CN107566981B - Indoor high-precision positioning method, device and system based on optimal path - Google Patents
Indoor high-precision positioning method, device and system based on optimal path Download PDFInfo
- Publication number
- CN107566981B CN107566981B CN201710762494.0A CN201710762494A CN107566981B CN 107566981 B CN107566981 B CN 107566981B CN 201710762494 A CN201710762494 A CN 201710762494A CN 107566981 B CN107566981 B CN 107566981B
- Authority
- CN
- China
- Prior art keywords
- node
- positioning
- optimal
- branch tree
- optimal path
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Landscapes
- Position Fixing By Use Of Radio Waves (AREA)
Abstract
The invention relates to an indoor high-precision positioning method, device and system based on an optimal path, wherein the method comprises the following steps: adopting a multi-branch tree to segment a region to be positioned, and establishing a dictionary matrix corresponding to all father nodes in the multi-branch tree, wherein the dictionary matrix comprises actual measurement distances between each child node corresponding to each father node in the multi-branch tree and each reference base station; a target point to be positioned transmits a positioning signal for receiving by the M reference base stations; receiving measurement distances sent by M reference base stations to form observation vectors, determining an optimal node of each layer of the multi-branch tree in an area to be positioned according to the dictionary matrix and the observation vectors, determining an optimal path, and taking a position corresponding to a leaf node at the bottom of the optimal path as a positioning result of the target point; and repeatedly positioning for many times, and taking out the position corresponding to the bottom leaf node with the maximum probability as the final positioning result of the target point.
Description
Technical Field
The invention belongs to the field of wireless communication, and discloses an indoor high-precision positioning method, device and system based on an optimal path.
Background
In recent years, communication technology and mobile internet are rapidly developed, meanwhile, the demand of people for indoor positioning information is increased day by day, and indoor accurate positioning plays a vital role in numerous fields such as navigation, production management and realization of the internet of things. For example: medical treatment, precision instrument manufacturing, accurate navigation and positioning of indoor robots, positioning and monitoring of dangerous goods, positioning and monitoring of personnel and equipment in dangerous areas and the like all contain urgent demands on positioning. However, in a relatively complex Non-Line of Sight (NLOS) environment, due to the existence of multipath effect, NLOS propagation and other factors, the existing indoor positioning technology has bottleneck problems of low positioning accuracy, high complexity and the like, and thus the increasingly growing demand of each field for indoor positioning accuracy is far from being met. Therefore, the research on high-precision indoor positioning technology has become one of the important research tasks in this field.
Global Positioning System (GPS) is the most widely used and popular Positioning technology. However, since the GPS signal is easily affected by propagation factors such as occlusion and multipath, the attenuation of the positioning signal after reaching the ground is very severe. Therefore, the positioning accuracy of the GPS is very low, and is not suitable for positioning an indoor target. The current wireless positioning technology capable of realizing indoor positioning mainly comprises: a positioning system based on Wireless Fidelity (Wi-Fi), a positioning technology based on infrared ray, a positioning technology based on ultrasonic wave, a positioning technology based on Radio Frequency Identification (RFID), a bluetooth positioning technology, a positioning system based on Wireless Local Area Network (WLAN), a positioning technology based on ZigBee, and the like. Although these techniques are widely used, they cannot achieve centimeter-level indoor positioning accuracy, and are far from meeting the increasing demands of people for high-accuracy positioning. An Ultra-wideband (UWB) positioning technology is a new indoor positioning technology, and is very different from a traditional indoor positioning technology. The pulse width adopted by UWB positioning is extremely narrow, only nanosecond level, and is far less than multipath propagation delay, so that the UWB positioning has higher time and multipath resolution, and is very suitable for indoor precise positioning. However, in practical application, the positioning accuracy is still relatively high, reaching 5-20cm, so that centimeter-level accurate positioning cannot be realized in the true sense.
The 60GHz pulse duration is extremely short, usually about hundred picoseconds or even shorter, and is far less than multipath propagation delay, and multipath signals can be effectively separated at a receiving end, so that the receiving end has higher time resolution and multipath resolution, centimeter-level or even millimeter-level positioning accuracy can be realized theoretically, and an effective way is provided for designing and realizing a high-accuracy indoor positioning system.
In the aspect of indoor wireless positioning, currently commonly used positioning algorithms can be roughly divided into two types: geometric positioning algorithms and non-geometric positioning algorithms. The former is mainly suitable for positioning under the conditions of good channel environments such as Line of Sight (LOS) and few reflection; the latter is mainly suitable for positioning under severe environments with serious factors such as multipath effect, NLOS propagation and the like. The commonly used geometric positioning algorithm mainly comprises: a positioning algorithm based on Time of Arrival (TOA) estimation, a positioning algorithm based on Time difference of Arrival (TDOA) estimation, a positioning algorithm based on RSS (Received Signal Strength) estimation, a positioning algorithm based on angle of Arrival (angle of Arrival, AOA) estimation, or a joint positioning algorithm of the above positioning algorithms, for example: the joint positioning algorithm of AOA and TOA, and the joint positioning algorithm of TOA and TDOA. The TOA and the TDOA can fully utilize the characteristic of high time resolution of 60GHz pulse signals to obtain higher ranging accuracy.
However, ranging and positioning under NLOS of 60GHz also face many problems and challenges due to the fact that 60GHz pulse signals are attenuated by propagation, do not easily penetrate obstacles, and have weak multipath effects. The fingerprint positioning algorithm is a common method for positioning 60GHz pulses under NLOS at present. The fingerprint positioning algorithm can fully utilize adverse factors such as multipath effect and NLOS (non-linear operating system) to create a position fingerprint feature library, and the position coordinates of the target node are determined by matching the measured feature information of the target node with the information in the fingerprint feature library. Compared with other positioning methods, the fingerprint positioning method can be suitable for positioning areas with serious factors such as multipath effect, NLOS (non-line-of-sight) propagation and the like, and has the defects of high positioning accuracy, high calculation complexity and poor environmental adaptability. Therefore, it is necessary to solve the problem of accurate indoor positioning of the 60GHz pulse in NLOS.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides an indoor high-precision positioning method based on an optimal path, which overcomes the problems of low positioning precision, high complexity and the like in an NLOS environment and realizes indoor precise positioning in the NLOS environment.
An indoor high-precision positioning method based on an optimal path comprises the following steps:
adopting a multi-branch tree to segment a region to be positioned, and establishing a dictionary matrix corresponding to all father nodes in the multi-branch tree, wherein the dictionary matrix comprises actual measurement distances between all child nodes corresponding to the father nodes and each reference base station;
a target point to be positioned transmits a positioning signal for receiving by the M reference base stations;
the positioning calculation center server receives the measurement distances sent by the M reference base stations to form an observation vector, determines an optimal node of each layer of the multi-branch tree in the area to be positioned according to the dictionary matrix and the observation vector, determines an optimal path, and takes a position corresponding to a leaf node at the bottom layer of the optimal path as a positioning result of the target point;
and repeatedly positioning for many times, and taking out the position corresponding to the bottom leaf node with the maximum probability as the final positioning result of the target point.
Further, the positioning signal transmitted by the target point to be positioned is a 60GHz pulse.
Further, the dividing the region to be located by using the multi-way tree includes:
taking the center of a region to be positioned as a root node, averagely dividing the region to be positioned into N sub-regions, wherein the center of each sub-region and the center of the region to be positioned are N +1 sub-nodes of the root node;
and taking each child node as a father node, iteratively establishing a multilayer multi-branch tree in the N +1 child regions, and dividing the region to be positioned.
Further, the dictionary matrix established by each parent node in the multi-branch tree and containing the actual measured distance between each child node and each reference base station comprises:
setting a characteristic vector for each node of the multi-branch tree, wherein the characteristic vector comprises an actual measured distance between the node and each reference base station;
and constructing a dictionary matrix of the parent node by taking the feature vector of each child node corresponding to the parent node as a column vector.
Further, the step of receiving, at the positioning calculation center server, the measurement distances sent from the M reference base stations to form an observation vector, and determining an optimal node of each layer of the multi-way tree in the region to be positioned according to the dictionary matrix and the observation vector includes:
step (51): determining an optimal node of each layer according to the dictionary matrix and the observation vector by adopting a Dice coefficient matching criterion, connecting the optimal nodes of each layer to form an optimal path, and taking a position corresponding to a node of the bottom layer of the optimal path as a positioning result of the iteration to a target point;
step (52): updating the optimal path set, updating the father node, returning to the step (52) to continue the iteration if the iteration times is less than the depth of the multi-branch tree, and stopping the iteration and entering the next step if the iteration times is more than or equal to the depth of the multi-branch tree;
step (53): and taking the position corresponding to the optimal node selected by the last iteration as a positioning result of the current positioning process on the target point.
Further, an evaluation algorithm is adopted to analyze the positioning error of the final target point and evaluate the positioning performance.
Furthermore, each node in the multi-branch tree is uniquely identified by adopting a path sequence, and the path sequence represents all paths which need to be passed by reaching the node.
The invention also provides an indoor high-precision positioning device based on the optimal path, which comprises:
the segmentation module is used for segmenting the region to be positioned by adopting a multi-branch tree;
the dictionary matrix establishing module is used for establishing a dictionary matrix corresponding to all father nodes in the multi-branch tree, wherein the dictionary matrix comprises actual measurement distances between all child nodes corresponding to each father node in the multi-branch tree and each reference base station;
the signal transmitting device is used for transmitting a positioning signal at a target point to be positioned;
a reference base station for receiving the positioning signal;
the optimal path determining module is used for receiving the measuring distance sent by the reference base station to form an observation vector, determining an optimal node of each layer of the multi-branch tree in the area to be positioned according to the dictionary matrix and the observation vector, determining an optimal path, and taking a position corresponding to a leaf node at the bottom layer of the optimal path as a positioning result of the target point;
and the repeated positioning module is used for repeatedly positioning for multiple times, and the position corresponding to the bottom leaf node with the maximum probability is taken out to be used as the positioning result of the final target node.
Further, the device also comprises an evaluation module for analyzing the positioning error of the final target point by adopting an evaluation algorithm and evaluating the positioning performance.
The invention also provides an indoor high-precision positioning system based on the optimal path, which comprises the following components:
the signal transmitting equipment is used for transmitting a positioning signal at a target point to be positioned;
the reference base station is used for receiving the positioning signal and sending a measurement distance;
the calculation center server is used for adopting the multi-branch tree to segment the region to be positioned, establishing a dictionary matrix of all father nodes in the multi-branch tree, receiving the measurement distance sent by the reference base station to form an observation vector, determining the optimal node of each layer of the multi-branch tree in the region to be positioned according to the dictionary matrix and the observation vector, determining the optimal path, and taking the position corresponding to the leaf node at the bottom layer of the optimal path as the positioning result of the target node; and repeatedly positioning for many times, and taking out the position corresponding to the bottom leaf node with the maximum probability as the positioning result of the final target node.
The invention has the beneficial effects that:
the invention selects a carrier pulse scheme to generate 60GHz pulses. The carrier wave pulse generation scheme is a pulse design scheme for shifting a pulse-based baseband signal spectrum to a 60GHz frequency band by using a sine or cosine carrier wave with a center frequency of 60 GHz. When the scheme is used for generating 60GHz pulses, narrow pulses with nanosecond pulse width can be generated, the requirement is relatively low, the center frequency and the power spectral density of the whole signal can be adjusted easily by controlling carrier frequency and baseband pulse waveforms, and the pulse generation difficulty is simpler than that of a carrier-free pulse mode under the existing technical condition.
In NLOS environment, during signal propagation, obstruction of obstacles and the like can cause phenomena such as signal reflection and diffraction, which make signal reach differentThe intensity of the location, the time difference of the same signal arriving at the location via different paths, and other information have differences, and the superposition of the factors forms the characteristics of the location, which are unique in all location areas and have almost zero possibility of complete coincidence. Therefore, the invention utilizes adverse factors such as multipath effect and NLOS between each node and each reference base station to create a characteristic vector P (O) of each nodeΛ,i)。
The invention mainly adopts the strategy of optimal path layering according to an observation vector Y of online measurement and a search space established in an offline stage, and completes the positioning of a target point by selecting an optimal node in each layer and taking a corresponding path to form an optimal path. If the optimal node is not selected in one of the layers, the optimal node selected in each layer below will be wrong, which may result in a very large positioning error and even a positioning failure. In order to overcome the problem and improve the robustness and stability of the system, the invention adopts multiple positioning, selects the coordinate position corresponding to the leaf node with the highest probability as the final positioning result of the target point, and can greatly improve the probability of successful positioning.
The invention adopts the strategy of optimal path layering, selects an optimal node on each layer, and takes the corresponding coordinate as the positioning result of the target point after the iteration, wherein the path corresponding to the optimal node is the optimal path selected by the iteration on the layer. Therefore, the process of searching for the optimal path is essentially a process of hierarchically selecting the optimal node. By continuous iteration, the positioning result of the target point can continuously approach to the actual position of the target node. And finally, the position corresponding to the leaf node at the bottommost layer of the optimal path is the positioning result of the target point. When the area to be positioned is changed or the required positioning accuracy is changed, the iteration times can be adjusted according to the actual situation so as to meet the requirement of the actual positioning accuracy.
According to the method, the target point is positioned by adopting a Dice coefficient matching criterion, and the correlation can be calculated by fully utilizing each element in the vector, so that the Dice coefficient can effectively distinguish two similar atoms in a dictionary with higher similarity between the atoms, the atoms matched with the measurement signal Y are more accurately selected, and the probability of correctly selecting the optimal node is improved.
Drawings
Fig. 1 is a positioning flow chart of the present invention.
Fig. 2 is a region to be located of the present invention.
Fig. 3 is a 3-level 9-ary tree built by the present invention.
Fig. 4 is a 60GHz pulse waveform used with the present invention.
The specific implementation mode is as follows:
the invention will be further illustrated with reference to the following examples and drawings:
it should be noted that the following detailed description is exemplary and is intended to provide further explanation of the disclosure. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments according to the present application. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
An exemplary embodiment of the present invention is an indoor high-precision positioning method based on an optimal path, including:
adopting a multi-branch tree to segment a region to be positioned, and establishing a dictionary matrix corresponding to all father nodes in the multi-branch tree, wherein the dictionary matrix comprises actual measurement distances between all child nodes corresponding to the father nodes and each reference base station;
a target point to be positioned transmits a positioning signal for receiving by the M reference base stations;
the positioning calculation center server receives the measurement distances sent by the M reference base stations to form an observation vector, determines an optimal node of each layer of the multi-branch tree in the area to be positioned according to the dictionary matrix and the observation vector, determines an optimal path, and takes a position corresponding to a leaf node at the bottom layer of the optimal path as a positioning result of the target point;
and repeatedly positioning for many times, and taking out the position corresponding to the bottom leaf node with the maximum probability as the final positioning result of the target point.
In the design of the embodiment in the offline stage, the step of dividing the region to be located by using the multi-way tree generally needs to be divided into two steps:
A. establishing a search space, wherein the establishment of the search space is realized by constructing a 9-branch tree by using an area to be positioned, and the coordinate position of each node in the 9-branch tree is saved in a calculation center server. And then, randomly installing M reference base stations in the area to be positioned, determining the coordinate position of each reference base station, and storing the coordinate position to a positioning calculation center server.
The specific method comprises the following steps: suppose that the three-dimensional region to be positioned selects a cubic region with the side length of L. The invention selects a 9-fork tree which is established from a central point O of an area to be positioned as a search space. The establishment method comprises the following steps: taking the central point O of the area to be positioned and recording as (x)0,y0,z0) Dividing the area to be positioned into 8 large grids as shown in fig. 2 according to the separation rule of 8 quadrants by taking the point O as the center, and respectively recording the central points of the 8 large grids as (O)1,O2,K,O8) And the coordinate position of the center point of the 8 large grids is taken as the coordinate of the grid and is marked as (x)i,yi,zi) (1 ≦ i ≦ 8), and the positional coordinate relationship with the O point may be expressed as:
selecting the coordinate position corresponding to the O point as the initial position of TN, taking the O as the root node of the 9-fork tree, and selecting the { O, O1,O2,K,O 89 points are used as child nodes of the root node. Then, each child node is taken as a father node, and the large grid corresponding to each father node is planned againDividing into 8 sub grids to construct a new sub node for each father node, and after 2 iterations, establishing a 3-layer 9-fork number as shown in fig. 3. According to the method, iteration is carried out continuously, and after h iteration, the grid center point coordinates of 8 child nodes corresponding to each parent node can be represented as follows:
wherein i is more than or equal to 1 and less than or equal to 8, H is more than or equal to 1 and less than or equal to (H-1), (x)h-1,yh-1,zh-1) The coordinate value representing the parent node, H is called the depth of the 9-way tree, and (H-1) represents the total number of iterations.
B. Constructing a feature vector P (O) corresponding to each node in the search space one by oneΛI), then, constructing a corresponding dictionary matrix in the father node according to the feature vectors of the 9 child nodes corresponding to each father node
The specific method comprises the following steps: first, the 9 paths corresponding to each parent node are numbered in the manner shown in FIG. 3. Therefore, each node in the search space can be uniquely identified by a sequence of numbers, which is called a path sequence and is denoted as Λ, if the node is located at the h-th level of the search tree, then Λ is a row vector with 1 × (h-1) dimension, and the corresponding node is denoted as OΛThis clearly indicates all the paths that need to be taken to reach this node. For example, the 1 st node at layer 3 may be represented as O[1,1]Then Λ ═ 1,1](ii) a The 2 nd node of layer 3 may be represented as O[1,2]，Λ＝[1,2]。
Constructing a feature vector P (O) corresponding to each node in the search spaceΛ,i)＝[d1,i,d2,i,...,dm,i,...,dM,i]T. Wherein, P (O)ΛI) represents the O thΛThe characteristic vector of each node, i represents the ith (i is more than or equal to 0 and less than or equal to 9) child node of the node which is the parent node of the node, and Λ is the path sequence of the node.The construction method comprises the following steps: element d in feature vectorm,iIn order to send 60GHz pulse signals at the node, after transmission through an IEEE802.15.3c channel, the time delay tau between the node and the mth reference base station is measured by utilizing the ranging based on the TOAm,iThen according to the delay taum,iObtaining the actual measurement distance d between the node and the mth reference base stationm,i＝c*τm,i. That is, the actual measured distances between the node and the M reference base stations constitute a distance feature vector P (O) of the nodeΛ,i)。
With OΛFor the father node, take OΛFeature vector construction node O of corresponding 9 child nodesΛCorresponding dictionary matrix
As follows:
measurement matrix OΛElement d in (1)m,nTo be at node OΛThe nth sub-node sends a 60GHz pulse signal, and after transmission through an IEEE802.15.3c channel, the propagation delay tau between the mth reference base station and the nth sub-node is measured by the mth reference base station through ranging based on TOAm,nThen according to dm,n＝c*τm,nAnd calculating the actual measurement distance between the nth grid point and the mth reference base station.
The information that each node in the 9-way tree needs to store includes: node name, node coordinate, path sequence, and feature vector P (O) corresponding to nodeΛI) and dictionary matrix
As shown in table 1:
TABLE 1 representation of nodes
Next, we enter the on-line positioning stage, and first we perform the following steps
A. Randomly select 1 target point, assuming the actual position of the target point is (x, y, z). Initializing parameters: n is 1, time is 10, locations0Phi is given. Wherein n represents the nth positioning; time represents the total number of fixes; locations are used to save the time's time-based location results.
B. Transmitting 60GHz pulse signals by a target point to be positioned, respectively receiving the pulse signals at M reference base stations, and obtaining a measurement vector Y ═ Y1,y2,...,yM]T. The specific scheme is as follows: y ismIn order to transmit a 60GHz pulse signal at the target point, after transmission through an IEEE802.15.3c channel, a signal is received at the mth reference base station, and the actual measured distance between the target point and the mth reference base station is measured by using TOA-based ranging.
Then, determining the optimal node of each layer of the multi-branch tree in the region to be positioned according to the dictionary matrix and the observation vector, and specifically adopting the following steps:
(a) initialization: lambda0＝φ，Γ＝0，Φ＝φ，posit＝O，h＝1(1≤h≤(H-1))。
Wherein, Λ0The optimal path for storing the selection of each layer is a row vector of (H-1) dimension; Γ represents the optimal path selected at each iteration; phi is an M multiplied by 9 matrix used for storing a dictionary matrix constructed in each iteration; posit is used for storing the optimal node selected in each iteration, and the initial value is set as a central point O; h denotes the H-th iteration, and the maximum number of iterations is (H-1).
(b) Selecting feature vectors corresponding to 9 child nodes corresponding to posit to construct a dictionary matrix phi as follows:
and then selecting an optimal node from 9 sub-nodes of posit according to a Dice coefficient matching criterion by using the observation vector Y and the dictionary matrix phi.
The specific selection steps are as follows: first of the matrix phi
The Dice coefficient matching criterion between a column and an observation vector Y may be expressed as:
take max [ dice (i)]The corresponding node is the optimal node, and the path corresponding to the optimal node is represented as gammahAnd the coordinate position corresponding to the optimal node is the positioning result of the algorithm on the target node after the iteration.
(c) Updating optimal path set Lambdah＝Λh-1∪ΓhUpdating the parent node
H is H +1, if H is less than or equal to (H-1), returning to the step (b) for continuing the iteration, and if H is greater than (H-1), stopping the iteration and entering the next step.
(d) Optimal node selected with last iteration
As the positioning result of the positioning process to the target point, the positioning result locates of this time is savedn＝locatesn-1∪posit。
D. Positioning for multiple times, and finally taking out the leaf node with the maximum occurrence probability as the final positioning result of the target point
(a) Updating the parameter n to n + 1; if n is less than or equal to time, returning to the step B, and if n is more than time, stopping iteration and entering the next step.
(b) Taking the leaf node with the maximum occurrence probability in locaets as the final positioning result of the target point, and recording the result as the final positioning result of the target point
Step three, the actual coordinates of the target node are (x, y, z), and the positioning result of the positioning algorithm on the target point is
The mean square error is used to evaluate the positioning performance, and therefore, the positioning error can be expressed as:
in the process a of the first step, the depth H of the tree depends on the positioning accuracy required by positioning, and the depth determination method of the 9-fork tree is as follows:
where l represents the side length of the small grid at the last iteration. The higher the precision required by positioning, the smaller the side length l of the small grid during the last iteration, the more the required iteration times, and the deeper the depth of the 9-fork tree. Meanwhile, along with the continuous reduction of the side length l of the small grid, the storage space required by positioning is also continuously increased, the calculation complexity of positioning is higher and higher, and the positioning time is longer and longer. Therefore, in the actual positioning process, factors in various aspects such as positioning accuracy, positioning complexity, positioning time and the like are comprehensively considered, and the iteration number is reasonably selected.
Another embodiment of the present invention is an indoor high-precision positioning apparatus based on an optimal path, including:
the segmentation module is used for segmenting the region to be positioned by adopting a multi-branch tree;
the dictionary matrix establishing module is used for establishing a dictionary matrix corresponding to all father nodes in the multi-branch tree, wherein the dictionary matrix comprises actual measurement distances between all child nodes corresponding to each father node in the multi-branch tree and each reference base station;
the signal transmitting device is used for transmitting a positioning signal at a target point to be positioned;
a reference base station for receiving the positioning signal;
the optimal path determining module is used for receiving the measuring distance sent by the reference base station to form an observation vector, determining an optimal node of each layer of the multi-branch tree in the region to be positioned according to the dictionary matrix and the observation vector, determining an optimal path, and taking a position corresponding to a leaf node at the bottom layer of the optimal path as a positioning result of the target node;
and the repeated positioning module is used for repeatedly positioning for multiple times, and the position corresponding to the bottom leaf node with the maximum probability is taken out to be used as the positioning result of the final target node.
Further, the device also comprises an evaluation module which is used for analyzing the positioning error of the final target node by adopting an evaluation algorithm and evaluating the positioning performance.
Another embodiment of the present invention is an indoor high-precision positioning system based on an optimal path, including:
the signal transmitting equipment is used for transmitting a positioning signal at a target point to be positioned;
the reference base station is used for receiving the positioning signal and sending a measurement distance;
the positioning calculation center server is used for adopting the multi-branch tree to segment the region to be positioned, establishing a dictionary matrix of all father nodes in the multi-branch tree, receiving the measurement distance sent by the reference base station to form an observation vector, determining the optimal node of each layer of the multi-branch tree in the region to be positioned according to the dictionary matrix and the observation vector, determining the optimal path, and taking the position corresponding to the leaf node at the bottom layer of the optimal path as the positioning result of the target node; and repeatedly positioning for many times, and taking out the position corresponding to the bottom leaf node with the maximum probability as the positioning result of the final target node.
The invention selects a carrier pulse scheme to generate 60GHz pulses. The carrier wave pulse generation scheme is a pulse design scheme for shifting a pulse-based baseband signal spectrum to a 60GHz frequency band by using a sine or cosine carrier wave with a center frequency of 60 GHz. When the scheme is used for generating 60GHz pulses, narrow pulses with nanosecond pulse width can be generated, the requirement is relatively low, the center frequency and the power spectral density of the whole signal can be adjusted easily by controlling carrier frequency and baseband pulse waveforms, and the pulse generation difficulty is simpler than that of a carrier-free pulse mode under the existing technical condition. The pulse expression obtained using this scheme is:
P(t)＝p(t)cos(2πfct)
wherein f iscThe frequency of the signal for spectrum shifting, p (t), is a baseband pulse, and in practical application, a pulse which is relatively easy to generate is generally used as the baseband pulse. Thus, the expression for the pulse waveform is:
where a is 1, α is 0.25ns, pulse width is about 0.5ns, and center frequency fc＝60.5GHz。
The method selects a typical 60GHz channel model IEEE802.15.3c, the channel environment adopts a channel model CM2 of an indoor living environment under NLOS, and an MATLAB is adopted to carry out simulation research on the positioning scheme provided by the invention, the positioning research of the invention aims at 3-dimensional, a square area with the side length of 5m is selected for an area to be positioned, 20 reference base stations are adopted for positioning, the SNR (signal to noise ratio) is 20dB, the method utilizes 60GHz pulses to realize centimeter-level indoor positioning accuracy, and a specific result of 20 positioning tests is listed in a table 1.
TABLE 1 Single target location results (unit: meter) based on optimal path under NLOS environment
Through calculation, the average positioning error of the results of 20 times of positioning simulation experiments is 0.0972m, wherein 11 times of positioning accuracy are below 0.05m, 6 times of positioning accuracy are between 0.05m and 0.1m, 1 time is between 0.1m and 0.15m, and only 2 times of error is more than 0.2 m. It can be seen that the positioning algorithm provided by the invention can realize centimeter-level indoor positioning accuracy in NLOS environment at a large probability.
In NLOS environment, signal is transmittedDuring the broadcasting process, the blocking of obstacles and the like can cause phenomena such as signal reflection and diffraction, which cause differences in information such as the intensity of signals arriving at different places, the time difference of the same signal arriving at a place after propagating through different paths, and the superposition of the factors forms the characteristics of the position, and the characteristics are unique in all position areas, and the possibility of complete coincidence is almost zero. Therefore, the invention utilizes adverse factors such as multipath effect and NLOS between each node and each reference base station to create a characteristic vector P (O) of each nodeΛ,i)。
The invention mainly adopts the strategy of optimal path layering according to an observation vector Y of online measurement and a search space established in an offline stage, selects an optimal node in each layer, and takes a corresponding path to form an optimal path so as to complete the positioning of a target point. If the optimal node is not selected in one of the layers, the optimal node selected in each layer below will be wrong, which may result in a very large positioning error and even a positioning failure. In order to overcome the problem and improve the robustness and stability of the system, the invention adopts multiple positioning, selects the coordinate position corresponding to the leaf node with the highest probability as the final positioning result of the target point, and can greatly improve the probability of successful positioning.
The invention adopts the strategy of optimal path layering, selects an optimal node on each layer, and takes the corresponding coordinate as the positioning result of the target point after the iteration, wherein the path corresponding to the optimal node is the optimal path selected by the iteration on the layer. Therefore, the process of searching for the optimal path is essentially a process of hierarchically selecting the optimal node. By continuous iteration, the positioning result of the target point can continuously approach to the actual position of the target node. And finally, the position corresponding to the leaf node at the bottommost layer of the optimal path is the positioning result of the target point. When the area to be positioned is changed or the required positioning accuracy is changed, the iteration times can be adjusted according to the actual situation so as to meet the requirement of the actual positioning accuracy.
The selection of the optimal node is the key for realizing the positioning of the target point, and the algorithm selects a node corresponding to the atom which is most matched with the measurement vector Y from the matrix phi as the optimal node during each iteration. In the positioning process, the side length of the grid is smaller and smaller along with the increase of the iteration times, so that the similarity between atoms in the dictionary matrix is higher and higher, and any two similar atoms in the dictionary matrix can interfere with the matching process of the optimal atoms. The method selects the Dice coefficient matching criterion to complete the selection of the optimal atom, and selects the atom which is most matched with the measurement signal Y from the given dictionary matrix phi. For any two vectors x and y, the Dice matching coefficient is defined as:
wherein x is (x)1,x2,K,xn)，y＝(y1,y2,K,yn). The Dice coefficient matching criterion can fully utilize each element of the vector to calculate the correlation, so that the Dice coefficient can effectively distinguish two similar atoms in a dictionary with higher similarity between the atoms, the atoms matched with the measurement signal Y are more accurately selected, and the probability of correctly selecting the optimal node is improved.
The above description is only a preferred embodiment of the present application and is not intended to limit the present application, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims (9)
1. An indoor high-precision positioning method based on an optimal path is characterized by comprising the following steps:
adopting a multi-branch tree to segment a region to be positioned, and establishing a dictionary matrix corresponding to all father nodes in the multi-branch tree, wherein the dictionary matrix comprises actual measurement distances between all child nodes corresponding to the father nodes and each reference base station;
a target point to be positioned transmits a positioning signal for receiving by the M reference base stations;
the positioning calculation center server receives the measurement distances sent by the M reference base stations to form an observation vector, determines an optimal node of each layer of the multi-branch tree in the area to be positioned according to the dictionary matrix and the observation vector, determines an optimal path, and takes a position corresponding to a leaf node at the bottom layer of the optimal path as a positioning result of the target point;
repeatedly positioning for many times, and taking out the position corresponding to the bottom leaf node with the highest probability as the final positioning result of the target point;
the positioning calculation center server receives the measurement distances sent by the M reference base stations to form an observation vector, and the step of determining the optimal node of each layer of the multi-branch tree in the region to be positioned according to the dictionary matrix and the observation vector comprises the following steps:
step (51): determining an optimal node of each layer according to the dictionary matrix and the observation vector by adopting a Dice coefficient matching criterion, connecting the optimal nodes of each layer to form an optimal path, and taking a position corresponding to a node of the bottom layer of the optimal path as a positioning result of the iteration to a target point;
step (52): updating the optimal path set, updating the father node, returning to the step (52) to continue the iteration if the iteration times is less than the depth of the multi-branch tree, and stopping the iteration and entering the next step if the iteration times is more than or equal to the depth of the multi-branch tree;
step (53): and taking the position corresponding to the optimal node selected by the last iteration as a positioning result of the current positioning process on the target point.
2. The method of claim 1, wherein: and the positioning signal transmitted by the target point to be positioned is a 60GHz pulse signal.
3. The method of claim 1, wherein: the method for dividing the region to be positioned by adopting the multi-branch tree comprises the following steps:
taking the center of a region to be positioned as a root node, averagely dividing the region to be positioned into N sub-regions, wherein the center of each sub-region and the center of the region to be positioned are N +1 sub-nodes of the root node;
and taking each child node as a father node, iteratively establishing a multilayer multi-branch tree in the N +1 child regions, and dividing the region to be positioned.
4. The method of claim 1, wherein: the dictionary matrix established by each parent node in the multi-branch tree and containing the actual measured distance between each child node and each reference base station comprises the following components:
setting a characteristic vector for each node of the multi-branch tree, wherein the characteristic vector comprises an actual measured distance between the node and each reference base station;
and constructing a dictionary matrix of the parent node by taking the feature vector of each child node corresponding to the parent node as a column vector.
5. The method of claim 1, wherein: and analyzing the positioning error of the final target point by adopting an evaluation algorithm, and evaluating the positioning performance.
6. The method of claim 1, wherein: and each node in the multi-branch tree is uniquely identified by adopting a path sequence, wherein the path sequence represents all paths which need to be passed by reaching the node.
7. An indoor high-precision positioning device based on an optimal path is characterized by comprising:
the segmentation module is used for segmenting the region to be positioned by adopting a multi-branch tree;
the dictionary matrix establishing module is used for establishing a dictionary matrix corresponding to all father nodes in the multi-branch tree, wherein the dictionary matrix comprises actual measurement distances between all child nodes corresponding to each father node in the multi-branch tree and each reference base station;
the signal transmitting device is used for transmitting a positioning signal at a target point to be positioned;
a reference base station for receiving the positioning signal;
the optimal path determining module is used for receiving the measuring distance sent by the reference base station to form an observation vector, determining an optimal node of each layer of the multi-branch tree in the region to be positioned according to the dictionary matrix and the observation vector, determining an optimal path, and taking a position corresponding to a leaf node at the bottom layer of the optimal path as a positioning result of the target node;
the repeated positioning module is used for repeatedly positioning for multiple times, and the position corresponding to the bottom node with the highest current probability is taken out to be used as the positioning result of the final target node;
the optimal path determining module is specifically configured to receive measurement distances sent from M reference base stations to form an observation vector, and determining an optimal node of each layer of a multi-way tree in a region to be located according to the dictionary matrix and the observation vector includes:
step (51): determining an optimal node of each layer according to the dictionary matrix and the observation vector by adopting a Dice coefficient matching criterion, connecting the optimal nodes of each layer to form an optimal path, and taking a position corresponding to a node of the bottom layer of the optimal path as a positioning result of the iteration to a target point;
step (52): updating the optimal path set, updating the father node, returning to the step (52) to continue the iteration if the iteration times is less than the depth of the multi-branch tree, and stopping the iteration and entering the next step if the iteration times is more than or equal to the depth of the multi-branch tree;
step (53): and taking the position corresponding to the optimal node selected by the last iteration as a positioning result of the current positioning process on the target point.
8. The indoor high-precision positioning device based on the optimal path as claimed in claim 7, further comprising an evaluation module for analyzing the positioning error of the final target point by using an evaluation algorithm to evaluate the positioning performance.
9. An indoor high-precision positioning system based on an optimal path, comprising:
the signal transmitting equipment is used for transmitting a positioning signal at a target point to be positioned;
the reference base station is used for receiving the positioning signal and sending a measurement distance;
the positioning calculation center server is used for adopting the multi-branch tree to segment the to-be-positioned area, establishing a dictionary matrix of all father nodes in the multi-branch tree, receiving the measurement distance sent by the reference base station to obtain an observation vector, determining an optimal node of each layer of the multi-branch tree in the to-be-positioned area according to the dictionary matrix and the observation vector, determining an optimal path, and taking a position corresponding to a leaf node at the bottom layer of the optimal path as a positioning result of the target node; repeatedly positioning for many times, and taking the position corresponding to the bottom leaf node with the highest probability as the positioning result of the final target node;
receiving the measurement distances sent by the M reference base stations at a positioning calculation center server to form an observation vector, and determining the optimal node of each layer of the multi-branch tree in the region to be positioned according to the dictionary matrix and the observation vector comprises the following steps:
step (51): determining an optimal node of each layer according to the dictionary matrix and the observation vector by adopting a Dice coefficient matching criterion, connecting the optimal nodes of each layer to form an optimal path, and taking a position corresponding to a node of the bottom layer of the optimal path as a positioning result of the iteration to a target point;
step (52): updating the optimal path set, updating the father node, returning to the step (52) to continue the iteration if the iteration times is less than the depth of the multi-branch tree, and stopping the iteration and entering the next step if the iteration times is more than or equal to the depth of the multi-branch tree;
step (53): and taking the position corresponding to the optimal node selected by the last iteration as a positioning result of the current positioning process on the target point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710762494.0A CN107566981B (en) | 2017-08-30 | 2017-08-30 | Indoor high-precision positioning method, device and system based on optimal path |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710762494.0A CN107566981B (en) | 2017-08-30 | 2017-08-30 | Indoor high-precision positioning method, device and system based on optimal path |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107566981A CN107566981A (en) | 2018-01-09 |
| CN107566981B true CN107566981B (en) | 2020-02-04 |
Family
ID=60978022
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710762494.0A Active CN107566981B (en) | 2017-08-30 | 2017-08-30 | Indoor high-precision positioning method, device and system based on optimal path |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107566981B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP7417500B2 (en) * | 2020-09-14 | 2024-01-18 | 株式会社東芝 | Information processing device, information processing method, computer program and travel management system |
| CN116976381B (en) * | 2023-09-25 | 2023-12-01 | 深圳市国芯物联科技有限公司 | Position positioning and tracking method in radio frequency identification system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106231549A (en) * | 2016-07-25 | 2016-12-14 | 青岛科技大学 | A kind of 60GHz pulse indoor orientation method based on restructing algorithm |
| CN106950537A (en) * | 2017-04-28 | 2017-07-14 | 浙江大学 | A kind of Distributed localization method based on UWB |
| CN107015193A (en) * | 2017-04-18 | 2017-08-04 | 中国矿业大学(北京) | A kind of binocular CCD vision mine movable object localization methods and system |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7479918B2 (en) * | 2006-11-22 | 2009-01-20 | Zimmerman Associates, Inc. | Vehicle-mounted ultra-wideband radar systems and methods |
-
2017
- 2017-08-30 CN CN201710762494.0A patent/CN107566981B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106231549A (en) * | 2016-07-25 | 2016-12-14 | 青岛科技大学 | A kind of 60GHz pulse indoor orientation method based on restructing algorithm |
| CN107015193A (en) * | 2017-04-18 | 2017-08-04 | 中国矿业大学(北京) | A kind of binocular CCD vision mine movable object localization methods and system |
| CN106950537A (en) * | 2017-04-28 | 2017-07-14 | 浙江大学 | A kind of Distributed localization method based on UWB |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107566981A (en) | 2018-01-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yan et al. | Review of range-based positioning algorithms | |
| Smith et al. | Tracking moving devices with the cricket location system | |
| US20100309059A1 (en) | Method and Apparatus of Positioning for a Wireless Communication System | |
| WO2006088599A1 (en) | Systems and methods for positioning using multipath signals | |
| WO2016112758A1 (en) | Method and apparatus for locating terminal | |
| CN106231549A (en) | A kind of 60GHz pulse indoor orientation method based on restructing algorithm | |
| Cheng et al. | Real time indoor positioning system for smart grid based on uwb and artificial intelligence techniques | |
| CN104076327A (en) | Continuous positioning method based on search space reduction | |
| Arsan et al. | A clustering-based approach for improving the accuracy of UWB sensor-based indoor positioning system | |
| Chen et al. | An indoor location system based on neural network and genetic algorithm | |
| CN107566981B (en) | Indoor high-precision positioning method, device and system based on optimal path | |
| CN101986756A (en) | Time-reversal signal-based wireless positioning scheme | |
| Zhao et al. | Comparing centralized Kalman filter schemes for indoor positioning in wireless sensor network | |
| Jung et al. | Indoor localization using particle filter and map-based NLOS ranging model | |
| Sala et al. | Using neural networks and Active RFID for indoor location services | |
| Kasmi et al. | Algorithms and position optimization for a decentralized localization platform based on resource-constrained devices | |
| Shareef et al. | Localization using extended Kalman filters in wireless sensor networks | |
| Lou et al. | High-accuracy positioning algorithm based on uwb | |
| Arsan | Accurate indoor positioning with ultra-wide band sensors | |
| Elfadil et al. | Indoor navigation algorithm for mobile robot using wireless sensor networks | |
| Iacono et al. | Distance measurement characterization for ultra wide band indoor localization systems | |
| CN115397012A (en) | Realization method of UWB positioning tracking system based on TWR-TDOA estimation and MPGA layout optimization | |
| Darakeh et al. | An accurate distributed rage free localization algorithm for WSN | |
| Ma et al. | A TDOA localization method for complex environment localization | |
| Ismail et al. | Comparison of wireless sensor node localisation between trilateration and multi-lateration methods using rssi |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |