CN108981686B

CN108981686B - P2P indoor navigation system and P2P indoor navigation method

Info

Publication number: CN108981686B
Application number: CN201810754717.3A
Authority: CN
Inventors: 姚英彪; 林维维; 潘雷; 金勇�; 冯维; 许晓荣; 刘兆霆
Original assignee: Hangzhou Dianzi University
Current assignee: Hangzhou Dianzi University
Priority date: 2018-07-11
Filing date: 2018-07-11
Publication date: 2021-03-16
Anticipated expiration: 2038-07-11
Also published as: CN108981686A

Abstract

The invention discloses a P2P indoor navigation system and a P2P indoor navigation method, comprising a leader reference path generation module, a path information processing module, a follower navigation path generation module and a server; the path information processing module is used for detecting turning point actions, walking path length, course angle and geomagnetic sequences of the leading person and the follower; the route leader reference path generation module is used for carrying out path diversification according to the turning point action, the walking path length, the course angle and the geomagnetic sequence of the route leader, generating the reference path information and uploading the diversified path and the reference path information to a server; and the follower navigation path generation module is used for matching the turning point action, the walking path length, the course angle and the geomagnetic sequence of the follower with the corresponding meta path and reference path information in the server in real time. The invention has no host/client, can complete the indoor navigation function only by using the inertial sensor in the smart phone, avoids the site deployment in advance, and can provide navigation service in most scenes.

Description

P2P indoor navigation system and P2P indoor navigation method

Technical Field

The invention relates to the technical field of indoor navigation and positioning, in particular to a P2P indoor navigation system and a P2P indoor navigation method.

Background

In recent years, with the popularization of smart phones, indoor positioning and navigation methods based on smart phones are becoming popular. Currently, there are two common indoor positioning and navigation methods:

(1) fingerprint technology based on wireless Received Signal Strength (RSS).

Although the method can obtain better positioning accuracy and navigation service at later nodes after the positioning infrastructure is deployed and the early-stage fingerprint acquisition is completed, the positioning and navigation system is limited in two aspects. First, this approach is highly dependent on accurate and reliable infrastructure pre-deployed by a particular vendor, and deploying accurate and reliable indoor positioning and navigation infrastructure requires significant human and material costs. Secondly, when the indoor building structure changes, updating the fingerprint database data brings a large amount of cost, and if the fingerprint database data is not updated in time, the quality of the positioning and navigation service is affected. Finally, detailed indoor maps always need to provide accurate architectural structure and semantic place information, but are difficult to obtain.

(2) Pedestrian Dead Reckoning (PDR) techniques that rely on inertial sensors.

Inertial sensor-based PDR is another way of indoor location and navigation, which determines the position at the current time based on the pedestrian's position, step size, and heading angle from the previous time. But because it is a relative positioning, positioning errors can accumulate. Therefore, in a short time, the PDR method can provide very high positioning accuracy; however, over a long period of time, as the number of walking steps increases, positioning errors accumulate, resulting in unreliable conventional PDR positioning over a long period of time.

In recent years, a class of P2P (Peer to Peer) based indoor positioning and navigation services has begun to be popular. This P2P mode may be applied in social and personal scenarios and conditionally implements a leader-follower navigation mode. The navigation system navigates specific route followers by allowing users to record and share specific route tracking information in a self-motivated manner. For example, a group of users arrive in tandem at an appointed place of appointment or the provider wishes to direct the customer to his own store.

Disclosure of Invention

The invention aims to solve the problem that navigation cannot be provided for pedestrians in a building with incomplete deployment of partial facilities, and provides a P2P indoor navigation system based on PDR and geomagnetism. The navigation system shares the key reference information on the reference path to the follow-up follower by tracking the moving track of the leader, provides timely walking prompt and reminds the potential path deviation for the follower, and ensures that the follower follows the moving track of the leader to reach the same destination.

The technical scheme of the invention is as follows:

a P2P indoor navigation system comprises a leader reference path generating module, a path information processing module, a follower navigation path generating module and a server;

the path information processing module comprises a turning detection unit, a path calculation unit and a geomagnetic processing unit, wherein the turning detection unit is used for detecting turning point actions of a leading person and a follower, the path calculation unit is used for calculating the length and the course angle of a walking path, and the geomagnetic processing unit is used for acquiring geomagnetic sequences of the leading person and the follower;

the pathfinder reference path generating module comprises a path metarization unit and a path information generating unit, wherein the path metarization unit metarizes paths according to the turning point action of the pathfinder, numbers the metarized paths, and obtains metarized path information according to the metarized path numbers, the turning point action of the pathfinder, the walking path length, the course angle and the geomagnetic sequence,

the path information generating unit generates reference path information according to the metaalized path information and uploads the reference path information to a server;

the follower navigation path generation module comprises a starting point locking unit, a meta path matching unit and a path deviation detection unit; the starting point locking unit is used for locking the starting point and the initial azimuth for the follower, and the meta-path matching unit is used for matching the turning point action, the walking path length, the course angle and the geomagnetic sequence of the follower with the corresponding meta-path information in the server; the path deviation detecting unit is used for detecting whether the follower deviates from the reference path.

Furthermore, the turning detection unit adopts a magnetometer or a gyroscope to detect the turning point action, the path calculation unit adopts a PDR algorithm to calculate the length and the course angle of a walking path, and the geomagnetic processing unit adopts a dynamic time scaling matching algorithm to calculate whether the geomagnetic sequences of the leading person and the follower are matched or not.

The invention also provides a P2P indoor navigation method, which comprises the following steps:

s1, the leader selects a starting point and an end point, and the leader starts from the starting point and starts to generate a reference path;

s2, detecting the turning point action of the leader in real time, calculating the path length and the heading angle of the leader, and acquiring the geomagnetic sequence of the leader;

s3, when the turning of the leader is detected, numbering the paths traveled by the leader to form a meta path; recording the average course angle, the path length, the turning point action, the geomagnetic sequence and the number on each meta path, and uploading the average course angle, the path length, the turning point action, the geomagnetic sequence and the number to a server;

s4, the leader confirms whether the terminal is reached, if yes, the generation of the reference path is finished, if no, the step S2 is returned;

s5, the follower selects the starting point and the end point and searches the starting point;

s6, after the follower determines the starting point, the walking direction of the follower is detected, and when the error between the walking course angle of the follower and the course angle of the leader in the server is smaller than a first threshold value, the step S7 is executed; if not, repeating the step;

s7, detecting whether the length of the walking path of the follower reaches the requirement for prompting the front steering in real time, if so, prompting the follower to make the same turning point action as the leader at the front intersection, and entering the step S8; if not, detecting whether the follower deviates from the reference path in real time, and returning to the step S6;

s8, detecting whether the follower turns in real time, if yes, entering the step S9;

and S9, detecting whether the geomagnetic sequence of the follower is matched with the geomagnetic sequence of the leader, if so, starting the matching of the next section of meta-path until the follower reaches the end point, and ending the navigation.

Further, the path length and the heading angle of the walking of the leader and the follower are calculated by adopting a PDR algorithm, and the geomagnetic sequences of the leader and the follower are collected by adopting a magnetometer.

Further, step S6 is specifically: estimating the course angle of the follower in real time by using a PDR algorithm, and entering the step S7 when the error between the real-time course angle of the follower and the average course angle on the meta-path corresponding to the reference path of the leader is less than a first threshold value; if not, the step is repeated.

Further, in step S7, detecting whether the length of the walking path of the follower has reached the requirement for prompting the front steering in real time specifically includes:

and calculating the error between the accumulated path length of the user walking and the accumulated path length of the leader walking on the meta path by using a PDR algorithm, wherein the error is reached if the error is less than a second threshold value.

Further, in step S7, detecting whether the follower deviates from the reference path in real time specifically includes:

and (4) estimating the course angle of the follower in real time by using a PDR algorithm, judging that the follower deviates from the reference path when the error between the real-time course angle of the follower and the average course angle on the meta-path corresponding to the reference path of the leader is greater than a first threshold value, continuously reminding the follower, and returning to the step S6.

Further, step S8 is specifically:

s8.1, detecting whether the user turns in real time, if so, entering the step S8.2, otherwise, calculating the error between the accumulated path length of the walking of the follower and the path length of the walking of the leader in the server on the diversified path by using a PDR algorithm, and if the error is greater than the second threshold value, reminding the user of missing a turning point and prompting to return to the previous intersection;

s8.2, detecting whether the turning point action of the follower is matched with the turning action of the leader at the turning point or not in real time by using a PDR algorithm, if so, entering a step S9, and if not, reminding the follower of a turning error.

Further, in step S9, detecting whether the geomagnetic sequence of the follower matches the geomagnetic sequence of the leader includes:

s9.1, respectively representing the geomagnetic intensity subsequence on a certain section of the meta-path of the follower and the geomagnetic intensity subsequence on the corresponding matched meta-path of the leader as follows:

wherein n and k respectively represent the total sequence length on the follower meta-path and the total sequence length on the leader corresponding meta-path;

s9.2 subsequence

And

the DTW distance between them is expressed as

S9.3 the difference in DTW distance between the previous and subsequent sample points in the follower geomagnetic sequence is expressed as

S9.4 setting a fixed threshold ε_MWhen the follower turns at a turning point close to the target turning point, the difference ε in DTW distance between the front and rear sampling points on the navigation path_DTWGreater than epsilon_MAnd the navigation system reminds the user of the occurrence of path deviation, the PDR algorithm stops calculating the path length until the course angles of the two are matched again, and the PDR continues to run.

Compared with the prior art, the invention has the beneficial effects that:

1) the system has no host/client. Each user may act as a host, providing routes to other users. 2) The navigation system can complete the indoor navigation function only by using the inertial sensor in the smart phone. No external infrastructure (e.g., WLAN access points, bluetooth access points, etc.) is required, nor is a fingerprint repository established. 3) Navigation information is rapidly provided for the next pedestrian according to the path of the previous pedestrian, the site is prevented from being deployed in advance, and a large amount of human resources and construction resources are saved. 4) Navigation services can be provided in most scenarios, particularly for emergency navigation (e.g., a moderator can provide conference participants with routes to a conference room in advance when a temporary meeting is made, a merchant can provide customers with routes from various entrances to a store, etc.).

Drawings

FIG. 1 is a navigation system technology framework.

Fig. 2 shows an example of indoor navigation in P2P.

Fig. 3 is a flow of generating a leader reference path.

Fig. 4 is a state transition diagram.

FIG. 5 is a flow chart for generating a follower navigation path.

Detailed Description

The present invention will be described in further detail below with reference to the accompanying drawings and examples.

Example one

As shown in fig. 1, the P2P indoor navigation system of the present invention includes a leader reference path generating module, a path information processing module, a follower navigation path generating module, and a server;

the route-leader reference path generation module comprises a path metaplasia unit and a path information generation unit, wherein the path metaplasia unit metaplasia is used for metaplasia of paths according to the turning point action of the route leader, numbering the metaplasia paths, and obtaining metaplasia path information according to the metaplasia path number, the turning point action of a follower, the length and the course angle of a walking path and a geomagnetic sequence,

In this embodiment, the turn detection unit detects the turn point motion by using a magnetometer or a gyroscope, the path calculation unit calculates the length and the heading angle of the walking path by using a PDR algorithm, and the geomagnetic processing unit calculates whether the geomagnetic sequences of the leading person and the follower are matched by using a dynamic time scaling (DTW) matching algorithm.

Example two

The embodiment provides a P2P indoor navigation method, which specifically includes the following steps:

s3, when the turning of the leader is detected, numbering the paths traveled by the leader to form a plurality of diversified paths; recording the average course angle, the path length, the turning point action, the geomagnetic sequence and the number on each meta path, and uploading the average course angle, the path length, the turning point action, the geomagnetic sequence and the number to a server;

The invention adopts PDR algorithm to calculate the path length and course angle of the walking of the leader and the follower, and adopts magnetometer to collect the geomagnetic sequences of the leader and the follower.

The invention utilizes the gyroscope to detect whether the user turns left and right or turns around in real time, and converts the complex path into a multi-section short straight line meta-path so as to reduce the accumulated error of the PDR algorithm.

In step S5, after the follower selects the start point and the end point, the follower searches for the start point by using the outdoor GPS.

Step S6 specifically includes: estimating the course angle of the follower in real time by using a PDR algorithm, and entering the step S7 when the error between the real-time course angle of the follower and the average course angle on the meta-path corresponding to the reference path of the leader is less than a first threshold value; if not, the step is repeated.

In step S7, detecting in real time whether the length of the walking path of the follower has reached the requirement for prompting the front steering, specifically: and calculating the error between the accumulated path length of the user walking and the accumulated path length of the leader walking on the meta path by using a PDR algorithm, wherein the error is reached if the error is less than a second threshold value.

In step S7, detecting whether the follower deviates from the reference path in real time specifically includes: and (4) estimating the course angle of the follower in real time by using a PDR algorithm, judging that the follower deviates from the reference path when the error between the real-time course angle of the follower and the average course angle on the meta-path corresponding to the reference path of the leader is greater than the first threshold value, continuously reminding the follower, and returning to the step (S6).

In step S8, detecting whether the follower turns in real time specifically includes:

s8.1, detecting whether the follower turns or not in real time, if so, entering the step S8.2, otherwise, calculating the error between the accumulated path length of the walking of the follower and the path length of the walking of the leader on the diversified path by using a PDR algorithm, and if the error is larger than the second threshold value, reminding the user of missing a turning point and prompting to return to the previous intersection;

In step S9, the DTW geomagnetic sequence matching algorithm is a dynamic geomagnetic time sequence matching algorithm, that is, inputting a dynamic geomagnetic time sequence of a follower, matching the dynamic geomagnetic time sequence with the geomagnetic sequence of the leader in real time, and determining whether the matching is successful according to the similarity of the geomagnetic sequences. The method comprises the following specific steps:

where n and k represent the total sequence length on the follower's meta-path and the total sequence length on the leader's corresponding meta-path, respectively, which are not necessarily equal.

S9.2 subsequence

And

the DTW distance between can be expressed as

Further, a DTW geomagnetic sequence matching algorithm is used for detecting whether the user turns at the intersection close to the turning point, and the specific steps are as follows:

the difference in DTW distance between the previous and next sample points in the S9.3 follower geomagnetic sequence can be expressed as

S9.4 setting a fixed threshold ε_MWhen the follower turns at a turning point close to the target turning point, the difference epsilon between the DTW distances of the front and rear sampling points on the navigation path is caused due to different geomagnetic intensities on different paths_DTWGreater than epsilon_MAt the moment, the navigation system reminds the user of the path deviation, meanwhile, the PDR algorithm stops calculating the path length until the course angles of the path deviation and the path deviation are matched again, namely, the follower returns to the nearest meta-path and faces the leader, and the PDR continues to run.

Referring to fig. 2-5, a detailed description will be given of an exemplary embodiment of the present invention by taking zhang three (a leader) and lie four (a follower) as examples. In a shopping building, a salesperson in a P brand shop opens a three-hand mobile phone to operate a navigation system, and arrives at the shop from a west entrance of the building along a route A- > B- > C- > D- > E, and the route is uploaded to a server through processing and is shared with all customers using the system. Customer lie four may use the navigation system to reach the store along the same route a- > B- > C- > D- > E.

As shown in fig. 3, the specific steps of generating the reference path of the third leader include:

s1, selecting a starting point (west entrance of a certain shopping building) and an end point (P brand shop) by Zhang III, and uploading the information to a server by the system. Zhang three starts from the starting point and starts to generate the reference path.

And S2, calculating the path length and the course angle of the Zhang III walking by using the PDR algorithm. At the same time, the magnetometer collects geomagnetic intensity values.

S3, when Zhang III turns are detected by a turning detection method, the system numbers a straight path passed by Zhang III, calculates an average course angle on the metaplastic path and uploads the course angle, the path length passed by Zhang III, the action of a turning point, a geomagnetic sequence and the metaplastic path number to a server, wherein a route A- > B in FIG. 2 is a metaplastic path 1; b- > C is a metaplasia path 2; c- > D is a meta path 3; d- > E is the metapath 4.

S4, confirming whether the terminal is reached or not by Zhang III, and if the terminal is not reached, repeating the step 2-3. And if the terminal point is reached, the reference path is generated by three ends.

The PDR algorithm in step S2 calculates the walking length and the estimated heading angle, and includes the following specific steps:

2.1 estimating the heading angle using an electronic compass.

2.2 finish the self-adaptation to count the step on the basis of the state machine, its concrete step is as follows:

2.2.1 calculate the magnitude of the linear acceleration vector using the three-axis linear accelerometer data collected by the accelerometer.

2.2.2 to remove high frequency components, a cut-off frequency f is used_LPA finite impulse response 13 th order gaussian filter at 3Hz filters the acceleration values.

2.2.3 use state machines to simulate the walking pattern of a person and calculate the number of steps based on the period of the state machine. The different states correspond to different phases of a step, which form the whole process of a step. C0 is the equilibrium state, C1 is the increasing state, C2 is the peak state, C3 is the decreasing state, C4 is the trough state, C5 is the completion state. To drive the state machine, four thresholds Str1, Str2, Str3, Str4 are set, which correspond to the upper limit of the equilibrium stage, the lower limit of the peak, the upper limit of the valley and the lower limit of the equilibrium stage, respectively. State machine operation as shown in the process of the figure fig. 4, the state machine executes the instructions (a) - (f):

(a) if the state machine is at state C0, it is determined whether the current linear acceleration is less than Str 1. If so, the state machine will remain at C0; otherwise, the state machine will proceed to C1.

(b) If the state machine is at state C1, it is determined whether the current linear acceleration is less than Str 2. If so, the state machine will remain in C1; otherwise, the state machine will proceed to C2.

(c) If the state machine is at state C2, it is determined whether the current linear acceleration is less than Str 3. If so, the state machine will remain in C2; otherwise, the state machine will proceed to C3.

(d) If the state machine is at state C3, it is determined whether the current linear acceleration is less than Str 3. If so, the state machine will remain at C3; otherwise, the state machine will proceed to C4.

(e) If the state machine is at state C4, it is determined whether the current linear acceleration is less than Str 4. If so, the state machine will remain in C4; otherwise, the state machine will proceed to C5.

(f) In the C5 state, it is determined whether to count one step according to the satellite conditions. If the state machine is in state C5, it needs to be determined whether the periodic information satisfies the subject condition. If so, add 1 to the counter and jump to state C0; otherwise, the value of the counter remains unchanged and the state machine jumps to state C0. The accessory conditions are as follows: 1) the time of one cycle should be within the time period of the pedestrian walking through one normal step; 2) the time interval between the peaks and the valleys should satisfy a normal step interval; 3) the difference between the peak and the valley should satisfy a normal range. Therefore, the influence of vibration or other interference factors can be reduced, and the precision is improved.

2.3 adopting two step length estimation methods, and fusing the two methods by a weighting method. The first step estimate is the fourth root of the acceleration variance:

wherein the content of the first and second substances,

and

the acceleration peak and the wave trough of the k step are respectively. Beta is a fixed proportion, and gamma is a compensation value. According to practice, we set β -1.479 and γ -1.259 respectively. Another step size estimate is to consider both pedestrian height and step frequency:

l₂＝h(af_step+b)+c

wherein a, b and c are related to the height and the step frequency of the pedestrian, f_stepIs the pedestrian stride frequency. The step size is estimated using a weight-based fusion method.

Wherein, ω is₁+ω₂1. The values of the two weights can be adjusted in real time. For example, when the pedestrian is of unknown height, ω₁Is set to zero; when the step frequency cannot be obtained, ω₂Set to zero at the beginning of the movement; any outliers can be removed from the fusion process by setting the corresponding weights to zero.

The path is diversified by using a turning detection method, the turning detection method is based on output data of an accelerometer and a gyroscope, and the method is implemented by the following steps:

wherein the content of the first and second substances,

is an intelligent mobile phone in actual sceneCollected gyroscope data; (a)_x,a_y,a_z) The average data of the acceleration sensor collected by the intelligent mobile phone in the actual scene;

is the vertical rotation speed. n and m are window sizes. In practical application, for the turning detection method, when the rotation accumulated angle theta is larger than theta_thTime (theta)_thIs a preset threshold), i.e., a turn can be considered to occur.

The meta path information uploaded to the server is recorded as:

wherein, L is the track length calculated according to PDR; h is the average course angle in the path; e is the turning of the turning point, such as left turn 'L' and right turn 'R'. M is the geomagnetic intensity sequence on the path; n is a radical of^RA set of paths after the metaplazation for the reference path.

The metatrack of follower LieIV is recorded as

As shown in fig. 5, the specific steps of generating the navigation path of follower lie four are as follows:

in the follower calibration stage, the following steps are sequentially carried out:

and S5, after selecting the same starting point and end point as Zhang III, the fourth plum walks to the west entrance of the shopping building by the outdoor GPS.

S6, after determining the starting point, lie four further determines the walking direction (consistent with zhang san) according to the heading angle estimated in the PDR algorithm. And repeating the steps until the heading angle error is smaller than the threshold value, and then carrying out the next stage.

Heading angle offset error is defined as

Wherein

For the course angle offset error on the ith metapath,

for the average heading angle on the ith metapath on the navigation path,

the real-time course angle on the ith metapath on the reference path is determined. The threshold value is set to a fixed value epsilon_HThe size of the swing can be determined according to the maximum swing in the actual walking process of the user.

In the follower straight-going matching stage, the following steps are sequentially carried out:

s7, the PDR algorithm starts to detect gait and estimate step size. Meanwhile, the system prompts the plum four to go straight.

And (3) detecting whether the path length of the four-leg walking has reached the requirement of prompting forward steering in real time by using a PDR (product data Rate) matching algorithm.

Detecting whether the Liquan deviates from the path in real time by using a PDR (product data Rate) matching algorithm before reaching the prompt point; after the prompting point is reached, the system prompts the Liquan to make the same steering action as Zhang three phases at the front intersection.

The error of path length matching is defined as:

wherein

For the path length offset error on the ith quantized path,

for the path length on the ith metapath on the navigation path,

the path length on the ith metapath on the reference path is used. The threshold value is set to a fixed value epsilon_LThe size of the material can be determined according to actual experimental results.

And (4) detecting whether the Liquan deviates from the direction on the straight path in real time by using a PDR matching algorithm, wherein the heading angle error and the threshold are defined in the same step S5. When the heading angle error of the lie IV is larger than the threshold value, the system judges that the turn deviation of the lie IV occurs and continuously reminds the lie IV until the heading angle error is smaller than the threshold value.

In the turning matching stage of the follower on the leader, the following steps are sequentially carried out:

and S8.1, detecting whether the plum four turns in real time, and executing the next step after detecting turning.

S8.2, detecting whether the Li IV is in wrong steering at the turning point by using a PDR matching algorithm. And executing the next step after detecting that the steering is correct.

And S9, detecting whether the geomagnetic sequence of Liqu matches with the geomagnetic sequence of Zhang three by using a dynamic time scaling (DTW) geomagnetic sequence matching algorithm.

Any one of the steps is not satisfied, which indicates that the path of the follower lie IV in the turning matching stage is not successfully matched with the path of the leader Zhang III, and the navigation can timely prompt the lie IV to have path deviation. With reference to fig. 2, when step S8.1 is not satisfied, the route of lie quad walking is a- > B- > C- > D- > F; when the step S8.2 is not satisfied, the route of the Liquan walking is A- > B- > C- > D- > G; when the step S9 is not satisfied, the route of Liquan walking is A- > B- > C- > G- > I.

Said step S8.1 fails to detect lie four turns, i.e. the wrong path a- > B- > C- > D- > F, in the matching phase. The specific method is that the PDR algorithm is used for calculating the accumulated path length of the four-Li walking and is matched with the path length of the three-Lee walking on the diversified path in the database. The path length error and the threshold are defined in the same step S7. The system reminds the intersection (point D) in front of the Li IV to turn right at the point G, the Li IV continuously advances to the point F after missing the point D, at the moment, the path length difference value is larger than the threshold value, the system judges that the Li IV misses the turning area, reminds the Li IV to miss the turning point and prompts the Li IV to return to the previous intersection (point D).

In the step S8.2, a PDR matching algorithm is used to detect whether the turn point steering error occurs to lie IV, namely, the error path A->B->C->D->G. The specific method is to compare the normalized paths of the ith segment of lie

Whether to be on the third ith section of the meta-path with the leader

And (5) the consistency is achieved. Three-in-one metaroad route C for leaders>The action saved on D is a right turn

If detecting follower Lifour in meta-path C->D-turn action as left turn

The system alerts Liquan to a steering error.

In the step S9, a DTW geomagnetic sequence matching algorithm is used to detect whether lie iv turns at an intersection close to a turning point, that is, an error path a- > B- > C- > G- > I. The method comprises the following specific steps:

s9.1, respectively representing the geomagnetic intensity subsequence on a certain section of the meta-path of the follower Li and the geomagnetic intensity subsequence on the corresponding matched meta-path of the leader Zhang III as follows:

S9.2 a two-dimensional matrix g (1k,1n) is established for calculating the DTW distance. Based on the idea of dynamic programming, the DTW algorithm utilizes a recursion equation to fill the matrix, and the expression mode is

To solve for

And

the euclidean distance between them. Magnetic strength sequence M₁And M₂The distance between the two DTW is g (k, n), i.e. DTW (M)₁,M₂)＝g(k,n)。

Subsequence(s)

And

the DTW distance between can be expressed as

S9.4 setting a fixed threshold ε_MWhen the follower lie four turns at a turning point (G point) close to the target turning point (D point), due to the different paths (D->E and G->I)The difference of the DTW distance between the front and rear sampling points on the navigation path is caused by different geomagnetic intensities_DTWGreater than epsilon_MAt the moment, the navigation system reminds the lee-IV route to deviate and prompts the user to return to the last intersection (G point), meanwhile, the PDR algorithm suspends calculating the route length until the course angles of the two are matched, namely the lee-IV route returns to the intersection G point and the heading is east, and the PDR continues to run.

While the preferred embodiments and principles of this invention have been described in detail, it will be apparent to those skilled in the art that variations may be made in the embodiments based on the teachings of the invention and such variations are considered to be within the scope of the invention.

Claims

1. A P2P indoor navigation system, comprising: the system comprises a leader reference path generation module, a path information processing module, a follower navigation path generation module and a server;

the follower navigation path generation module comprises a starting point locking unit, a meta path matching unit and a path deviation detection unit; the starting point locking unit is used for locking the starting point and the initial azimuth for the follower, and the meta-path matching unit is used for matching the turning point action, the walking path length, the course angle and the geomagnetic sequence of the follower with the corresponding meta-path information in the server; the path deviation detection unit is used for detecting whether the follower deviates from the reference path;

the turning detection unit detects the turning point action by adopting a magnetometer or a gyroscope, the path calculation unit calculates the length and the course angle of a walking path by adopting a PDR algorithm, and the geomagnetic processing unit calculates whether the geomagnetic sequences of the leading person and the follower are matched by adopting a dynamic time scaling matching algorithm.

2. A P2P indoor navigation method is characterized in that: the method comprises the following steps:

s9, detecting whether the geomagnetic sequence of the follower is matched with the geomagnetic sequence of the leader, if so, starting the matching of the next section of meta-path until the follower reaches the end point, and ending the navigation;

and calculating the length and the heading angle of the walking path of the leader and the follower by adopting a PDR algorithm, and collecting the geomagnetic sequences of the leader and the follower by adopting a magnetometer.

3. The P2P indoor navigation method, according to claim 2, wherein:

4. The P2P indoor navigation method, according to claim 2, wherein: in step S7, detecting in real time whether the length of the walking path of the follower has reached the requirement for prompting the front steering, specifically:

5. The P2P indoor navigation method, according to claim 4, wherein: in step S7, detecting whether the follower deviates from the reference path in real time specifically includes:

6. The P2P indoor navigation method, according to claim 5, wherein: step S8 specifically includes:

7. The P2P indoor navigation method, according to claim 3, wherein: in step S9, it is detected whether the geomagnetic sequence of the follower matches the geomagnetic sequence of the leader, specifically:

s9.2 subsequence

And

the DTW distance between them is expressed as