;

FIG. 24 is a schematic diagram of a map mapping scenario provided by an embodiment of the present application;

FIG. 25 is a schematic diagram of a scenario of a fingerprint map provided by an embodiment of the present application.

detailed description

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. It should be noted that the terms used in the embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to limit the present application.

In the embodiments of the present application, words such as "exemplary" or "for example" are used to represent examples, illustrations or illustrations. Any embodiments or designs described in the embodiments of the present application as "exemplary" or "such as" should not be construed as preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary" or "such as" is intended to present the related concepts in a specific manner.

As used in the embodiments of this application or the appended claims, the singular forms "a," "the," and "the" are intended to include the plural forms as well, unless the context clearly dictates otherwise. It will also be understood that the term "and/or" as used herein refers to and includes any and all possible combinations of one or more of the associated listed items. As used herein, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly dictates otherwise. It will also be understood that the terms "comprising", "having", "including" and/or "comprising" when used herein designate the presence of stated features, integers, steps, operations, means, and/or components, but do not The presence or addition of one or more other features, integers, steps, operations, elements, components and/or groups thereof is not excluded.

It will also be understood that the term "and/or" as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

In order to better understand the solutions of the present application, some terms involved in the embodiments of the present application are explained first.

(1) Crowdsourced data:

Crowdsourcing technology is a method of obtaining resources through a large number of user terminals. Through a certain trigger mechanism, the available data such as sensors, radio frequency signals or network positioning of user terminals are collected without the user's perception. Such data Can be called crowdsourced data. For example, the crowdsourced data collected by the mobile terminal in this embodiment of the present application may include motion data and location fingerprint data, and the motion data may be data collected by sensors such as a speedometer, a gyroscope, a magnetometer, and a GPS, such as motion speed, attitude, rotation, etc. Angle, moving direction, etc., the location fingerprint data can be the radio frequency information scanned by WiFi technology and Bluetooth technology, such as WiFi signal strength (Received Signal Strength, RSS) information, the number of wireless access points (Access Point, AP) ,etc. In addition, the crowdsourced data can also include altitude data collected by barometers. In an indoor environment, a user can walk around with a mobile terminal such as a smartphone. With the authorization of the user in advance, the mobile terminal can autonomously collect crowdsourced data during the movement.

(2) Fingerprint map:

Fingerprint maps are fingerprint datasets with absolute location information that can be used for indoor positioning. Fingerprints, also known as location fingerprints or signal fingerprints, refer to the characteristic sequence of wireless signals received at a specific location indoors. The wireless signal may be a WiFi signal, a Bluetooth signal, or the like. For example, the fingerprint of the WiFi signal can be characterized by the statistical value of the received signal strength (Received Signal Strength, RSS) of the WiFi signal from different wireless access points (Access Point, AP) and the number of APs. It can be understood that the WiFi signal strength and the number of APs detected at each location in the indoor space are different. Therefore, when the indoor AP is fixedly deployed, each location point in the indoor space is different from that location. The location fingerprint data collected by the point has a one-to-one correspondence.

(3) Multi-story buildings, floors, passages and landmarks:

Referring to the left figure in FIG. 1 , a multi-story building is a building with multiple floors, and the solution of the present application is suitable for the construction of a fingerprint map of a multi-story building. The floors in the multi-storey building may be floors above the surface such as the first, second, and third floors, or may be floors below the surface such as the basement, minus 1 floor, and minus 2 floors.

The passage is a passage facility used to connect different floors of a multi-storey building. The passage is connected between different floors, and people and goods can be transferred between different floors through the passage. The connection relationship between the passages between floors is determined by the attributes of the passage, which include: the connection position of the passage on the floor, and the structural type (or configuration) of the passage.

There are various configurations of passageways, such as stairs, elevators, escalators, ladders, steps, ramps, etc. in:

Stairs are the main access facilities in multi-storey buildings for people to walk between different floors and for emergency evacuation.

An elevator is a passage facility that uses electricity to drive the car to run in a vertical direction to transport people or goods.

Escalator, also known as escalator, is a passage facility that automatically transports people through chain conveyors.

A ladder refers to a passage facility with a gradient of more than 45° and requires the help of both hands to go up and down.

The steps are stair-shaped for different indoor elevations, which is convenient for people to climb up the steps.

A ramp is a ramp-type passage facility for wheeled vehicles in a multi-storey building.

The connection position of the passage on the floor is the end point of the passage on the floor, which is simply understood as the position of the entrance and exit of the passage on the floor. For example, for the elevator between the 1st floor and the 2nd floor, the connection position of the elevator on the 1st floor is the entrance and exit position of the elevator on the 1st floor, and the connection position on the 2nd floor is the entrance and exit position of the elevator on the 2nd floor.

For the convenience of description, in this article, "the location of the passage on the floor", "the connection position of the passage on the floor", "the end point of the passage on the floor", and "the location of the entrance and exit of the passage on the floor" can also be uniformly used as "ground". Punctuation" for short.

It can be understood that, for channels of different configurations, the connection relationship between floors is also different. According to the connection relationship between the passage and the floor, it can be divided into directly connected passages (such as elevators, ladders, some stairs folded in the middle) and inclined connected passages (such as inclined stairs, escalators, steps, ramps). For a direct channel, its endpoints on different floors are vertically aligned with no horizontal offset. For the inclined channel, its endpoints on different floors are horizontally offset.

It should be noted that the configurations, quantities and connection positions of passages between different floors may be the same or different, which is not limited in the present application. For example, the 1st and 2nd floors can be connected by elevators, escalators and stairs, and the 2nd and 3rd floors can be connected by elevators and stairs.

It should also be noted that passages are not necessarily connected only to adjacent floors, but may also be connected across floors. For example, an escalator may be designed to directly connect the 1st floor and the 3rd floor; another example, an elevator may be configured to be configured to go from the 1st floor to the 4th floor without stopping at other floors, and so on.

The fingerprint map constructed in the embodiment of the present application may be a three-dimensional (3D) fingerprint map. That is to say, there is not only the fingerprint map information corresponding to the floor, which is used for the positioning of the mobile terminal on the floor, but also the fingerprint map information corresponding to the channel, which is used for the positioning of the mobile terminal on the channel.

In addition, the fingerprint map constructed in the embodiment of the present application may be a fingerprint map corresponding to a part of floors in a multi-story building. For example, a multi-story building has 5 floors, but based on the deployment of APs, only the fingerprint map of floors 1-3 can be constructed to realize the positioning of floors 1-3.

(4) 3D topological map and layer skeleton:

Referring to the right figure in FIG. 1 , in the embodiment of the present application, the three-dimensional (3D) topology map is a three-dimensional topology structure generated by an aggregation algorithm based on the movement trajectories formed by crowdsourcing data of multiple mobile terminals, and the 3D topology map includes multiple Each leveling skeleton corresponds to a floor, and each position point in the leveling skeleton is bound with the position fingerprint data. It can be seen that the landmark points in the floor are also projected to the corresponding positions on the leveling skeleton. After the 3D topological map is generated in the embodiment of the present application, the fingerprint map (or 3D fingerprint map) can be obtained by mapping the 3D topological map to the world coordinate system (eg, the WGS84 coordinate system).

The fingerprint map construction method described in this application can be executed by a server, and the mobile terminal provides crowdsourced data to the server. The positioning method described in this application can also be jointly implemented through the interaction between the mobile terminal and the server.

Among them, the mobile terminal can be a smart phone, a tablet computer, a notebook computer, a handheld computer, a mobile internet device (MID, mobile internet device), a wearable device (such as a smart bracelet, a smart watch, etc.), a special AR device, a special VR equipment, camera equipment (such as video recorders, smart cameras, digital cameras, video cameras, etc.) or other equipment.

A mobile terminal may also be referred to as user equipment (UE), subscriber station, mobile unit, subscriber unit, wireless unit, remote unit, mobile device, wireless device, wireless communication device, remote device, mobile subscriber station, terminal device, access Terminal, electronic device, wireless terminal, smart terminal, remote terminal, handset, user agent, mobile client, client, or some other appropriate term.

The server can be an independent server, a cloud platform, a data center, or a server cluster. The type of server can be, for example, a database server, an application server, a general purpose server, a dedicated server, and the like. The server may also include one or more processing nodes, or include one or more virtual machines running on the server.

Referring to FIG. 2 , FIG. 2 shows a system architecture in which a terminal interacts with a server. The system architecture includes a mobile terminal 10 and a server 20 . Communication between the mobile terminal 10 and the server 20 is possible through a network.

When the system architecture is the system architecture of the fingerprint map construction process, the server 20 can receive crowdsourced data from a large number of mobile terminals 10, and construct fingerprint maps of one or more multi-story buildings according to the crowdsourced data of these mobile terminals. Wherein, the present application does not limit the type and quantity of these mobile terminals 10 , nor does it limit the communication mode between the mobile terminal 10 and the server 10 .

When the system architecture is the system architecture of the indoor positioning process, the real-time positioning of the terminal can be realized in at least two ways:

(1) The method of cloud positioning service: the server 20 provides the mobile terminal 10 with a positioning service interface, and the server 20 uses the pre-built fingerprint map to locate the mobile terminal according to the received positioning request and the position fingerprint data of the mobile terminal 10 through the positioning algorithm. The terminal 10 performs positioning and returns the positioning result to the mobile terminal 10 . The server 20 may also push the fingerprint map of the building where the mobile terminal 10 is located to the mobile terminal 10, so as to support the positioning function of the mobile terminal 10 in the case of no network or poor network signal.

(2) The method of local positioning of the terminal: the mobile terminal 10 downloads the fingerprint map from the server 20 in advance, and when it is judged that there is no current network signal or the network signal is poor, it no longer needs to interact with the server, but directly uses the current acquisition The location fingerprint data and the locally stored fingerprint map can be used to achieve positioning. In addition, it can also be further combined with data such as motion sensors and GPS for fusion positioning.

Please refer to Figure 3 to better understand the internal structure of the device under the device-cloud system. As shown in FIG. 3 , the end-cloud system includes a mobile terminal 10 and a server 20 . The mobile terminal and the server 20 establish a communication connection through their respective communication devices, and the communication method can be a wireless communication method. The wireless communication method includes but is not limited to: WiFi method, radio frequency (Radio Frequency, RF) method, data communication method, Bluetooth method, etc. one or more of etc.

For mobile terminal 10:

The mobile terminal 10 comprises: at least one processor 13, memory 15, at least one sensor 11, satellite positioning means 14, display means 17, signal detection means 12 and communication means 16, which components can communicate on one or more communication buses to The functions of the mobile terminal 10 are realized.

The processor 13 may also be referred to as a central processing unit (CPU, central processing unit), and the processor 13 may specifically include one or more processing units, for example, the processor 13 may include a general-purpose processor, an application processor (application processor, AP) ), modem processor, graphics processor (graphics processing unit, GPU), image signal processor (image signal processor, ISP), controller, video codec, digital signal processor (digital signal processor, DSP) , baseband processor, and/or neural-network processing unit (NPU), etc. Wherein, different processing units may be independent devices, or may be integrated in one or more processors.

In some embodiments, processor 13 may include one or more interfaces. The interface may include an integrated circuit (inter-integrated circuit, I2C) interface, an integrated circuit built-in audio (inter-integrated circuit sound, I2S) interface, a pulse code modulation (pulse code modulation, PCM) interface, a universal asynchronous transceiver (universal asynchronous transmitter) receiver/transmitter, UART) interface, mobile industry processor interface (MIPI), general-purpose input/output (GPIO) interface, subscriber identity module (SIM) interface, and / or universal serial bus (universal serial bus, USB) interface, etc.

The communication device 16 is used for receiving and transmitting data, and mainly integrates the receiver and the transmitter of the mobile terminal 10 . In one embodiment, the communication device 16 may include, but is not limited to, an antenna system, an RF transceiver, one or more amplifiers, a tuner, one or more oscillators, a digital signal processor, a CODEC chip, a SIM card, and storage medium, etc. In some embodiments, the communication device 16 may be implemented on a separate chip. The communication device 16 may, for example, support data network communication through at least one of 2G/3G/4G/5G, etc., and/or support at least one of the following short-range wireless communication methods: Bluetooth (Bluetooth, BT) communication, Wireless Fidelity (WiFi) communication, Near Field Communication (NFC), Infrared (Infrared, IR) wireless communication, Ultra Wide Band (UWB, Ultra Wide Band) communication, ZigBee communication. The mobile terminal 10 communicates and interacts with the server 20 through the communication device 16 . For example, crowdsourced data or a positioning request is sent to the server 20 , and for example, a positioning result or a fingerprint map sent by the server 20 is received.

The signal detection device 12 is used for scanning/detecting the radio frequency information of the wireless access point (Access Point, AP) in the indoor environment through WiFi technology, Bluetooth technology, etc., such as detecting WiFi signal strength (Received Signal Strength, RSS) information, The number of wireless access points (Access Point, AP), etc. to obtain the location fingerprint data of the current location.

In a possible embodiment, the function of the signal detection device 12 can also be integrated into the communication device 16, that is, the communication device 16 can not only communicate with the server 20, but also be used to obtain the location fingerprint data of the current location. For example, after the communication device 16 enters the WiFi communication mode, it can not only scan the WiFi signal in the indoor environment to obtain the location fingerprint data, but also send the crowdsourced data to the server in the form of WiFi.

In yet another possible embodiment, the functions of the signal detection device 12 can also be integrated into the processor 13, and after the communication device 16 scans the indoor environment for signals, the processor 13 performs signal analysis to obtain location fingerprint data.

The sensor 11 may include one or more sensors, such as a barometer, a magnetometer, a gyroscope, an accelerometer, a wheel speedometer, an inertial measurement unit (Inertial Measurement Unit, IMU), etc., for detecting when the mobile terminal 10 moves ( For example, the user carries the movement data of the mobile terminal 10, such as movement speed, posture, rotation angle, movement direction, etc., so as to form crowdsourced data in combination with the above-mentioned position fingerprint data. Motion data can also be used to assist positioning during indoor positioning. Of course, the sensor 11 may also include more or other sensors.

The satellite positioning device 14 is used to realize satellite positioning of the mobile terminal 10. For example, the satellite positioning device 14 may be provided with a Global Navigation Satellite System (Global Navigation Satellite System, GNSS), and GNSS is not limited to the Global Positioning System (Global Positioning System, GPS) , GLONASS positioning system (GLONASS), Galileo satellite navigation and positioning system (Galileo satellite navigation system), Beidou positioning system, etc. to assist positioning. For example, the mobile terminal 10 may obtain GPS signals at the entrances and exits of a multi-story building through the satellite positioning device 14 (referred to as entrance and exit GPS), or may also obtain GPS signals (referred to as opportunistic GPS) when the signals at certain indoor locations allow.

The mobile terminal 10 may include one or more display devices 17 . The mobile terminal 10 may jointly implement the display function through the display device 17, a graphics processing unit (GPU) in the chip 310, an application processor (AP), and the like. The GPU is a microprocessor for image processing, and connects the display device 17 and the application processor. The GPU is used to perform mathematical and geometric calculations for graphics rendering. The display device 17 is used to display the interface content currently output by the system, the interface content may include the interface of the running application program and the system level menu, etc., and may be specifically composed of the following interface elements: input interface elements, such as buttons (Button), text Input box (Text), slide bar (Scroll Bar), menu (Menu), etc.; and output interface elements, such as window (Window), label (Label), image, video, animation, etc.

In a specific implementation, the display device 17 may be a display panel, a lens (eg, VR glasses), a projection screen, and the like. The display panel may also be called a display screen, for example, it may be a touch screen, a flexible screen, a curved screen, etc., or other optical components. That is to say, when the electronic device has a display screen in this application, the display screen can be a touch screen, a flexible screen, a curved screen or other forms of screen, and the display screen of the electronic device has the function of displaying images. Shape This application does not make any limitation in this regard.

For example, when the display device 17 includes a display panel, the display panel may use a liquid crystal display (LCD), an organic light-emitting diode (OLED), an active matrix organic light emitting diode or an active matrix Organic light emitting diodes (active-matrix organic light emitting diodes, AMOLED), flexible light emitting diodes (flex light-emitting diodes, FLED), Miniled, MicroLed, Micro-oLed, quantum dot light emitting diodes (quantum dot light emitting diodes, QLED), etc.

In addition, in some specific implementations, the touch panel and the display panel in the display device 17 may be coupled together and disposed, for example, the touch panel may be disposed below the display panel, and the touch panel may be used to detect a user's input of a touch operation through the display panel Touch pressure acting on the display panel (such as clicking, sliding, touching, etc.), and the display panel is used for content display. The content displayed in the embodiment of the present application may be, for example, a map application operation interface, a fingerprint map, an indoor positioning interface, and the like.

The memory 15 can be connected with the processor 13 through a bus, or can be coupled with the processor 13, and is used for storing various software programs and/or groups of instructions. In a specific implementation, the memory 15 may include a high-speed random access memory (such as a cache memory), and may also include a non-volatile memory, such as a random access memory (RAM), a read-only memory (Read- Only Memory, ROM), Erasable Programmable Read Only Memory (EPROM), or Portable Read-Only Memory (Compact Disc Read-Only Memory, CD-ROM) and so on.

The memory 15 can store an operating system, such as ANDROID, IOS, WINDOWS, Hongmeng operating system or LINUX embedded operating system, and so on.

Memory 15 may be used to store data (eg crowdsourced data, fingerprint map data).

Memory 15 may also store communication programs that may be used to communicate with one or more servers or other devices.

The memory 15 may also store one or more application programs. As shown in the figure, these applications may include: map applications, virtual scene applications such as AR/VR/MR, shopping applications, life service applications, and the like.

The memory 15 can also store a user interface program, which can vividly display the content of the application program through the graphical operation interface and present it through the display device 17, and realize the reception through input controls such as menus, dialog boxes and keys. The user's control operation on the application.

In specific embodiments of the present application, the memory 15 may store codes including methods of any of the embodiments discussed herein, and/or codes of functional modules. The processor 13 is configured to invoke the code in the memory 15 to implement the functions on the side of the mobile terminal 10 in any of the embodiments herein.

For server 20:

The server 20 may include one or more processors 21 , one or more memories 23 , a communication device 22 . These components can be connected via a bus.

The processor 21 may be the processor 13 and may include a general-purpose processor, an application processor (application processor, AP), a modem processor, a graphics processor (graphics processing unit, GPU), an image signal processor (image signal processor, ISP), controller, video codec, digital signal processor (DSP), baseband processor, and/or neural-network processing unit (NPU), etc. The processor 21 is mainly used to realize the construction of fingerprint maps of one or more multi-story buildings.

The communication device 22 is used to realize the communication between the server 20 and the mobile terminal, for example, a receiver and a transmitter can be integrated, wherein the receiver is used to receive data (such as positioning requests, crowdsourcing data, etc.) sent by various mobile terminals, and the transmitter is used to send data to the mobile terminal. The mobile terminal sends data (eg, positioning results, fingerprint map data, etc.).

The memory 23 can be connected with the processor 21 through a bus, or can be coupled with the processor 21 for storing various software programs and/or multiple sets of instructions, and data (eg, fingerprint map data, crowdsourced data, etc.). In a specific implementation, the memory 23 includes, but is not limited to, a random access memory (Random Access Memory, RAM), a read-only memory (Read-Only Memory, ROM), an erasable programmable read-only memory (Erasable Programmable Read Only Memory, EPROM), or portable read-only memory (Compact Disc Read-Only Memory, CD-ROM).

In specific embodiments of the present application, the memory 15 may store codes including methods of any of the embodiments discussed herein, and/or codes of functional modules. The processor 21 may be configured to invoke program instructions in the memory 23 to perform the server-side functions in any of the embodiments herein.

It should be understood by those skilled in the art that the mobile terminal 10 and the server 20 may also include more or less components than those shown, or combine certain components, or arrange different components. The device structure shown in FIG. 2 does not constitute a limitation to the present application.

Additionally, the term "coupled" as used herein means connected directly to, or connected through one or more intervening components or circuits. Any of the signals described herein provided on the various buses may be time multiplexed with other signals and provided on one or more shared buses. Additionally, the interconnections between various circuit elements or software blocks may be shown as buses or as single signal lines. Each bus is alternatively a single signal line, and each single signal line is alternatively a bus, and the single line or bus may represent any one or more of a number of physical or logical mechanisms for communication between components .

The following further describes the system architecture provided by the embodiments of the present application from the level of function implementation.

Referring to FIG. 4 , FIG. 4 is a functional implementation block diagram of a possible system architecture provided by an embodiment of the present application. The system architecture includes a mobile terminal 10 and a server 20. As shown in the figure, the mobile terminal 10 may include a local positioning module 101, a database 102 for saving crowdsourced data collected by the mobile terminal 10, and a database 103 for saving offline fingerprint maps ; The functional modules in the mobile terminal 10 may, for example, run on the processor 13 shown in the embodiment of FIG. 3 , and the database in the mobile terminal 10 may, for example, store data through the memory 15 shown in the embodiment of FIG. 3 . The server 20 may include a fingerprint map construction module 201 , a database 202 for storing crowdsourced data from the mobile terminal 10 , a database 203 for storing the constructed fingerprint map, a data normalization module 204 and a cloud location service module 205 . In addition, although not shown in the figure, the mobile terminal 10 may include a receiving module and a sending module, which are respectively used to receive data and send data from the outside world (such as a server), and the server 20 may also include a receiving module and a sending module, which are respectively used to send data from the outside world. (such as a mobile terminal) to receive data and send data. The functional modules in the server 20 may run on, for example, the processor 21 shown in the embodiment of FIG. 3 , and the database in the server 20 may, for example, store data through the memory 23 shown in the embodiment of FIG. 3 . The relevant descriptions are as follows:

(1) Database 102: In the map construction stage, the mobile terminal 10 can store the crowdsourced data collected by the mobile terminal 10 in the multi-story building through the database 102. The crowdsourced data includes two parts: motion data collected by sensors and data collected by signals The location fingerprint data collected by the device. The sensor here is, for example, the sensor 11 shown in the embodiment of FIG. 3 , for example, relevant data is obtained through a speedometer, gyroscope, magnetometer, barometer, GPS, etc., and the signal acquisition device here is, for example, shown in the embodiment of FIG. 2 . The signal collection device 12 detects the radio frequency information of the surrounding APs by means of WiFi, Bluetooth and the like. The mobile terminal 10 can upload its own crowdsourced data to the cloud server 20 to support the server's collection of the crowdsourced data.

(2) Database 103: used to store the fingerprint map data downloaded by the mobile terminal 10 from the server 20 in advance to support the offline positioning of the mobile terminal 10 when there is no network signal or the network signal is poor, so the fingerprint map data can also be called Fingerprint map for offline.

(3) Local positioning module 101 : In the indoor positioning stage, the mobile terminal 10 sends a positioning request to the server 20 according to the position fingerprint data collected at the current position, and receives the positioning result returned by the server 20 . In the case of no network signal or poor network signal, offline positioning can be performed by using the offline fingerprint map cached in the database 203 in advance, and fusion positioning can also be performed in combination with data such as sensors and GPS to further improve the accuracy of indoor positioning.

(4) Database 202: used to store crowdsourced data uploaded by various mobile terminals, the database 102 can be used as a data source for the server 20 to construct a fingerprint map.

(5) Database 103: used to store the fingerprint maps of various multi-story buildings constructed by the server 20 to support indoor positioning service of the mobile terminal by the server 20, and also support the download of part or all of the fingerprint maps by the mobile terminal.

(6) Data regularization module 204: used to implement regularization of crowdsourced data uploaded by various mobile terminals. For example, for a large amount of crowdsourced data uploaded by different mobile terminals, the data regularization module 204 can verify the valid crowdsourced data through data standard verification; it can also identify valid crowdsourced data according to GPS information and AP information (such as WiFi signal characteristics, Bluetooth signal Features, etc.), according to the initial grid division, the coarse granularity is classified and stored according to the country, city and other information. It supports the construction of fingerprint maps of different multi-story buildings through efficient classification of data.

(7) Fingerprint map building module 201 : for building a fingerprint map of a multi-story building according to the database 202 . Specifically, when constructing a fingerprint map of a certain multi-story building, the respective leveling skeletons of at least two floors in the multi-story building and the connections of the channels connected to the respective leveling skeletons can be determined according to crowdsourcing data The connection relationship of the channel is determined by the connection position of the channel on the leveling frame and the configuration of the channel; and then according to the connection relationship between the leveling frame of the at least two floors and the channel, Generate a fingerprint map of the multi-story building.

In a specific implementation, the fingerprint map construction module 201 may further include the following functional modules:

The channel self-learning module is used to determine the position of the channel between each leveling frame in the multi-storey building on the leveling frame according to the crowdsourced data, and can also be used to determine the configuration of the channel. For example, the channel self-learning module can fuse the mobile behavior recognition results and the cross-layer event detection results, and use the leveling segment to splicing the leveling skeleton to initially locate the candidate landmark points. Combined with the WiFi information similarity across layers and the spatial distribution of candidate landmark points, joint clustering is used to accurately locate landmark points on the flat skeleton of multi-story buildings.

The 3D topology building module is used to generate a 3D topology map according to the leveling skeleton of the multi-story building, the location points of the channel and the configuration of the channel. For example, the 3D topology building module can identify the uplink and downlink connections between the leveling skeletons based on the uplink and downlink connections of the candidate landmark points located on the leveling skeletons, and then sort the leveling skeletons according to their high-low relationship. , identify the absolute floor, and finally align the floor with the coordinate system according to the configuration of the identified passage, so as to obtain the 3D topology map of the multi-story building.

The map mapping module is used for mapping the three-dimensional topological map to the world coordinate system to obtain the fingerprint map of the multi-storey building. For example, the map mapping module can convert the 3D topology map of the multi-story building to the WGS84 coordinate system according to the identified entrance and exit GPS, opportunity GPS points, output control points or local floor maps, so as to obtain the final fingerprint map (or 3D map). Fingerprint map) to achieve seamless indoor and outdoor positioning.

(8) Cloud positioning service module 205: The cloud positioning service module 205 can provide the mobile terminal positioning service interface, and according to the received positioning request and position fingerprint data of the mobile terminal, through the positioning algorithm, the fingerprint map in the database 203 is used to locate the mobile terminal. The terminal performs positioning and returns the positioning result to the mobile terminal. In addition, the fingerprint map of the building can also be pushed to the mobile terminal to support the offline positioning function of the mobile terminal when there is no network or poor network signal.

For the method embodiments described below, for the sake of convenience, they are all expressed as a combination of a series of action steps, but those skilled in the art should know that the specific implementation of the technical solutions of the present application is not affected by the described series of steps. restrictions on the sequence of action steps.

Based on the system architecture described above, the following describes a method for constructing a fingerprint map of a multi-story building provided by an embodiment of the present application. Referring to FIG. 5 , the method is described from the perspective of the server side, including but not limited to the following steps:

S301. Receive crowdsourced data from multiple mobile terminals, where the crowdsourced data includes motion data and location fingerprint data collected by the mobile terminals in the process of moving across floors in a multi-story building.

The motion data may be data collected by sensors such as a speedometer, a gyroscope, a magnetometer, and a GPS, such as motion speed, attitude, rotation angle, movement direction, and the like. The location fingerprint data can be radio frequency information scanned by WiFi technology or Bluetooth technology, such as WiFi signal strength (Received Signal Strength, RSS) information, the number of wireless access points (Access Point, AP), and so on. In addition, the crowdsourced data can also include altitude data collected by barometers.

The multiple mobile terminals may be different types of mobile terminals, and the sensors configured on each mobile terminal may also be different. In the present application, the collection of crowdsourced data by various users' mobile terminals replaces the collection of manual data and manual annotation, and there is no need to deploy a special data collection device indoors, which greatly reduces the cost of labor and material resources.

S302. Determine, according to the crowdsourcing data, the leveling frames of at least two floors in the multi-storey building, the position points on the leveling frames of the passages connected between the leveling frames, and the channel configuration.

Each leveling skeleton corresponds to a floor, and each location point in the leveling skeleton is bound with the location fingerprint data. The passage is a passage facility used to connect different floors of a multi-storey building. The configuration of the passage includes stairs, elevators, escalators, ladders, steps, ramps, etc. The location point of the passage on the floor is the end point of the passage on the floor. The simple understanding is the position of the entrance and exit of the passage on the floor. In this article, the location point of the passage can also be called the landmark point, so each passage has a pair of ground punctuation. For the concepts of the leveling skeleton, the landmark point, and the channel configuration, reference may be made to the relevant descriptions before and after this application, which will not be repeated here.

In a specific embodiment of the present application, the process of generating a leveling skeleton may include: obtaining movement trajectories of the plurality of mobile terminals in the at least two floors according to motion data of the mobile terminals; The movement trajectories of the terminal are aggregated to obtain the leveling skeleton of the at least two floors; wherein, each position point of the leveling skeleton has corresponding position fingerprint data.

In a specific embodiment of the present application, the process of determining the location of the landmark point of the passage may include: obtaining the movement trajectories of the plurality of mobile terminals in the at least two floors according to the movement data of the mobile terminal; distribution of the position points on each of the moving trajectories; matching each of the moving trajectories with the leveling skeletons of the at least two floors to obtain the position points of the passage.

In a specific embodiment of the present application, the process of determining the configuration of the channel may include: performing pedestrian dead reckoning (PDR) according to the motion data of the mobile terminal, and obtaining PDR information when the mobile terminal moves in the channel; The PDR information includes one or more of changes in speed, number of steps, distance, and height; according to the PDR information, the configuration of the channel is determined.

S303: Generate a fingerprint map of the multi-storey building according to the leveling framework of the at least two floors, the location points of the passage, and the configuration of the passage.

Specifically, a 3D topology map can be generated according to the leveling framework of the at least two floors, the location points of the passage, and the configuration of the passage, and the construction process of the 3D topology map does not need to depend on the existing indoor plane. map. Then, the 3D topology map is mapped to the world coordinate system to obtain the fingerprint map of the multi-story building, which can be used by the server to provide online or offline indoor full-floor location services.

It can be seen that in the embodiment of the present application, through the interaction between multiple mobile terminals and the server, the server can make full use of the crowdsourced data collected by various mobile terminals to construct a fingerprint map of a multi-story building, replacing the collection of manual data and manual labor. It can also realize the construction of 3D topology map without map dependence, without deploying special data acquisition devices indoors, which greatly reduces the cost of labor and material resources. In addition, the embodiment of the present application automatically creates a multi-floor 3D topology map by using the landmark position of the channel, the uplink and downlink connection relationship of the paired landmark points, and the channel configuration, so as to solve the problem that the map construction in the prior art relies on manual editing, Improved the efficiency and accuracy of fingerprint map generation.

The indoor positioning method based on fingerprint map is described below, and the method includes two types: online positioning method and offline positioning method. The following descriptions are respectively described in the way of terminal-cloud interaction.

Referring to FIG. 6A, FIG. 6A shows an online positioning method, which includes but is not limited to the following steps:

S311. The mobile terminal collects real-time location fingerprint data of the mobile terminal in a multi-storey building; the location fingerprint data includes the number and/or signal strength of wireless access points (APs). For example, the AP is used to generate WiFi signals, and the location fingerprint data may include the number of APs and/or WiFi signal strengths scanned/detected by the mobile terminal in the current location.

Wherein, the current position of the mobile terminal may be in the floor, or may be in the passage of the floor.

S312. The mobile terminal sends a positioning request and the position fingerprint data to the server, where the positioning request is used to request the server to provide an online positioning service.

S313. The server matches the location fingerprint data with a pre-built fingerprint map, thereby obtaining a positioning result; wherein the fingerprint map may be the multiple pre-determined by the server according to crowdsourcing data of multiple mobile terminals. It is generated from the leveling skeleton of the floor building, the position points on the leveling skeleton and the configuration of the passage; for the specific generation process of the fingerprint map, please refer to the description of the relevant embodiments in the context of this application , which will not be repeated here.

S314. The server sends a positioning result to the mobile terminal. Accordingly, the mobile terminal receives the positioning result, and the positioning result is used to indicate the real-time position of the mobile terminal in the multi-storey building, and the real-time position corresponds to the mobile terminal. The location of the terminal in a floor of the multi-story building or in a passageway between two floors.

The positioning result may support subsequent services such as user indoor navigation, location search, AR/VR/MR or advertisement push, which is not limited in this application.

In a practical application of the present application, the method of the present application may be presented in the form of indoor positioning service software, deployed in the operating system layer of the mobile terminal, and used by the application layer LBS through an API interface. When the user needs positioning, the operating system layer of the intelligent mobile terminal system initiates a positioning request to the server in the cloud, the server returns the positioning result to the system layer of the terminal device, and finally uploads it to the application layer through the API interface, so that the user can obtain the positioning result.

It can be seen that through the method of online positioning, the cloud server can provide the mobile terminal positioning service interface. According to the received positioning request and position fingerprint data of the mobile terminal, the mobile terminal can be positioned by using the latest fingerprint map through the positioning algorithm. The positioning result is returned to the mobile terminal, which improves the convenience of indoor positioning of the mobile terminal and saves data storage overhead.

Referring to FIG. 6B, FIG. 6B shows an offline positioning method, which includes but is not limited to the following steps:

S321. The mobile terminal downloads and saves a fingerprint map from the server in advance, and the fingerprint map may be a leveling skeleton of a multi-story building determined by the server in advance according to crowdsourcing data of multiple mobile terminals, on the leveling skeleton The location point of the fingerprint map and the configuration of the channel are generated; for the specific generation process of the fingerprint map, reference may be made to the description of the relevant embodiments in the context of the present application, which will not be repeated here.

S322. The mobile terminal collects real-time location fingerprint data of the mobile terminal in the multi-storey building; the location fingerprint data includes the number and/or signal strength of wireless access points (APs). For example, the AP is used to generate WiFi signals, and the location fingerprint data may include the number of APs and/or WiFi signal strengths scanned/detected by the mobile terminal in the current location.

S323. The mobile terminal matches the location fingerprint data with the stored fingerprint map to obtain a positioning result, for example, obtains the real-time location of the first mobile terminal in the multi-storey building, where the real-time location is in the A location in a floor in a multi-story building or in a passageway between two floors.

Specifically, the positioning result may support subsequent services such as user indoor navigation, location search, AR/VR/MR or advertisement push, which is not limited in this application.

It can be seen that the offline positioning method can use the fingerprint map cached in advance for offline positioning when there is no network signal or the network signal is poor. In addition, it is also beneficial to combine the data of sensors, GPS, etc. for fusion positioning. , to further improve the accuracy of indoor positioning.

In order to better understand the positioning method of the present application, an application scenario is used as an example for description below. It should be noted that this application scenario is only used for an exemplary explanation of the solution of the present application, rather than a limitation of the solution of the present application.

Referring to FIG. 7 , a mobile phone 10 is used as an example for the mobile terminal. The mobile phone 10 is located in a multi-floor building (eg, a shopping mall with multiple floors), for example, a user is walking in the shopping mall with the mobile phone 10 The WiFi positioning function has been activated. The application scenario can be an online positioning scenario or an offline positioning scenario.

As shown in FIG. 7 , a certain map application is installed in the mobile phone 10 , and ( 1 ) in FIG. 7 shows the desktop 101 on the display panel of the mobile phone 10 . When the electronic device detects that the user clicks the icon 102 of the map application on the desktop 101, a user interface 104 of the application as shown in (2) in FIG. 7 is displayed on the display panel of the electronic device 10, where the The user interface 104 includes an electronic map interface and a plurality of controls, such as an electronic map control, a satellite map control and an indoor map control 103 as shown in the figure.

When the mobile phone 10 detects that the user clicks on the indoor map control 103, the user interface 105 of the application as shown in (3) in FIG. 7 is displayed on the display panel, and the user interface 105 shows that the user is currently on the first floor of the shopping mall real-time location in . When the user continues to walk in the mall, the real-time location of the user displayed in the user interface 105 will also be updated accordingly.

When the user moves across floors, the user interface 105 will also seamlessly display the process of moving across floors. As shown in (4) in FIG. 7 , when the user takes the escalator (of course, other passages, such as elevators, stairs, etc.) can be used to go to the second floor of the shopping mall, the user interface 106 displays the current location as being on the first floor. Go to the escalator on the 2nd floor, and the current real-time altitude can also be displayed.

After the user arrives on the second floor, as shown in (5) in FIG. 7 , the user's current real-time location on the second floor of the shopping mall continues to be displayed through the user interface 107 .

It can be seen that, by implementing the embodiment of the present application, it is possible to realize accurate positioning of a single point in each floor, such as the positioning in the first floor, the second floor, and the passage, and in the process of moving across floors, it can also display the process of crossing floors in real time ( For example, the 3D effect of continuous positioning from the 1st floor to the 2nd floor can realize the accurate and seamless 3D transition effect between floors. It can be completed without switching when positioning between different floors, and the accurate height can be displayed in real time. Therefore, the embodiments of the present application can improve the efficiency and accuracy of indoor positioning, and improve user experience.

In order to better understand the fingerprint map construction solution of the present application, the specific functions of the fingerprint map construction module 201 in the server 20 are described below. Referring to FIG. 8, FIG. 8 is an exemplary flowchart of a fingerprint map construction process of a multi-story building.

As shown in FIG. 8 , after obtaining the crowdsourcing data of the multiple mobile terminals, the server may, on the one hand, generate the leveling skeleton of at least two floors of the multi-storey building based on the crowdsourcing data. For example, the server obtains the movement trajectories of the plurality of mobile terminals in the at least two floors according to the movement data in the crowdsourced data, and then aggregates the movement trajectories of the plurality of mobile terminals to obtain the The leveling skeleton of at least two floors; wherein, each position point of the leveling skeleton has corresponding position fingerprint data.

On the other hand, the crowdsourced data of the mobile terminal can be processed through the channel self-learning module, the three-dimensional topology building module and the map mapping module in the fingerprint map building module respectively, and then the fingerprint map of the multi-storey building can be output. The processing of the channel self-learning module may further include processing stages of single-track channel detection, skeleton candidate channel detection, and skeleton accurate channel positioning. In the processing of the 3D topology building block, it may further include several processing stages of multi-layer skeleton sorting, absolute floor recognition and multi-layer skeleton alignment. In the processing process of the map mapping module, it may further include several processing stages of base-level entrance and exit detection and absolute coordinate mapping. The following will be described in detail through specific examples.

First, a specific embodiment about the channel self-learning module is described. In the embodiment of the present application, the channel self-learning module can be used to determine the position of the channel between the leveling frameworks of at least two floors in a multi-story building on the leveling framework according to crowdsourcing data, and also to determine the configuration of the channel .

Referring to FIG. 9, FIG. 9 shows a schematic flowchart of a possible implementation logic of the channel self-learning module. The process includes the following stages:

S401: Implement a single track channel detection stage based on cross-layer information fusion.

The crowdsourced data includes motion data, which is data collected by at least one sensor including at least one of an accelerometer, a gyroscope, a magnetometer, and an inertial measurement unit. Through this stage, the movement trajectories of the plurality of mobile terminals in at least two floors can be obtained according to the motion data, and the distribution of the location points (landmark points) of the passage on each of the movement trajectories can be identified.

In one implementation, a position where a user behavior change occurs on the movement track may be identified according to the motion data of the mobile terminal, where the user behavior change refers to the user behavior on the floor and the passageway. Then, according to the position where the user behavior change occurs, the position point distribution of the channel on the moving track is determined.

In yet another implementation, the crowdsourced data includes both motion data and altitude data collected by a barometer; then, according to the motion data of the mobile terminal, it is possible to identify the user behavior change on the movement track. position; the user behavior change refers to the transition between the user behavior on the floor and the user behavior on the passage; according to the height data of the mobile terminal, identify the occurrence of cross-floor events on the moving track Location. Then, fusion processing is performed according to the position where the user behavior change occurs and the position where the cross-layer event occurs, so as to determine the position point distribution of the channel on the moving track. For example, filtering and preprocessing the user behavior recognition results to obtain continuous and accurate cross-layer switching states. Kalman filtering is performed on the height change data retrieved from the height data of the barometer, and the cross-layer switching state is detected by the height change inflection point. According to the accuracy of the output behavior recognition state and the accuracy of the barometer cross-layer detection, the weighted fusion is carried out, the time series cross-layer state is fused, and the cross-layer state switching point is further accurately determined by setting the threshold, and corresponding to the PDR trajectory according to the timestamp information. , and output the single-track landmark detection results. The process can also be described in detail as follows:

(I) User behavior identification and processing:

The server can output various states of pedestrians walking with mobile phones through the behavior recognition program, including the states of going up and down stairs, elevators, escalators, etc. Therefore, cross-layer state detection can be performed with the help of behavior recognition output states. There is a discrete situation in the original behavior recognition output state, so it is first necessary to perform sequential filtering on the state to obtain a continuous floor switching state, and then detect the start and end points of the continuous state.

The processing flow is shown in Figure 10. In (1) of Figure 10, the sensor behavior model can be trained through machine learning, and the user behavior can be predicted at each moment. For example, the user's movement process includes walking on the floor. Behavior, also includes the behavior of going up in a corridor (such as an elevator). In (2) in Figure 10, the result input in (1) can be filtered, a segment of data window is taken during the filtering process, the state value with the most frequent occurrence in the right neighbor window is assigned to the current moment, and the current detection result is The confidence is set as the frequency ratio of the state in the current window, and the output result is shown in (3) in Figure 10. The multi-segment elevator upward behavior is integrated into a continuous elevator upward behavior, and the multi-segment user walks The behavior is integrated into a continuous walking behavior. The places where the user behavior changes (floor behavior and elevator behavior) are the start and end points of the passage (elevator).

(II) Cross-layer state detection based on barometer data:

The barometer can reflect the altitude change during the walking process, and the altitude of the mobile terminal can be inverted through the air pressure value output by the barometer during the walking process.

Among them, height represents the height value, and pressure represents the air pressure value.

The original height value may fluctuate greatly, and filtering processing is required first. In the embodiment of the present application, median filtering and Kalman filtering can be used to smooth the height curve, and the switching point of height change can be detected on the smooth height curve. The median filter is used to remove invalid data, and the Kalman filter is used to smooth the height curve.

The specific processing flow is shown in FIG. 10 . In (4) of FIG. 10 , the barometric pressure value change curve is obtained through the barometer. Since the curve has more invalid data and noise, filtering processing is required. After the filtering is completed, the height change curve can be obtained, as shown in (5) in Figure 10. In (6) in Figure 10, inflection point detection can be performed to detect the inflection point of height change, and the inflection point corresponds to the start and end time of floor switching, that is, it corresponds to landmark points such as stair entrance, elevator entrance, and escalator entrance. By detecting the peaks and valleys on the height change curve, it can be determined whether a cross-layer event occurs. The peak point of the height change curve greater than 0 represents the moment when the local height rises the fastest, corresponding to the upstairs event, and the valley point less than 0 corresponds to the moment when the local height drops the fastest, corresponding to the downstairs event. The occurrence of cross-layer can be detected by detecting the peak and valley points on the change curve. Then, the flat area of height change at both ends of the peak and valley points on the height curve is detected by scribing the window, and the cross-layer switching point can be located by locating the height change to the gentle inflection point. At the same time, the confidence of the detection state in the detection section is given according to the change rate of the change curve. The greater the absolute value of the change rate, the greater the confidence. The above implementation logic can be represented by the following formula ②:

Where h ₀ is the height change rate threshold, which is set as the mean value of the absolute value of the height change rate at the start and end points of detection (h _s + _{he )/2, and h Δ} _is the absolute value of the height change rate. γ is the sensitivity coefficient, which is set to 0.5.

(III) Fusion processing based on user behavior recognition and cross-layer event detection:

The state confidence obtained by behavior recognition and the state confidence obtained by barometer detection can be weighted and fused to obtain the final state confidence, and a higher weight can be given to the barometer detection result. The weight of the result is set to 0.7, the weight of the result of user behavior recognition is set to 0.3, and the threshold is set according to the final state confidence to determine the starting and ending points of the continuous state. The position of the starting and ending point is the position of the landmark point of the passage (eg, the elevator) on the moving track (eg, the PDR track).

In addition, pedestrian dead reckoning (PDR) can be performed according to motion data and/or altitude data to obtain PDR information when the mobile terminal moves in the channel, the PDR information including one or more of speed, number of steps, distance, and altitude change indivual. Based on the PDR information, the configuration of the channel is determined. For example, after the pedestrian dead reckoning (PDR) is performed to obtain the PDR trajectory, according to the detected start and end time stamps, corresponding to the PDR trajectory time stamps, the position of the start and end points in the trajectory is determined to realize single-track landmark point detection. At the same time, according to the behavior between the starting and ending points detected by the PDR, the status, speed/step number, distance, height change, etc. are identified, so as to distinguish the specific configuration of the passage, such as stairs, elevators or escalators, etc., and assign the passage corresponding up and down row direction property.

S402 : the detection stage of the candidate landmark points of the leveling skeleton.

After identifying the distribution of the location points (landmark points) of the channel on each moving track through the aforementioned S401, the server may divide the moving track into a flat segment track and a cross-segment segment based on the location point distribution on the moving track. Track; match the leveling segment track with the leveling skeleton through the K Nearest Neighbors (KNN) algorithm to obtain multiple candidate position points of the channel on the leveling skeleton.

Specifically, as shown in FIG. 11 , the detection results of the landmark points of a single movement track of the mobile terminal can be used to divide each cross-layer movement track (eg, PDR track) into two levels of leveling segment tracks (eg, the leveling track in the figure). Segment 1 and leveling segment 2) and a cross-level segment, the leveling segment trajectory is first calculated with the effective AP coincidence degree of each leveling skeleton. In the layer skeleton, according to the KNN registration algorithm, the track of the leveling segment is registered to the leveling skeleton, and the ground on the leveling skeleton can be obtained after coordinate transformation by using the marks of the landmark points carried by the track of the leveling segment. The location of the punctuation. Cross-segment trajectories are used for subsequent fingerprint clustering to obtain more precise landmark locations.

It should be noted that the traditional matching method generally adopts the most similar matching algorithm, that is, the point with the largest WiFi similarity with the point to be matched is selected as the matching point, where the WiFi similarity is defined as: assuming that the i-th step point in the reference trajectory scans The fingerprints of m APs are defined as:

For the jth step point in the track to be spelled to scan n APs, it is defined as:

but

and

Similarity of two fingerprint points

defined as:

where U is the union of two fingerprint APs, and when

Therefore, when the most similar matching algorithm is used, it will be affected by the burr trajectory on the skeleton, outliers, AP disorder on the skeleton, etc., resulting in the situation that the most similar points in WiFi are not matching points.

However, in this embodiment of the present application, the leveling segment is mainly used to find a matching leveling skeleton by using the number of AP intersections, selecting the leveling skeleton with the largest number of AP intersections as the matching leveling skeleton, and then passing KNN on the matching leveling skeleton. The registration algorithm finds matching points one by one. Using the KNN matching algorithm can reduce the interference of the above errors to a certain extent.

The following uses Fig. 12 as an example to explain a process of using the KNN matching algorithm to determine candidate landmark points. As shown in Fig. 12, the segmented leveling segment trajectory includes leveling segment 1 and leveling segment 1. number, and determine the leveling skeleton matching the leveling segment from the set of leveling skeletons of the multi-story building. For example, determine that leveling section 1 matches leveling frame 1, and determine that leveling section 2 matches leveling frame 2. Then, KNN registration is performed on the flattened segments with their matching flattened skeletons respectively. As shown in Figure 12, after KNN registration, the candidate landmark point 1 on the leveling skeleton 1 and the candidate landmark point 2 on the leveling skeleton 2 are obtained. The candidate landmark point 1 and the candidate landmark point 2 are the ground points at both ends of the same channel. punctuation.

The implementation process of the KNN matching algorithm is further explained below by taking FIG. 13 as an example. For any leveling segment and leveling skeleton that needs to be matched, K-nearest neighbor search is performed first, and the K points with the highest fingerprint similarity are selected, as shown in (1) in Figure 13. Then, the spatial coordinates of these K points can be used to perform meanshift clustering to eliminate the interference of outliers, and obtain points with better spatial distribution consistency, and then use similarity weighting to obtain the cluster center as the final matching point position, as shown in (2) in Figure 13. In this way, the results obtained by matching the fingerprint similarity can meet the neighbor consistency of the spatial distribution. By repeating the above process, the matching points of each point on the leveling section can be obtained, as shown in (3) in Fig. 13 . After matching, the coordinates of the matching points are obtained, and the RANSAC algorithm can be used to calculate the similarity transformation relationship between the coordinates (as shown in formula ④) to convert the leveling segment with registration to the leveling skeleton, that is, convert the coordinates of the landmark point to the leveling skeleton. The coordinate system of the skeleton is shown in (4) in Figure 13.

After KNN registration, pairs of landmark points (that is, landmark points at both ends of the same channel) on the same moving track can be recorded, so as to obtain the positions of the landmark points at both ends of the same channel on the leveling skeleton, and realize the candidate on the leveling skeleton. Detection of landmark points.

S403: The stage of precise positioning of landmark points on the leveling skeleton.

After the moving trajectory is divided into a leveling segment trajectory and a cross-segment trajectory through S402, and KNN registration is performed according to the leveling segment to obtain the positions of multiple candidate landmark points on each leveling skeleton, the multiple The cross-segment trajectories of the moving trajectory are clustered based on the similarity of the fingerprints, and the cross-segment clustering center is obtained. , and the obtained cluster center is used as the location point of the channel.

The following takes FIG. 14 as an example to further explain the implementation process of the WiFi similarity clustering algorithm. After obtaining multiple candidate landmark points (as shown in (1) in FIG. 14 ) through the registration of the horizontal segment trajectories, according to the position fingerprint data corresponding to the cross-segment trajectories of the multiple moving trajectories, the fingerprint The similarity is clustered, as shown in (2) in Figure 14, and before clustering, it is first distinguished according to the configuration of the channel (such as stairs, escalators, elevators), and then according to the configuration of the channel (such as stairs, escalators, elevators) Elevator) to perform cross-layer WiFi similarity clustering respectively. The category of the initial fingerprint point is obtained. This category distinction can perform initial classification on the candidate landmark points in each leveling skeleton. The initial classification result is shown in (3) in Figure 14. After that, on each leveling skeleton, meanshift clustering is performed using the spatial coordinates of the candidate landmark points in the initial classification to remove outliers in the initial classification, as shown in (4) in Figure 14. Accurate cluster centers are obtained through the above-mentioned clustering as the precise positions of landmark points, thereby realizing the optimization of channel positions. The optimized landmark positions are shown in (5) in FIG. 14 .

Among them, the WiFi similarity of the cross-layer trajectories in the above process can be defined by the intersection of trajectory APs, as shown in the following formula ⑤:

S ₁ ={bssid|rssi>-95}

The similarity represents the WiFi similarity, and S ₁ and S ₂ represent the set of APs whose Rssi value is greater than -95 in the APs of the two cross-layer trajectories.

It can be seen that, by implementing the embodiments of the present application, the server can use the motion data collected by sensors such as accelerometers, gyroscopes, and magnetometers to fuse with the altitude data collected by the barometer to detect landmark points of a channel on a single track. . Then, using the leveling skeleton generated by the aggregation of the leveling segment trajectories of each mobile terminal, the single-track cross-layer KNN is registered to the leveling skeleton, and then the candidate landmark points on the leveling skeleton are located. Finally, the fingerprint similarity of the cross-segment trajectories is used to perform the initial classification of the candidate landmark points, and the spatial distribution of the candidate landmark points is used to perform meanshift clustering to eliminate the interference of outliers, so as to accurately locate the landmark points on the leveling skeleton. In this way, it is helpful to improve the positioning accuracy of multiple floors in the subsequent establishment of fingerprint maps, especially the positioning accuracy in passages (such as elevators, stairs, escalators, etc.), realize the function of fine positioning between floors, and enhance seamless indoor and outdoor multi-floor The user experience of positioning, and supports providing positioning height in indoor positioning to improve the user experience.

Specific examples of three-dimensional topology building blocks are described below. In this embodiment of the present application, the 3D topology building module can be used to generate a 3D topology map according to the leveling skeleton of the multi-story building, landmark points of the passage, and the configuration of the passage.

Referring to FIG. 15, FIG. 15 shows a schematic flowchart of a possible implementation logic of the three-dimensional topology building module. The process includes the following stages:

S501 : a stage of sorting the obtained leveling skeletons.

The server may sort the leveling frameworks according to the connection relationship of the channels between the leveling frameworks, so that the order of the leveling frameworks conforms to the order of floors in the multi-story building. In addition, the floor height corresponding to each leveling skeleton can also be obtained according to the height clustering value of the floor.

Specifically, a pair of landmark points from the same cross-layer movement trajectory (that is, the landmark points at both ends of the same channel) will be registered to the leveling skeleton through the trajectory. In this way, the connectivity relationship between many landmark points can be used for clustering, and the upstream and downstream connectivity relationships between the two-level skeletons can be obtained.

As shown in Figure 16, after the landmark points on the trajectory are obtained according to the single movement trajectory, the movement trajectory is divided according to the landmark points, and each end of the divided two leveling layers will have a landmark point (that is, in the figure start and end points), and obtain candidate landmark points on the leveling skeleton by registering to the leveling skeleton. In this way, according to numerous leveling candidate landmark points, numerous connectivity relationships between two leveling skeletons can be marked.

Then, a directed graph can be constructed by the connection relationship between the two-level skeletons, and the directed graph can be detected by ring detection, and the directed acyclic graph can be segmented. The upper and lower order of each segmented sub-graph is performed to realize the upper and lower order of each leveling skeleton. In addition, the height difference represented by a pair of landmark points between the two leveling skeletons can also be used for clustering to obtain the height difference between the two leveling skeletons, and then the floor height of each leveling skeleton relative to the ground can be obtained.

In order to better understand the above sorting process, the following is a detailed description.

(I) Clustering of connection relationships and establishment of directed graphs:

Referring to Fig. 17 and Fig. 18 comprehensively, the connection relationship between the leveling skeleton T ₁ and T ₂ is denoted as R(T ₁ , T ₂ ), assuming that the landmark points L ¹ ∈ T ₁ , L ² ∈ T ₂ , the landmark point L ¹ to L ² are stairs going up, then d(L ₁ , L ₂ )=1; landmark points L ¹ to L ² are stairs going down, then d(L ₁ , L ₂ )=-1; and it can be obtained from the actual meaning d(L ₁ , L ₂ )=−d(L ₂ , L ₁ ). Then R(T ₁ , T ₂ )={d(L ₁ , L ₂ )|L ¹ ∈ T ₁ , L ² ∈ T ₂ ,} and L ¹ and L ² have a connection relationship. Use the relationship set R(T ₁ , T ₂ ) to perform clustering, and select the category with the largest number of clusters to represent the relationship, that is, the number of 1 and -1 in the statistical set, and the relationship with the largest number is selected as The up-down relationship d(T ₁ , T ₂ ) between the leveling skeletons T ₁ and T ₂ . In this way, we can obtain the uplink and downlink connectivity between the two-level skeletons, as shown in (1) in Figure 17. With this, a directed graph is constructed, and the directed graph is shown as (1) in FIG. 18 .

(II) Directed acyclic graph segmentation:

As shown in (2) in Figure 18, after completing the construction of the directed graph of the leveling skeleton, it needs to be segmented according to the uplink and downlink transfer relationship between the leveling skeletons to segment the directed acyclic graph. The ring graph is sorted up and down. When defining the up-down relationship of the leveling skeleton, the up-down direction of the vast majority of connection pairs can be taken as the up-down direction of the leveling skeleton among all the connection landmarks. Specifically, the direction is first converted into a consistent direction (both up or down), the directed acyclic graph is segmented using a depth-first search algorithm, starting from a certain node, the set is divided into a searched set and an unsearched set Set, when there is a connection relationship between the current search node and the node other than the direct parent node in the combination of the searched node, disconnect the connection relationship, until there is no connection relationship to continue the search, then save the currently divided subgraph, and continue in the remaining Search in the nodes of , and finally obtain a directed acyclic graph including each leveling skeleton.

(III) Sorting up and down:

After the directed acyclic graph segmentation is completed, the upstream and downstream sorting relationships can be sorted out according to the upstream and downstream connection relationships between the leveling skeletons. Specifically, for each directed acyclic graph, first find the node with only out-degree, which is the first layer, and then put it into the sorted set, and then disconnect all nodes in the sorted set and the unsorted set. Connect the relationship, and then use the same method to find the first layer of the remaining set, put the sorted set as the highest layer, and process the method in turn until all the leveling skeletons are sorted, then the upper and lower leveling skeletons of all floors can be completed. Row sorting, a sorting result is shown in (2) in Figure 17.

(IV) High clustering of flat skeletons:

This process is optional. After realizing the clustering of the connection relationship of the landmark points, the connection relationship between the skeletons of the two levels can be obtained, and the ground pressure value of each pair of connection relationship can be calculated according to the air pressure value of the landmark points with the connection relationship between the two floors. The height difference between the punctuation points, the absolute value of the height difference between the landmark points of the connectivity relationship between the floors is clustered, and the height difference value of the cluster center is selected as the height difference between the two floors. Combined with the upper and lower ordering relationship of floors, the height of each floor relative to the lowest floor can be obtained.

It should be noted that because the barometer is easily affected by the external environment and mobile phone equipment, the height value cannot be used for accurate floor sorting and height calculation; the height of a single calculated floor relative to the height is easily deviated due to the inaccuracy of the air pressure value. , and it is also difficult to reliably and efficiently sort and calculate heights. However, the embodiment of the present application utilizes the connection relationship between landmark points to realize the ordering of the leveling skeleton, which can effectively overcome the uncertainty caused by the inaccuracy of the barometer, and realize a more robust floor ordering.

In addition, the embodiment of the present application calculates the relative height of each leveling frame according to the floor sorting result, and performs high clustering according to a large amount of crowdsourced data, which can effectively eliminate unstable factors such as environment and collection status, and achieve a more reliable floor height. estimate.

S502: The absolute floor identification stage corresponding to each leveling skeleton.

The server may determine the mapping relationship between the sorted leveling skeletons and the floors in the multi-story building. After the upper and lower relationship of the leveling skeleton is sorted, it is only necessary to identify the leveling skeleton corresponding to the reference floor. For example, the reference floor may be the first floor, and then the leveling skeletons of other floors can be determined. Specifically, the server determines the number of entrances and exits on each leveling skeleton according to the crowdsourcing data, maps the leveling skeleton with the largest number of entrances and exits to the first floor of the multi-story building, and according to the sorting result of each leveling skeleton, assigns the remaining leveling skeletons. Floor skeletons are mapped to other floors in the multi-story building, enabling identification of absolute floors.

Generally speaking, most buildings will set the most entrances and exits on the first floor for the convenience of passage, transportation, fire protection, etc., and use the detected number of valid entrances and exits to select the floor with the most entrances and exits as the first floor, and sort according to the floor. relationship to get the floor corresponding to the real physical building. Therefore, the absolute floor recognition first uses the entrance and exit identification results to perform meanshift clustering on the entrances and exits to obtain effective entrance and exit clusters, and identify the reference floor - the first floor according to the number of entrance and exit clusters, and finally use the floor sorting relationship to obtain the absolute floor. The specific process is described as follows:

(I) Clustering of entrances and exits on the leveling skeleton:

Under the action of the entrance and exit detection module, the distribution of entrances and exits on the leveling skeleton can be obtained. However, due to the existence of detection errors, the distribution of entrances and exits is relatively discrete, and it only represents the entrances and exits of a single passage, which requires spatial clustering to identify and real buildings. The distribution of the corresponding entrances and exits. The entrance and exit identification adopts the meanshift clustering method. It is not necessary to set the number of clusters. It only needs to limit the maximum distance of the entrance and exit to obtain the clustering of the entrance and exit. Then, by setting the cluster size threshold, the category of a certain number of clusters can be satisfied. It is determined as a valid entry and exit. At the same time, the discrete detection points can be further eliminated by clustering, and the accurate cluster center can be obtained as the position of the entrance and exit.

(II) Identify the reference floor:

The identification of the reference floor (such as the first floor) uses the number of clusters of entrances and exits as the primary determining factor, and the one with the largest number of clusters is determined as the first floor. When the number of clusters is the same, it is further judged according to the orientation of the entrance and exit. Because the actual first floor generally has entrances and exits in all directions, and the entrances and exits will not be concentrated in a certain area, while the second and third floors or underground generally have cost or planning considerations, and the entrances and exits will be concentrated on a certain side, so they can be clustered according to the entrances and exits. The degree of dispersion of the center is further judged, and the position variance of the entrance and exit cluster centers is used to judge, and the one with the largest variance is judged as the first floor.

(III) Identify the absolute floor corresponding to each leveling skeleton:

After obtaining the floor sorting result and the reference floor, we can correspond the other floors to the real floor in turn according to the first floor identified in the reference floor, and according to the floor ascending and descending sorting relationship. In this way, when it is applied to the real positioning scene, it can accurately perceive the exact floor where it is located.

As shown in FIG. 19 , in an example, by performing the above process, a floor order is “leveling skeleton 3-leveling skeleton 2-leveling skeleton 1”, and the entrance and exit of leveling skeleton 3 are determined through the clustering of entrances and exits The number is 3, and the number of entrances and exits of other leveling skeletons is 0, then it can be determined that leveling skeleton 3 corresponds to the reference floor (1st floor), thus, based on the sorting order, it can be determined that leveling skeleton 2 corresponds to the 2nd floor, and leveling skeleton 1 corresponds to 3rd floor.

It should be noted that, in other embodiments of the present application, the identification of the absolute floor may also be implemented in other ways. For example, as shown in Figure 20, if the relative height values of each leveling frame have been calculated in S501, then the floor with the lowest height can be used as the reference floor, and the height of the reference floor is defined as 0m, so that other levels can be known. The absolute height of the layer skeleton. Then divide the height of each leveling frame by the floor spacing value to estimate the absolute floor corresponding to the leveling frame. Among them, the floor spacing value can be preset according to the actual building design, for example, the floor spacing value can be set to 3.5m. In the example shown in Figure 20, the height of the leveling frame 2 relative to the ground is about 3.5m, and the height of the leveling frame 3 relative to the ground is about 7.2m, then it can be determined that the leveling frame 2 corresponds to the 2nd floor, and the leveling frame 1 corresponds to the 3rd floor. building.

S503: Perform an alignment stage on the sorted leveling skeleton.

Since each leveling skeleton belongs to a local two-dimensional rectangular coordinate system, there is no corresponding connection relationship between each leveling skeleton, and it is necessary to restore the connection relationship of each leveling skeleton through cross-level information. The alignment of the leveling skeleton is to reconstruct the connection relationship between the leveling skeletons through the self-learned landmark positions and spatial configurations of the channels in the leveling skeleton, and unify the multi-level leveling skeletons into the same coordinate system. , in order to achieve seamless indoor positioning on multiple floors. That is, in the embodiment of the present application, the server may align the sorted leveling skeletons in the three-dimensional coordinate space according to the landmark points of the channel and the configuration of the channel to obtain the three-dimensional topological map.

Specifically, by identifying the characteristics of different channels, different leveling skeletons can be aligned from coarse to fine according to the characteristics of the channels. Floor alignment is a necessary step to construct a multi-floor 3D topology map. It mainly uses the configuration and connection relationship of elevators, stairs, escalators and other passages between floors to convert the coordinate systems of different leveling skeletons into a unified coordinate space. The directly connected channels correspond to the same coordinate point on the X and Y planes, so that the leveling skeletons reconstructed from different floors can be aligned into the same three-dimensional coordinate space, and the same coordinate system mapping of the multi-floor fingerprint map can be realized. Give the floor height, so as to realize the construction of 3D topology map. The 3D topology map established in this way can realize seamless positioning when moving across floors and fine positioning on passages between floors, which greatly improves the user experience. The alignment process is described in detail below:

(I) Channel configuration identification:

Since different channels have different spatial configurations, and the leveling skeletons need to be aligned in different ways for channels with different configurations, the spatial configuration of each channel needs to be determined before performing the alignment of the leveling skeletons. For example, the configuration of the channel can be a direct connection type and an oblique connection type; the configuration of the channel includes a direct connection type or an oblique connection type; the direct connection type means that the position points at both ends of the channel are in the vertical direction Collinear, the oblique connection type means that the position points at both ends of the channel are not collinear in the vertical direction. Therefore, the landmark points at both ends of the direct connection type channel need to be aligned in the vertical direction, and the landmark points at both ends of the oblique connection type channel are not aligned in the vertical direction and are offset in the horizontal direction. A direct-connection type of passage is, for example, an elevator, a ladder, or some kind of staircase that folds in the middle, and an inclined type of passage is, for example, a sloped staircase, an escalator, a ramp, and the like.

Generally speaking, elevators and ladders have a strict straight-through alignment relationship, while for stairs and escalators, there may be various spatial configurations. Referring to Figure 21, Figure 21 shows five common configurations of stairs and escalators in the real world. in:

The stairs in (1) in Figure 21 are of the direct connection type, and the stair landmarks detected by cross-floor detection will be vertically aligned on different floors; in (2) in Figure 21, the even-numbered interval floors (1 -The stairs between floors 2, 1-4) are inclined connection types with horizontal offsets, while the stairs between floors with odd intervals (floors 1-3, 2-4, 1-5) can be regarded as Direct connection type; the escalator in (3) in Figure 21 is an inclined connection type, and the distance of this inclined connection can be estimated by sensors; in Figures (4) and (5) in Figure 21, 1-2 The escalators on the 2nd and 3rd floors are of the oblique connection type, but it should be noted that there may be implicit direct connections between the escalator landmarks on the other floors. The so-called implicit direct connection refers to the landmarks without connection relationship. There may also be vertical alignment relationships between. For example, in (5) in Figure 21, although there is no channel connection between the landmark point on the 1st floor and the landmark point on the 3rd floor, there may also be an implicit direct connection between these two landmark points. Also assists in the alignment of the flat skeleton in this application.

Referring to Figure 22, the landmark points that can be used for cross-floor alignment can be divided into the following four categories: (1) For straight elevators and stairs, they can be directly used for straight alignment; (2) For diagonally connected stairs , the hypotenuse distance can be estimated by PDR, the horizontal offset can be calculated by using the floor height, and the horizontal offset can be optimized and corrected; (3) For the implicit directly connected landmarks, including the escalator landmarks, the ICP algorithm can be used to align the implicit The relationship can be used to achieve further alignment correction; in addition, for the escalator with simple inclined span, it is difficult to use it for floor alignment, but the speed of the escalator can be estimated by the detected starting and ending heights and time intervals for fine inter-floor positioning .

Since the elevator can determine that there is a straight-through relationship, when there is an elevator between the two leveling skeletons, the coordinate system transformation matrix can be calculated by using the elevator landmark points. Due to the error uncertainty of the PDR step size estimation, there may be uneven scales between different leveling skeletons, so the similarity transformation can be used to calculate the transformation matrix. The calculation of similarity transformation requires at least two pairs of corresponding points, so at least two pairs of elevator landmark points that are far enough apart are required when calculating the preliminary transformation matrix. Generally, landmark points from two different elevator clusters are selected for calculation. When multiple elevator clusters or landmark points in multiple clusters exist, RANSAC is used for robust parameter estimation.

After calculating the initial transformation matrix, transform the landmark points of the stairs and the landmark points of the escalator, and calculate the back-projection error between the landmark points of the connection relationship, as shown in formula ⑥:

e=∑||L _i -St*L _j || ⑥

Among them, St is the calculated similarity transformation matrix, and L _i and L _j are the landmark points of a pair of connection relationships.

For stairs, the stairs with the back-projection error less than a certain threshold can be selected as the stairs of the direct connection type, otherwise the stairs of the inclined connection type can be selected.

For escalators, the connection relationship is basically the oblique connection type, and in the subsequent weighted ICP correction process, there is basically no need to know the corresponding relationship between the escalator landmarks, so there is no need to distinguish the type for the escalator. In the subsequent ICP The correction process can be applied directly.

It should be noted that, in the process of implementing the alignment solution provided by the present application, since there are multiple floors in a multi-story building, and the number and configuration of passages between floors may be different, the more directly connected floors are, the higher the alignment is. The reliability of the results is better, and a certain floor may have a connection relationship with multiple other floors. Therefore, after the identification of the passage configuration is completed, the sequential alignment strategy with multiple floors needs to be considered.

A possible alignment strategy is: firstly initialize the first floor into the aligned floor set, then select the unaligned connected floor in the unaligned floor set that has the most direct connection with each floor in the aligned floor set for processing, and initialize it first Directly connect and align, then optimize the horizontal distance estimation, use weighted ICP fine alignment, and add the current processing floor to the set of lower aligned floors, and so on until all floors are aligned. The related description is further expanded below.

(II) The leveling skeleton is aligned based on the direct connection type channel:

When there is a direct connection type channel between the two leveling frameworks, the two leveling frameworks can be aligned by aligning the position points at both ends of the channel in the vertical direction.

Specifically, the identified elevators and stairs of the direct connection type can be used to calculate the similarity transformation between the points with the same name by using the RANSAC algorithm, and then the skeleton coordinates to be aligned are transformed into the aligned leveling skeleton coordinate system using the similarity transformation. Among them, since the calculation of the similarity transformation matrix requires at least two connected point pairs, and the distance between the two pairs of points is far enough, if the similarity transformation matrix is too close, the error of the fitted similarity transformation matrix will be large.

Therefore, when RANSAC random sampling is performed in the process of direct connection alignment, the selection is made according to different directly connected stairs and elevator clusters. First, two landmark clusters with a direct relationship are selected, and the distance between the two cluster centers should be greater than 10. m, and then randomly select a point in each of the two pairs of landmark clusters to form two pairs of sampling points, use the two pairs of points to calculate the similarity transformation matrix, and then use all the directly connected points to calculate the back projection error, so that the projection error is less than a certain value Threshold points as interior points. In this way, after several times of sampling, the set of results with the largest number of interior points is selected, and the similarity transformation matrix is calculated by using its interior points as the final fitting result output.

(III) The horizontal distance optimization of the leveling skeleton is based on the oblique connection type channel:

When there is an oblique connection type channel between the two leveling frameworks, the alignment of the two leveling frameworks is further optimized according to the horizontal offset between the position points at both ends of the channel. That is to say, after the straight-connected alignment is performed, the alignment result can be further corrected according to the identified inclined-connected stairs.

Referring to Figure 23, taking a sloped staircase as an example, first, the PDR distance between the connected landmark points is used as the hypotenuse distance, and the estimated height of the connected floor is used as a right-angle side, so that the distance between the diagonally connected landmark points can be estimated. By compensating the calculated horizontal offset to the obliquely connected landmark points, and incorporating the compensated obliquely connected landmark points into the directly-aligned stairs and elevators, further direct-connected alignment calculations are performed to obtain further corrected alignments. result. Since the landmark point L2 is directly projected onto the ^floor F1 according to the calculated similarity transformation, there will be projection points

This projection point will also have an offset between the landmark point L ¹ on F1 to which it is connected. If the calculated similarity transformation matrix has good accuracy, the two offset values will be equal, so it is possible to pass these two The difference between the offset values can determine the quality of the similar transformation matrix fit.

Furthermore, by compensating for this distance offset, it can be added to the straight-through landmark point fitting program to further optimize the fitting accuracy of the similarity transformation matrix. Assume that the stair direction is based on the projected point

Consistent with the direction between L ¹ , through the principle of collinearity, the projected point coordinates of the compensation coordinates on F1 can be obtained

The coordinates of can be calculated according to formula ⑦:

By inverse projection, the

The coordinates are projected to the F2 plane to obtain the compensation coordinate point ΔL ² :

(IV) For implicitly connected types, the alignment results are optimized by a weighted iterative Closest Point (ICP) algorithm:

This step is optional. After using the direct-connected elevators, stairs and inclined-connected stairs, there are still a large number of implicit direct-connected landmarks that are not used, that is, the stairs, elevators and escalators that have no connection relationship but are actually directly connected. To use this part of the implicit direct connection relationship, the ICP algorithm can be used for iterative optimization. After several iterations of optimization, the global direct connection alignment result will be output.

According to the previous direct alignment and inclined stairs correction, a better alignment effect can be obtained as the initial alignment value of the ICP algorithm. However, in order to avoid a large number of unaligned landmark points, resulting in a sharp deterioration of the final result, in the iterative calculation process Weighting is performed, in which straight stairs and elevators are given higher weights. On the basis of the overall data, the directly connected elevator and stair landmark point pairs are repeatedly injected to increase the proportion of directly connected landmark points; each time the nearest neighbor point is selected, it is preferred to form a team of points with a connected relationship in the K adjacent points. If not, select the nearest neighbor point; use all directly connected elevator stairs for distance error evaluation, and select only the minimum error of other cross-floor points that account for one-third of the number of directly connected stairs and elevators for calculation.

By performing the above process, a better alignment effect of each leveling skeleton can be obtained, and finally a 3D topology map can be output.

It can be seen that, in the embodiment of the present application, firstly, according to the configuration of the detection channel, it is determined whether the landmark point of the channel belongs to a direct connection type (such as a direct connection staircase and a direct connection elevator), an oblique connection type (such as an inclined connection staircase) or an implicit direct connection type. connection type (such as escalator); then, the direct connection type channel can be preferentially used for the initial direct connection alignment of the leveling frame; then, the horizontal offset is calculated using the PDR track distance and floor height difference in the inclined connection type channel, so that Compensate the horizontal offset of the leveling skeleton to further optimize the calculation of the alignment relationship; finally, use all landmark points to perform ICP fine alignment, assign higher weights to directly connected elevators and stairs, and use the escalator with implicit direct relationship to further accurate Align floors to get a 3D topology map. By adopting the coarse-to-fine floor coordinate system alignment strategy in stages, the fine alignment of floors can be realized without resorting to indoor floor plans or manual editing, which is conducive to the realization of automatic multi-floor fingerprint map construction and improves the efficiency and accuracy of map construction. , and the construction process of the 3D topological map does not need to rely on the existing indoor flat map.

The following continues to describe specific embodiments of the map mapping module. In this embodiment of the present application, the 3D topology building module can be used to map the 3D topology map to the world coordinate system to obtain a fingerprint map of a multi-story building. Specifically, the functional realization of the 3D topology building module can include two aspects:

On the one hand, as shown in Figure 24, the mobile terminal that can be included in the crowdsourcing data is based on the information obtained by the satellite navigation module. Occasional GPS information in the environment (opportunistic GPS for short). The server obtains the GPS of the entrance or the indoor opportunity GPS by identifying the position of the entrance and exit of the reference floor. Then, the 3D topology map can be mapped to the world coordinate system, such as the WGS84 coordinate system, by using the entrance and exit GPS or indoor opportunistic GPS, so as to realize the absolute coordinate mapping of the whole floor, that is, to obtain the fingerprint map of the multi-story building in the world coordinate system.

On the other hand, as shown in FIG. 25 , the embodiment of the present application also supports using pre-obtained indoor map, vector map, grid map, or control point information to map with the 3D topology map, so that the 3D topology map can be mapped with the 3D topology map. The actual indoor floor map of the multi-story building is associated, thereby obtaining a fingerprint map corresponding to the actual map of the multi-story building.

It can be seen that by implementing the embodiments of the present application, a fingerprint map of a multi-storey building under the WGS84 coordinate system can be obtained without (or only based on part of) an indoor floor plan, thereby realizing the map coordinate mapping of all floors. In addition, the fine connection relationship between floors can be reconstructed, so as to achieve refined and seamless switching of intra-floor positioning and intra-channel positioning. The present application is compatible with two ways of realizing the absolute coordinate mapping of the 3D topology map without a map source and with a map source, so as to realize the seamless indoor and outdoor positioning effect of multi-floor and full-scene.

Combining the various embodiments described above, it can be seen that implementing the solution of the present application can at least bring about the following effects:

On the one hand, the present application replaces manual data collection and manual annotation by collecting crowdsourced data on various users' mobile terminals, and realizes the construction of 3D topological maps without map dependence, without deploying special data collection devices indoors. , greatly reducing the cost of labor and material resources.

On the other hand, the embodiment of the present application identifies landmark points on the trajectory of the mobile terminal by fusing data from multiple sensors, is compatible with scenarios with and without a barometer, and uses the KNN registration algorithm to self-learn candidates on the leveling skeleton. Landmark points, combined with the fingerprint space characteristics of the cross-layer trajectories on the channel and the Euclidean spatial distribution of the landmark points, can accurately locate the landmark points on the flat-layer skeleton, thereby avoiding the influence of the diversity of crowdsourcing data and the quality of crowdsourcing data collection. .

On the other hand, the application makes full use of the characteristics of the landmark points of the passage, automatically sorts and aligns multiple leveling skeletons, and identifies absolute floors. The whole process does not require manual editing and labeling, and has no map dependence. The entire process can be automated and highly efficient. The fingerprint map of the multi-storey building can be outputted in place, which improves the efficiency and accuracy of the fingerprint map generation. Practice has proved that the multi-floor fingerprint map constructed based on this patent has an average floor alignment accuracy of 2-3.5m. Therefore, the fingerprint map based on this application can achieve high-precision positioning in actual indoor positioning. Since the generated multi-floor fingerprint map is represented in the same coordinate space, the positioning between different floors can be completed without switching. Combined with the map mapping module, it can realize the seamless positioning experience of indoor and outdoor multi-floor full-scene.

Finally, the existing methods mainly rely on indoor maps to achieve absolute coordinate mapping, and many multi-story buildings currently do not provide indoor maps, resulting in the failure of map construction. However, the embodiment of the present application proposes a method based on entry-exit GPS and opportunistic DPS mapping to realize the mapping of fingerprint map to absolute coordinates. This method does not need to rely on indoor maps, has strong universality, and is effective when there is an indoor map or no indoor map.

It should be understood that in the various method embodiments of the present application, the size of the sequence numbers of the above-mentioned processes does not mean the sequence of execution, and the execution sequence of each process should be determined by its functions and internal logic, and should not be implemented in the present application. The implementation of the examples constitutes no limitation.

Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Skilled artisans may implement the described functionality using different methods for each particular application, but such implementations should not be considered beyond the scope of this application.

In the above-mentioned embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described systems, devices and units may refer to the corresponding processes in the foregoing method embodiments, which will not be repeated here. In the several embodiments provided in this application, it should be understood that the disclosed system, apparatus and method may be implemented in other manners. For example, the apparatus embodiments described above are only illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components may be combined or Can be integrated into another system, or some features can be ignored, or not implemented. On the other hand, the shown or discussed mutual coupling or direct coupling or communication connection may be through some interfaces, indirect coupling or communication connection of devices or units, and may be in electrical, mechanical or other forms.

In addition, each functional module in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically alone, or two or more units may be integrated into one unit. The functions, if implemented in the form of software functional units and sold or used as independent products, may be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be embodied in the form of a software product in essence, or the part that contributes to the prior art or the part of the technical solution. The computer software product is stored in a storage medium, including Several instructions are used to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk and other media that can store program codes .

The above embodiments are only used to illustrate the technical solutions of the present application, but not to limit them. Although the present application has been described in detail with reference to the above embodiments, those of ordinary skill in the art should also understand that: any modification, deformation, or modification of some of the technical features based on the technical solutions described in the above embodiments The equivalent replacements of all should belong to the spirit and scope of the technical solutions of the various embodiments of the present application.

Claims

A method for constructing a fingerprint map of a multi-story building, wherein the method is applied to a server, comprising:

receiving crowdsourced data from a plurality of mobile terminals, the crowdsourced data including motion data and location fingerprint data collected during cross-floor movement of the mobile terminal in a multi-story building;

According to the crowdsourced data, the leveling frames of at least two floors in the multi-storey building, the location points on the leveling frames of the passages connected between the leveling frames, and the location points of the passages are determined. structure;

A fingerprint map of the multi-story building is generated according to the leveling framework of the at least two floors, the location points of the passage, and the configuration of the passage.
The method according to claim 1, wherein the configuration of the channel includes a direct connection type or an oblique connection type; the direct connection type means that the position points at both ends of the channel are collinear in the vertical direction, so The above-mentioned oblique connection type means that the position points at both ends of the channel are not collinear in the vertical direction.
The method according to claim 1 or 2, wherein the generating a three-dimensional topological map according to the leveling skeleton, the location points of the channel and the configuration of the channel includes:

According to the connection relationship of the passages between the respective leveling frames, the respective leveling frames are sorted, so that the sequence of the respective leveling frames conforms to the upper and lower relationship of the floors in the multi-storey building; The position point and the configuration of the channel are aligned in the three-dimensional coordinate space of the sorted flat skeletons to obtain the fingerprint map.
The method according to claim 3, characterized in that, before aligning the sorted leveling skeletons in the three-dimensional coordinate space according to the position point of the channel and the configuration of the channel, further include:

A mapping relationship between the sorted leveling skeletons and the floors in the multi-story building is determined.
The method according to claim 4, wherein, according to the connection relationship of the channels between the respective leveling skeletons, sorting the respective leveling skeletons, comprising:

constructing a directed acyclic graph of each leveling framework according to the connection relationship of the channels between the leveling frameworks;

The respective flat skeletons are ordered according to the directed acyclic graph.
The method according to claim 4 or 5, wherein the determining the mapping relationship between the sorted leveling skeletons and the floors in the multi-story building comprises:

According to the crowdsourcing data, determine the number of entrances and exits of each floor corresponding to the leveling framework; the leveling framework with the largest number of entrances and exits corresponds to the first floor of the multi-story building;

According to the sorting result of each of the leveling frameworks, the floors corresponding to the remaining leveling frameworks are determined.
The method according to any one of claims 4-6, wherein, according to the position point of the channel and the configuration of the channel, each leveling skeleton after the sorting is arranged in a three-dimensional coordinate space Alignment, including:

When there is a direct connection type channel between the two leveling frameworks, aligning the two leveling frameworks by aligning the position points at both ends of the channel in the vertical direction;

When there is an oblique connection type channel between two leveling frameworks, the two leveling frameworks are aligned according to the horizontal offset between the position points at both ends of the channel.
The method according to any one of claims 1-7, wherein the process of determining the leveling skeleton of at least two floors in the multi-story building according to crowdsourcing data comprises:

obtaining movement trajectories of the plurality of mobile terminals in the at least two floors according to the motion data;

The movement trajectories of the multiple mobile terminals are aggregated to obtain the leveling skeletons of the at least two floors; wherein, the leveling skeletons are associated with the position fingerprint data.
The method according to any one of claims 1-8, wherein the process of determining, according to the crowdsourcing data, the location points of the channels connected between the skeletons of each leveling layer comprises:

obtaining movement trajectories of the plurality of mobile terminals in the at least two floors according to the motion data;

Identifying the position points belonging to the channel distributed on each of the moving trajectories;

Each of the moving trajectories is matched with the leveling skeletons of the at least two floors to obtain the position points of the passage on the leveling skeletons.
The method according to claim 9, wherein the matching each of the moving trajectories with the leveling skeletons of the at least two floors to obtain the position points of the passages comprises:

Based on the position point distribution on the moving track, the moving track is divided into a flat-segment track and a cross-segment track;

Obtain multiple candidate position points of the channel on the leveling frame according to the leveling segment trajectory and the leveling skeleton; obtain the position point of the channel according to the multiple candidate position points .
The method according to claim 9, wherein the identifying the position point distribution of the channel on the movement track comprises: for each mobile terminal in the plurality of mobile terminals, according to the movement The movement data of the terminal, identify the position where the user behavior change occurs on the moving track; the user behavior change refers to the transition between the user behavior on the floor and the user behavior on the passage; (instructions Interpretation is a dynamic process)

According to the location where the user behavior change occurs, the location point distribution of the location points at both ends of the channel on the movement track is determined.
The method of claim 9, wherein the crowdsourced data further comprises altitude data collected by a barometer;

The identifying the position distribution of the channel on the moving track includes:

For each mobile terminal in the plurality of mobile terminals, according to the motion data of the mobile terminal, identify the position where the user behavior change occurs on the moving track; the user behavior change refers to the movement on the floor. transitions between user behavior and user behavior on said channel;

According to the height data of the mobile terminal, identify the position where the cross-layer event occurs on the movement track;

According to the location where the user behavior change occurs and the location where the cross-layer event occurs, the location point distribution of the channel on the moving track is determined.
The method according to any one of claims 1 to 12, wherein the generating the multi-layer structure is performed according to the leveling framework of the at least two floors, the connection position of the passage and the configuration of the passage. The fingerprint map of the floor building, including:

generating a three-dimensional topological map according to the leveling skeleton, the location points of the channel and the configuration of the channel;

The three-dimensional topological map is mapped to the world coordinate system to obtain the fingerprint map of the multi-story building.
The method according to any one of claims 1-13, wherein the process of determining the configuration of the channel according to crowdsourcing data comprises:

Perform pedestrian dead reckoning (PDR) according to the motion data, and obtain PDR information when the mobile terminal moves in the channel; the PDR information includes one or more of speed, number of steps, distance, and altitude change;

Based on the PDR information, the configuration of the channel is determined.
The method according to any one of claims 1-14, wherein the motion data is data collected by at least one sensor, and the at least one sensor includes an accelerometer, a gyroscope, a magnetometer, and an inertial measurement unit. At least one of; the location fingerprint data includes the number and signal strength of wireless access points.
An indoor positioning method, wherein the method is applied to a first mobile terminal, comprising:

Download a fingerprint map from the server, the fingerprint map is the leveling skeleton of the multi-story building, the position points on the leveling skeleton and the passage determined by the server according to the crowdsourced data from a plurality of second mobile terminals generated from the configuration;

collecting real-time location fingerprint data of the first mobile terminal in the multi-storey building;

Matching the location fingerprint data with the fingerprint map to obtain a real-time location of the first mobile terminal in the multi-story building, where the real-time location is a location in a floor in the multi-story building Or the location in the passage between two floors.
An indoor positioning method, wherein the method is applied to a server, comprising:

receiving a location request and the location fingerprint data from the first mobile terminal;

The location fingerprint data is matched with the fingerprint map to obtain a positioning result; the fingerprint map is the flat skeleton of the multi-story building determined by the server according to the crowdsourcing data of a plurality of second mobile terminals, in the generated from the position points on the flat skeleton and the configuration of the channel;

The first mobile terminal sends the positioning result, and the positioning result is used to indicate the real-time position of the first mobile terminal in the multi-storey building, and the real-time position is the location in the multi-storey building. A position in a floor or in a passageway between two floors.
A device for fingerprint map construction for multi-story buildings, characterized in that the device comprises:

a receiving module, configured to receive crowdsourcing data from a plurality of mobile terminals, where the crowdsourcing data includes motion data and location fingerprint data collected by the mobile terminal during a process of moving across floors in a multi-story building;

A fingerprint map building module for determining, according to the crowdsourcing data, the leveling skeletons of at least two floors in the multi-storey building and the positions of the passages connected between the leveling skeletons on the leveling skeletons point, and the configuration of the passage; generating a fingerprint map of the multi-storey building according to the leveling framework of the at least two floors, the location points of the passage, and the configuration of the passage.
The device according to claim 18, wherein the configuration of the channel includes a direct connection type or an oblique connection type; the direct connection type means that the position points at both ends of the channel are collinear in the vertical direction, so The above-mentioned oblique connection type means that the position points at both ends of the channel are not collinear in the vertical direction.
The device according to claim 18 or 19, wherein the fingerprint map building module is specifically used for:

According to the connection relationship of the passages between the respective leveling frameworks, the respective leveling frameworks are sorted, so that the order of the respective leveling frameworks conforms to the upper and lower relationship of the floors in the multi-story building;

According to the position point of the channel and the configuration of the channel, the sorted leveling skeletons are aligned in the three-dimensional coordinate space to obtain the fingerprint map.
The device according to claim 20, wherein the fingerprint map building module is further configured to: at the location point according to the channel and the configuration of the channel, the sorted Before the floor skeletons are aligned in the three-dimensional coordinate space, the mapping relationship between the sorted leveling skeletons and the floors in the multi-story building is determined.
The device according to claim 21, wherein the fingerprint map building module is specifically used for:

constructing a directed acyclic graph of each leveling framework according to the connection relationship of the channels between the leveling frameworks;

The respective flat skeletons are ordered according to the directed acyclic graph.
The device according to claim 21 or 22, wherein the fingerprint map building module is specifically used for:

According to the crowdsourcing data, determine the number of entrances and exits of the floors corresponding to each of the leveling frameworks; the leveling framework with the largest number of entrances and exits corresponds to the first floor of the multi-story building;

According to the sorting result of each of the leveling frameworks, the floors corresponding to the remaining leveling frameworks are determined.
The device according to any one of claims 21-23, wherein the fingerprint map building module is specifically used for:

When there is a direct connection type channel between the two leveling frameworks, aligning the two leveling frameworks by aligning the position points at both ends of the channel in the vertical direction;

When there is an oblique connection type channel between two leveling frameworks, the two leveling frameworks are aligned according to the horizontal offset between the position points at both ends of the channel.
The device according to any one of claims 18-24, wherein the fingerprint map building module is specifically used for:

obtaining movement trajectories of the plurality of mobile terminals in the at least two floors according to the motion data;

The movement trajectories of the multiple mobile terminals are aggregated to obtain the leveling skeletons of the at least two floors; wherein, the leveling skeletons are associated with the position fingerprint data.
The device according to any one of claims 18-25, wherein the fingerprint map building module is specifically used for:

obtaining movement trajectories of the plurality of mobile terminals in the at least two floors according to the motion data;

Identifying the position points belonging to the channel distributed on each of the moving trajectories;

Each of the moving trajectories is matched with the leveling skeletons of the at least two floors to obtain the position points of the passage on the leveling skeletons.
The device according to claim 26, wherein the fingerprint map building module is specifically used for:

Based on the position point distribution on the moving track, the moving track is divided into a flat-segment track and a cross-segment track;

According to the leveling segment trajectory and the leveling skeleton, multiple candidate position points of the channel on the leveling skeleton are obtained; according to the multiple candidate position points, the position point of the channel is obtained .
The device according to claim 26, wherein the fingerprint map building module is specifically used for:

For each mobile terminal in the plurality of mobile terminals, according to the motion data of the mobile terminal, identify the position where the user behavior change occurs on the moving track; the user behavior change refers to the movement on the floor. The transition between user behavior and user behavior on the channel; (the specification explains that it is a dynamic process)

According to the location where the user behavior change occurs, the location point distribution of the location points at both ends of the channel on the movement track is determined.
The apparatus of claim 26, wherein the crowdsourced data further comprises altitude data collected by a barometer;

The fingerprint map building module is specifically used for:

For each mobile terminal in the plurality of mobile terminals, according to the motion data of the mobile terminal, identify the position where the user behavior change occurs on the moving track; the user behavior change refers to the movement on the floor. transitions between user behavior and user behavior on said channel;

According to the height data of the mobile terminal, identify the position where the cross-layer event occurs on the movement track;

According to the location where the user behavior change occurs and the location where the cross-layer event occurs, the location point distribution of the channel on the moving track is determined.
The device according to any one of claims 18-29, wherein the fingerprint map building module is specifically used for:

generating a three-dimensional topological map according to the leveling skeleton, the location points of the channel and the configuration of the channel;

The three-dimensional topological map is mapped to the world coordinate system to obtain the fingerprint map of the multi-story building.
The device according to any one of claims 18-30, wherein the fingerprint map building module is specifically used for:

Perform pedestrian dead reckoning (PDR) according to the motion data, and obtain PDR information when the mobile terminal moves in the channel; the PDR information includes one or more of speed, number of steps, distance, and altitude changes;

Based on the PDR information, the configuration of the channel is determined.
The device according to any one of claims 18-31, wherein the motion data is data collected by at least one sensor, and the at least one sensor includes an accelerometer, a gyroscope, a magnetometer, and an inertial measurement unit. At least one of; the location fingerprint data includes the number and signal strength of wireless access points.
An apparatus for indoor positioning, wherein the apparatus is applied to a first mobile terminal, comprising:

The receiving module is used for downloading a fingerprint map from the server, where the fingerprint map is the leveling skeleton of the multi-story building and the position on the leveling skeleton determined by the server according to the crowdsourcing data of a plurality of second mobile terminals generated by the configuration of the points and channels;

a data collection module, configured to collect real-time location fingerprint data of the first mobile terminal in the multi-storey building;

A local positioning module for matching the location fingerprint data with the fingerprint map to obtain the real-time location of the first mobile terminal in the multi-story building, where the real-time location is in the multi-story building A position in a floor in a , or a position in a passageway between two floors.
An apparatus for indoor positioning, characterized in that the apparatus is applied to a server, comprising:

a receiving module for receiving a positioning request and the position fingerprint data from the first mobile terminal;

The cloud positioning service module is used for matching the position fingerprint data with a fingerprint map to obtain a positioning result; the fingerprint map is the data of the multi-story building determined by the server according to the crowdsourcing data of a plurality of second mobile terminals. Generated by the leveling framework, the position points on the leveling framework and the configuration of the channel;

a sending module, configured to send the positioning result to the first mobile terminal, where the positioning result is used to indicate the real-time position of the first mobile terminal in the multi-storey building, where the real-time position is in the A position in a floor in a multi-story building or in a passageway between two floors.
A server, characterized in that the server comprises a memory, a communication device and a processor, wherein: the communication device is used for receiving data from the outside world or sending data to the outside world; the memory is used for storing programs; the processor is used for For executing the program stored in the memory, when the processor executes the program stored in the memory, it is specifically configured to execute the method according to any one of claims 1 to 15 .
A mobile terminal, characterized in that the server comprises a communication device, a signal detection device and a processor, wherein:

The communication device is used for the server to download a fingerprint map, and the fingerprint map is the leveling skeleton of the multi-story building and the position on the leveling skeleton determined by the server according to crowdsourcing data of a plurality of second mobile terminals. generated by the configuration of the points and channels;

The signal detection device is used to collect real-time location fingerprint data of the first mobile terminal in the multi-storey building;

The processor is configured to match the location fingerprint data with the fingerprint map to obtain a real-time location of the first mobile terminal in the multi-story building, where the real-time location is in the multi-story building The location in the floor of the , or the location in the passage between the two floors.
A server, characterized in that the server includes a communication device and a processor, wherein:

The communication device is configured to receive a positioning request and the position fingerprint data from the first mobile terminal;

The processor is configured to match the location fingerprint data with a fingerprint map to obtain a positioning result; the fingerprint map is the leveling of the multi-story building determined by the server according to crowdsourcing data from a plurality of second mobile terminals. The skeleton, the position points on the flat skeleton and the configuration of the channel are generated;

The communication device is further configured to send the positioning result to the first mobile terminal, where the positioning result is used to indicate the real-time position of the first mobile terminal in the multi-storey building, where the real-time position is at A location in a floor in the multi-story building or a location in a passageway between two floors.