A self-learning wireless signal positioning method, system, device and storage medium

Abstract

This application relates to artificial intelligence and wireless signal positioning technologies, including a self-learning wireless signal positioning method, system, computer device, and storage medium. Periodic signal parameter sets of a terminal to be positioned within a region to be recognized are obtained for parameter recognition to obtain state characterization parameters; training signal parameter sets that can be used for self-learning training are automatically and continuously generated using the state characterization parameters and the signal parameter sets; a pre-constructed self-learning positioning model to be trained is trained using the training signal parameter sets; the coordinate position of the terminal to be positioned is calculated using the trained self-learning wireless signal positioning model, and the terminal coordinates are output. The present invention proposes a solution that does not increase system costs, has a relatively low real-time engineering difficulty for the system, can improve the accuracy of wireless signal positioning, and can automatically adapt to the time-varying characteristics of wireless signals and changes in the on-site environment.

Description

Technical Field

[0001] The present application relates to the fields of artificial intelligence and wireless signal positioning, and particularly to a self-learning wireless signal positioning method, system, computer device, and storage medium. Background Art

[0002] At present, using wireless signals to implement positioning services is one of the common methods. However, due to the instability of wireless signals and the influence of various factors such as occlusion, shielding, reflection, and multipath, the accuracy of simply estimating the position through signal parameters for positioning often cannot achieve satisfactory results. The method of using signal strength to calculate the position of the positioned object has the lowest cost, but is most affected by environmental changes and has the worst accuracy; the method of using the angle of arrival or the angle of departure for positioning, although having relatively high positioning accuracy in an ideal environment, also has relatively high costs; using fingerprint positioning is one of the relatively preferred methods for implementing positioning services at present, but this method also has its limitations. When deploying the system each time, it is necessary to manually set multiple position fingerprint sampling points and measure the signal data at the sampling points, with a large implementation difficulty and workload. Moreover, due to the time-varying characteristics of wireless signals, the signal strength model measured in a certain period may change in another period, resulting in the invalidation of the position fingerprint. Therefore, in order to solve these problems existing in the above methods, it is necessary to propose a solution that does not increase the system cost, has a relatively low real-time engineering difficulty of the system, can improve the accuracy of wireless signal positioning, and can automatically adapt to the time-varying characteristics of wireless signals and changes in the on-site environment. Summary of the Invention

[0003] An object of an embodiment of the present application is to propose a self-learning wireless signal positioning method to solve the problems of relatively large difficulty in sampling training samples for the training model and relatively low signal positioning accuracy in the prior art.

[0004] To solve the above technical problems, the self-learning wireless signal positioning method provided by an embodiment of the present application adopts the following technical solutions:

[0005] Obtain a periodic signal parameter set of a to-be-positioned terminal in a to-be-identified area for parameter identification to obtain a state characterization parameter, where the state characterization parameter is used to characterize the motion state of the to-be-positioned terminal;

[0006] Obtain a signal parameter set of a to-be-positioned terminal in a to-be-identified area for parameter identification to obtain a state characterization parameter;

[0007] Automatically and continuously generate a training signal parameter set that can be used for self-learning training by using the state characterization parameter; train a pre-constructed to-be-trained self-learning positioning model by using the training signal parameter set to obtain a training result set;

[0008] If the training result set does not exceed the training error range, a trained self-learning wireless signal positioning model is obtained;

[0009] Use the self-learning wireless signal positioning model to calculate the coordinate position of the terminal to be positioned and output the terminal coordinates.

[0010] Furthermore, the method further includes:

[0011] Obtain the position data sets of multiple deployed receiving terminals to construct an identification area;

[0012] Based on a preset area division rule, use the position data set to geometrically re-divide the identification area to obtain multiple sub-areas;

[0013] Perform signal strength identification on the signal parameters of the terminal to be positioned received in each sub-area to obtain a signal strength value set;

[0014] Extract the sub-areas corresponding to the signal strength values that meet the conditions in the signal strength value set to form an area to be identified.

[0015] Furthermore, the method further includes:

[0016] Determine the motion state of the terminal to be positioned;

[0017] When the terminal to be positioned is in a stationary state, extract the signal parameter set of the terminal to be positioned in the area to be identified and perform probability accumulation calculation to obtain a probability accumulation value;

[0018] If the probability accumulation value exceeds a preset acceptable value, the state characterization parameter performs a steady-state characterization on the signal strength of the terminal to be positioned to form steady-state signal data;

[0019] If the probability accumulation value does not exceed the preset acceptable value and / or the terminal to be positioned is in a non-stationary state, the state characterization parameter performs a non-steady-state characterization on the signal strength of the terminal to be positioned to form non-steady-state signal data;

[0020] Summarize the steady-state signal parameters and / or the non-steady-state signal parameters to generate a training signal parameter set.

[0021] Furthermore, the method further includes:

[0022] Use any conventional signal strength positioning algorithm to calculate the position of the steady-state signal data to obtain a training coordinate comparison set;

[0023] Calculate whether the training error between the training coordinate result set and the training coordinate comparison set exceeds the training error range;

[0024] If it exceeds, retrain the to-be-trained self-learning wireless signal positioning model until the training error is within the training error range;

[0025] If it does not exceed, obtain the trained self-learning wireless signal positioning model.

[0026] Further, the method further includes:

[0027] Use any conventional signal strength positioning algorithm to calculate the position of the signal parameter set of the to-be-positioned terminal to obtain a first positioning coordinate;

[0028] When the to-be-positioned terminal is in a steady state representation, output the first positioning coordinate as the terminal coordinate;

[0029] And / or retrieve the trained self-learning wireless signal positioning model to calculate the position of the wireless signal parameter set of the to-be-positioned terminal to obtain a second positioning coordinate;

[0030] Assign weights to the first positioning coordinate and the second positioning coordinate for coordinate calculation to obtain a third positioning coordinate;

[0031] When the to-be-positioned terminal is in a non-steady state representation, output the third positioning coordinate as the terminal coordinate.

[0032] Further, the method further includes:

[0033] Extract the error value corresponding to the training error;

[0034] Use the error value to perform correction calculation according to a preset learning rate to obtain a correction amount value;

[0035] Use the correction amount value to correct the to-be-trained self-learning wireless signal positioning model to obtain the corrected to-be-trained self-learning wireless signal positioning model.

[0036] Further, the method further includes:

[0037] Retrain the corrected to-be-trained self-learning wireless signal positioning model until the corresponding error value is within the training error range;

[0038] Stop correcting the to-be-trained self-learning wireless signal positioning model.

[0039] To solve the above technical problems, an embodiment of the present application further provides a self-learning wireless signal positioning system, adopting the following technical solutions:

[0040] A self-learning wireless signal positioning system, the self-learning wireless signal positioning system includes:

[0041] Motion state detection and reporting module: Installed on the terminal to be located, it is used to detect the motion state data of the positioning terminal and report it to the base station through wireless communication.

[0042] Identification module: It is used to obtain the signal parameter set of the terminal to be located in the area to be identified, perform parameter identification, and obtain the state characterization parameters;

[0043] Positioning training module: It is used to train the pre-constructed self-learning positioning model to be trained by using the state characterization parameters and the signal parameter set to obtain the training result;

[0044] Correction module: If the training result does not exceed the training error range, it is used to obtain the trained self-learning wireless signal positioning model;

[0045] Positioning integration module: It is used to calculate the coordinate position of the terminal to be located by using the self-learning wireless signal positioning model and output the terminal coordinates.

[0046] To solve the above technical problems, the embodiment of the present application also provides a computer device, which adopts the following technical solutions:

[0047] A computer device includes a memory and a processor. Computer-readable instructions are stored in the memory. When the processor executes the computer-readable instructions, the steps of the self-learning wireless signal positioning method as described above are implemented.

[0048] To solve the above technical problems, the embodiment of the present application also provides a computer-readable storage medium, which adopts the following technical solutions:

[0049] A computer-readable storage medium, characterized in that computer-readable instructions are stored on the computer-readable storage medium. When the computer-readable instructions are executed by a processor, the steps of the self-learning wireless signal positioning method as described above are implemented.

[0050] Compared with the prior art, the embodiment of the present application mainly has the following beneficial effects:

[0051] In the present application, a certain number of terminal base stations are deployed in the area. The terminal base stations are used to determine whether the signal parameters of the terminal to be located are in a stable state. The signal parameters in the stable state are used to continuously train the self-learning signal positioning model to be trained. Through the above method, the cumbersome work of generating training samples through manual measurement is replaced, the deployment cost is reduced, and at the same time, the self-learning signal positioning model to be trained is further corrected and trained by correction values. The more self-learning training data accumulated according to the above method can greatly improve the positioning accuracy and better adapt to the laws of environmental changes and wireless signal fluctuations. Description of the Drawings

[0052] To more clearly illustrate the solutions in this application, the following will briefly introduce the drawings needed in the description of the embodiments of this application. Obviously, the drawings in the following description are some embodiments of this application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0053] Figure 1 is a flowchart of an embodiment of the self-learning wireless signal positioning method according to this application;

[0054] Figure 2 is a flowchart of an embodiment of step S130 in this application;

[0055] Figure 3 is a structural diagram of an embodiment of the self-learning wireless signal positioning system according to this application

[0056] Figure 4 is a schematic structural diagram of an embodiment of the computer device according to this application. Detailed implementation manners

[0057] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this application belongs; the terms used in the description of this application in the specification are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification, claims, and drawings of this application are intended to cover non-exclusive inclusion. The terms "first", "second", etc. in the specification, claims, or drawings of this application are used to distinguish different objects and not to describe a specific order.

[0058] Referring to "embodiment" herein means that the specific features, structures, or characteristics described in connection with the embodiment can be included in at least one embodiment of this application. The phrase appears in various places in the specification and does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those of ordinary skill in the art explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0059] To enable those in the technical field of this application to better understand the solutions of this application, the following will clearly and completely describe the technical solutions in the embodiments of this application in conjunction with the drawings.

[0060] Continue to refer to Figure 1, which shows a flowchart of an embodiment of self - learning wireless signal positioning proposed according to the present application. The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, Artificial Intelligence (AI) is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to obtain the best results.

[0061] Artificial intelligence basic technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, and mechatronics. Artificial intelligence software technologies mainly include several major directions such as computer vision technology, robotics, biometric technology, speech processing technology, natural language processing technology, and machine learning / deep learning. The self - learning wireless signal positioning method includes the following steps:

[0062] S100: Acquire a periodic signal parameter set of the terminal to be positioned in the area to be recognized for parameter recognition, and obtain a state characterization parameter, where the state characterization parameter is used to characterize the motion state of the terminal to be positioned;

[0063] In this embodiment, it should be noted that for the sake of simplicity in description and considering that most applications are planar positioning, the present application mainly takes two - dimensional planar positioning as an example for illustration. The solution proposed by the present application can also be applied to three - dimensional positioning, and only the corresponding dimensions need to be extended without affecting the essence of the present application.

[0064] Furthermore, the terminal to be positioned is also equipped with any motion state sensor. In this application, an acceleration sensor is taken as an example. The motion state of the terminal to be positioned is measured by the acceleration sensor, and the motion state data is periodically sent to the terminal base station through wireless communication. The motion sensor installed on the terminal to be positioned periodically measures the motion data and then reports the motion data to the base station through wireless communication.

[0065] Furthermore, in a preferred embodiment, before acquiring the signal parameter set of the terminal to be positioned in the area to be recognized for parameter recognition, it also includes the acquisition of the recognition area. First, acquire the position data set of multiple deployed receiving terminals to construct the recognition area; geometrically re - divide the recognition area based on a preset area division rule using the position data set to obtain multiple sub - areas; perform signal strength recognition on the signal parameters of the terminal to be positioned received in each sub - area to obtain a signal strength value set; finally, extract the sub - areas corresponding to the signal strength values that meet the conditions in the signal strength value set to form the area to be recognized.

[0066] Specifically, in this embodiment, the recognition area is constructed by a receiving terminal that can receive the signal parameters sent by the terminal to be located. The receiving terminal includes a terminal base station. After deploying multiple terminal base stations, any form of position connection is performed on the obtained position data of the terminal base stations to obtain a recognition area in a geometric shape. Then, the area to be recognized within the recognition area is further divided by a preset connection method of the terminal base stations. Among them, at least 3 terminal base stations are included in the divided area to be recognized. The signal parameters sent by the terminal to be located are received by multiple terminal base stations within the recognition area to form an initial signal parameter set. The initial signal parameter set is analyzed to obtain an initial intensity value set. The receiving terminals that meet the conditions are extracted from the initial intensity value set according to the preset signal intensity interval, and the corresponding area to be recognized is extracted by using the receiving terminals that meet the conditions.

[0067] Furthermore, parameter recognition is performed on the received signal parameter set. First, the motion state of the terminal to be located is determined; among them, the signal parameter set also includes motion state signal parameters. When the terminal to be located is in a stationary state within a preset time, stationary state data is generated. When the base station terminal determines that the terminal to be located is in a stationary state through the stationary state data, the signal parameter set of the terminal to be located within the area to be recognized is extracted for probability cumulative calculation to obtain a probability cumulative value; if the probability cumulative value exceeds the preset acceptable value, the state characterization parameter performs a steady-state characterization on the signal intensity of the terminal to be located to form steady-state signal data;

[0068] If the probability cumulative value does not exceed the preset acceptable value and / or the terminal to be located is in a non-stationary state, the state characterization parameter performs a non-steady-state characterization on the signal intensity of the terminal to be located to form non-steady-state signal data. It should be noted that when the stationary state data of the terminal to be located is not received, it is determined that the terminal to be located is in a moving state, and a non-steady-state characterization is also performed on the terminal to be located in the moving state.

[0069] Specifically, in this embodiment, first, the motion state of the terminal to be located is determined. When it is in a stationary state, the signal intensity parameters of the terminal to be located are continuously received, the mean value E and the standard deviation σ of the signal intensity are calculated, and the occurrence probability of all wireless signal intensities between [E - σ, E + σ] in the cumulative data is extracted to obtain a probability cumulative value P. When the probability cumulative value P is greater than the acceptable value, it is considered that this terminal has entered a steady state, and steady-state characterization data is embedded in the signal parameters to form steady-state characterization signal data. In this embodiment, the optional acceptable value is 68%. The duration of the cumulative data that does not exceed the acceptable value is statistically analyzed. When the preset statistical duration is exceeded and the acceptable value is not reached, the cumulative statistical data is cleared.

[0070] Specifically, in this embodiment, for signal parameters that do not exceed the preset statistical duration and do not reach the acceptable value, and signal parameters emitted by the to-be-positioned terminal in the moving state, they are included in the non-steady-state characterization data to form non-steady-state characterization signal data.

[0071] S110: Automatically and continuously generate a set of training signal parameters that can be used for self-learning training by using the state characterization parameters;

[0072] Specifically, in this embodiment, the terminal base station receives the state characterization parameters and signal strength sets of all to-be-positioned terminals. When the state characterization parameters indicate that the to-be-positioned terminal has a steady-state characterization, the corresponding signal strength set can be used as a training sample to form a set of training signal parameters; during the operation of the system, the to-be-positioned terminal continuously generates state characterization parameters and signal strength sets. As long as the state characterization parameters of the positioning terminal indicate that it has a steady-state characterization, the corresponding signal strength set can always be used as a set of training signal parameters and continuously used for the training of the self-learning positioning model.

[0073] S120: Train the to-be-trained self-learning positioning model pre-constructed by using the set of training signal parameters to obtain a training result;

[0074] In this embodiment, the learning and training of the to-be-trained self-learning wireless signal positioning model are described by taking the BP neural network method as an example. The to-be-trained self-learning wireless signal positioning model includes: an input layer, a hidden layer, and an output layer;

[0075] Use the signal strength parameter received by the terminal base station as the input of the input layer, represented by R, where Ri represents the signal strength received by base station i from the terminal and is used as the input of the i-th neuron in the input layer, i = (1,..., i).

[0076] Vih represents the weight from the i-th neuron in the input layer to the h-th neuron in the hidden layer, where h = 1,..., H, and αh is the input of the h-th neuron in the hidden layer. Then the input formula of the hidden layer includes:

[0077]

[0078] The activation function of the hidden layer selects the Sigmod function:

[0079] Use γh to represent the threshold of the h-th neuron in the hidden layer, and bh to represent the output of the h-th neuron in the hidden layer. Then the output formula of the hidden layer includes:

[0080]

[0081] Let Whj represent the weight from the h-th neuron in the hidden layer to the j-th neuron in the output layer, and let βj be the input to the j-th neuron in the output layer. Then the input formula for the output layer is as follows:

[0082]

[0083] Let X1 and X2 represent the outputs of the output layer respectively, and θ j be the threshold of the j-th neuron in the output layer.

[0084] The activation function of the output layer is selected as the purelin function: F2(x) = x;

[0085] Then the training result is: Xj = F2(βj - θ j ) = βj - θj.

[0086] Specifically, in this embodiment, the self-learning signal positioning model to be trained further includes: obtaining the initial values of the weight parameters Vi h and Wh j as well as the threshold parameters γ h and θ j . It should be noted that the initialization of the weights mainly adopts the Xavier initialization method, and the initial values are uniformly distributed values with a variance of , where n represents the number of inputs.

[0087] In the above embodiment, the weight Vi h first takes a random number in [-1, 1]. Set the number of input neurons and the number of output neurons to 3, and set the derivative of the activation function SigMod near 0 to be 0.25. Since the initial value is a uniformly distributed value of the variance of , the weight Vi h is uniformly distributed in the range [-4, 4].

[0088] The weight Wh j , set the number of input neurons to 3, the number of output neurons to 2, and the derivative of the activation function purelin near 0 to be 1. Since the initial value is a uniformly distributed value with a variance of , the weight Wh j is uniformly distributed in the range [-1.095, 1.095].

[0089] The threshold θ j , (where j = 1,..., J), takes its initial value as a random number between [-1, 1].

[0090] The threshold γ h(h＝1，…，h) The formula for obtaining the initial value is as follows:

[0091]

[0092] Among them, M is a random number between [-1, 1].

[0093] Combine Figure 2 , S130: If the training result does not exceed the training error range, then a trained self-learning wireless signal positioning model is obtained.

[0094] S131: Use any conventional signal strength positioning algorithm to calculate the position of the steady-state signal data to obtain a training coordinate comparison set;

[0095] In this application, a signal strength positioning algorithm can be selected to calculate the position of the steady-state signal data. By using the signal strength of the terminal to be detected received by the terminal base station, the distances from the terminal to the nearest three terminal base stations are calculated respectively using the wireless signal strength and distance formula, and then the trilateration method is used to calculate the position coordinates of the terminal to be located. That is, with these three terminal base stations as the centers and the distances from the terminal to be located to the terminal base stations as the radii, draw circles. The intersection of the three circles is the position where the terminal to be located is located.

[0096] Among them, the intensity and distance calculation formula is as follows:

[0097]

[0098] d is the distance from the terminal to the base station, d0 is the reference distance, usually taken as 1 meter; P T is the transmission power; P L(d0) is the signal strength received by the base station when the terminal is at the reference distance; P L(d0) is the signal strength received by the base station when the terminal is at a distance d from the base station; η is the path loss exponent, usually taken as 2 - 4.

[0099] S132: Calculate whether the training error between the training coordinate result set and the training coordinate comparison set exceeds the training error range;

[0100] Specifically, in this embodiment, any training sample K in the training sample set is extracted, and the training coordinate comparison set measured by the terminal base station in a conventional manner The training result is The signal strength value of the training sample is, Then the error value is expressed by the least squares method as the following formula:

[0101]

[0102] S133: If it exceeds, then retrain the self-learning wireless signal positioning model to be trained until the training error is within the training error range;

[0103] Specifically, in this embodiment, an error value corresponding to the training error is extracted; a correction value is obtained by performing correction calculation on the error value according to a preset learning rate; according to E k and by setting a learning rate η between 0.01 and 0.1, the correction values of the weights and thresholds of each self-learning signal positioning model to be trained are calculated by backtracking. The formulas for calculating the correction values of each weight sum and threshold are as follows:

[0104] ΔWhj is the correction value of Whj,

[0105] Δθj is the correction value of θj,

[0106] Specifically, in this embodiment, the activation function of the hidden layer is selected as the Sigmod function, then F1’(x) = F1(x)(1 - F1(x));

[0107] wherein, let

[0108] ΔVih is the correction value of Vih,

[0109] Δγh is the correction value of γh, Δγh = η * eh.

[0110] Furthermore, the self-learning wireless signal positioning model to be trained is corrected using the calculated correction values to obtain a corrected self-learning wireless signal positioning model to be trained; the corrected self-learning wireless signal positioning model is retrained using the corrected weights and thresholds until the corresponding error value E k is within the training error range; stop correcting the self-learning wireless signal positioning model to be trained.

[0111] S134: If it does not exceed, then a trained self-learning wireless signal positioning model is obtained.

[0112] S140: Calculate the coordinate position of the terminal to be located using the self-learning wireless signal positioning model, and output the terminal coordinates

[0113] Specifically, in this embodiment, the signal parameter set of the terminal to be located is calculated for its position using any conventional signal strength positioning algorithm to obtain a first positioning coordinate; when the terminal to be located is in a steady state representation, the first positioning coordinate is output as the terminal coordinates of the terminal to be located.

[0114] Further, in this embodiment, it also includes positioning a to-be-positioned terminal in an unsteady state. By invoking the trained self-learning wireless signal positioning model to calculate the position of the wireless signal parameter set of the to-be-positioned terminal, a second positioning coordinate is obtained; weights corresponding to the first positioning coordinate and the second positioning coordinate are assigned for coordinate calculation.

[0115] Specifically, it is determined whether the self-learning wireless signal positioning model in the current to-be-identified area is in an available state. If it is in an available state, a low weight is assigned to the first positioning coordinate and a high weight is assigned to the second positioning coordinate. The weights are used to calculate the final coordinate of the first positioning coordinate and the second positioning coordinate to obtain a third positioning coordinate, and the third positioning coordinate is output as the terminal coordinate of the to-be-positioned terminal.

[0116] If it is in an unavailable state, a high weight is assigned to the first positioning coordinate and a low weight is assigned to the second positioning coordinate. The weights are used to calculate the final coordinate of the first positioning coordinate and the second positioning coordinate to obtain a third positioning coordinate, and the third positioning coordinate is output as the terminal coordinate of the to-be-positioned terminal.

[0117] Compared with the prior art, the embodiments of the present application mainly have the following beneficial effects: By deploying a certain number of terminal base stations in the area, the present application uses the terminal base stations to determine whether the signal parameters of the to-be-positioned terminal are in a stable state, and uses the signal parameters in a stable state to continuously train the to-be-trained self-learning signal positioning model. Through the above method, the cumbersome work of generating training samples by manual measurement is replaced, the deployment cost is reduced, and at the same time, the to-be-trained self-learning signal positioning model is further corrected and trained by correction values. The more self-learning training data accumulated according to the above method can greatly improve the positioning accuracy and better adapt to the laws of environmental changes and wireless signal fluctuations.

[0118] Those of ordinary skill in the art can understand that all or part of the processes of implementing the methods in the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a magnetic disk, an optical disk, a read-only memory (ROM), etc., or a random access memory (RAM), etc.

[0119] It should be understood that although the steps in the flowchart of the accompanying drawings are shown sequentially in the direction of the arrows, these steps are not necessarily executed sequentially in the order indicated by the arrows. Unless otherwise clearly stated in this document, there is no strict order restriction for the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or sub-steps or stages of other steps.

[0120] Further referring to Figure 3 and as an implementation of the method shown above Figure 1 , this application provides an embodiment of a self-learning wireless signal positioning system 200. This system embodiment corresponds to the method embodiment shown in Figure 2 , and this system can be specifically applied to various electronic devices.

[0121] Motion detection and reporting module 201: It is used to measure the motion state of the terminal to be located and report it to the base station through wireless communication;

[0122] The motion detection and reporting module includes two parts: a motion monitoring sub-module and a wireless communication sub-module. The motion monitoring sub-module can be any device that realizes the monitoring of the motion state of the terminal. In this application, an acceleration sensor is used for illustration. When the terminal to be located is in a stationary state, the acceleration value measured by the acceleration sensor will be lower than a specific threshold; when the terminal to be located is in a motion state, the acceleration value measured by the acceleration sensor will be higher than a specific threshold. The motion detection and reporting module can directly report the measured acceleration value to the base station, or directly determine the motion state based on the acceleration value and only report the indication mark of whether it is in a stationary state to the base station.

[0123] Identification module 202: It is used to obtain the signal parameter set of the terminal to be located in the area to be identified, perform parameter identification, and obtain the state characterization parameters;

[0124] In this embodiment, the recognition module further includes a regional deployment sub-module. Before the recognition module obtains the signal parameter set of the terminal to be located in the area to be recognized and performs parameter recognition, it also includes obtaining the recognition area by using the regional deployment sub-module. First, the regional deployment sub-module obtains the position data sets of multiple deployed receiving terminals to construct the recognition area; based on the preset regional division rules, the recognition area is geometrically re-divided by using the position data sets to obtain multiple sub-areas; the signal strength of the signal parameters of the terminal to be located received in each sub-area is recognized to obtain a signal strength value set; finally, the sub-areas corresponding to the signal strength values that meet the conditions in the signal strength value set are extracted to form the area to be recognized.

[0125] Specifically, in this embodiment, the regional deployment sub-module constructs the recognition area through receiving terminals that can receive the signal parameters sent by the terminal to be located. The receiving terminals include terminal base stations. After deploying multiple terminal base stations, any form of position connection is performed on the position data of the terminal base stations to obtain a geometrically shaped recognition area. The recognition area is re-divided into the area to be recognized by the preset connection method of the terminal base stations. Among them, at least 3 terminal base stations are included in the divided area to be recognized. The multiple terminal base stations in the recognition area receive the signal parameters sent by the terminal to be located to form an initial signal parameter set, the initial signal parameter set is analyzed to obtain an initial strength value set, the receiving terminals that meet the conditions are extracted from the initial strength value set according to the preset signal strength interval, and the corresponding area to be recognized is extracted by using the receiving terminals that meet the conditions.

[0126] Furthermore, the recognition module performs parameter recognition on the received signal parameter set. First, according to the motion state data reported by the motion detection and reporting module, the motion state of the terminal to be located is determined; among them, the signal parameter set also includes motion state signal parameters. When the terminal to be located is in a stationary state within a preset time, stationary state data is generated. The recognition module determines that the terminal to be located is in a stationary state through the stationary state data, extracts the signal parameter set of the terminal to be located in the area to be recognized for probability accumulation calculation to obtain a probability accumulation value; if the probability accumulation value exceeds the preset acceptable value, the state characterization parameter performs a steady-state characterization on the signal strength of the terminal to be located to form steady-state signal data; if the probability accumulation value does not exceed the preset acceptable value and / or the terminal to be located is in a non-stationary state, the state characterization parameter performs a non-steady-state characterization on the signal strength of the terminal to be located to form non-steady-state signal data. It should be noted that when the recognition module does not receive the stationary state data sent by the motion monitoring sub-module, it is determined that the terminal to be located is in a motion state, and a non-steady-state characterization is also performed on the terminal to be located in the motion state.

[0127] Specifically, in this embodiment, first, the motion state of the terminal to be located is determined. When it is in a stationary state, the signal strength parameters of the terminal to be located are continuously received, including the mean value E and the standard deviation σ of the line signal strength. The occurrence probabilities of all wireless signal strengths between [E - σ, E + σ] in the accumulated data are extracted to obtain the probability accumulation value P. When the probability accumulation value P is greater than the acceptable value, it is considered that this terminal has entered a steady state, and then the signal parameters are included in the steady-state characterization data to form steady-state characterization signal data. In this embodiment, the optional acceptable value is 68%. The duration of the accumulated data that does not exceed the acceptable value is statistically analyzed. When the preset statistical duration is exceeded and the acceptable value is not reached, the accumulated statistical data is cleared.

[0128] Specifically, in this embodiment, for the signal parameters that do not exceed the preset statistical duration and do not reach the acceptable value, and the signal parameters sent by the terminal to be located in the motion state, non-steady-state characterization data is embedded to form non-steady-state characterization signal data.

[0129] The positioning training module 203: is used to train a pre-constructed self-learning positioning model to be trained by using the state characterization parameters and the signal parameter set, and obtain a training result;

[0130] Specifically, in this embodiment, the positioning training module receives the signal strength set of the signal parameter set of the terminal to be located through the terminal base station, and uses the steady-state characterization signal data and the corresponding signal strength set as training samples to train the self-learning wireless signal positioning model to be trained, and extracts the training result to obtain the training coordinate result set.

[0131] In another preferred embodiment, the positioning training module takes the BP neural network as an example to illustrate the learning and training of the self-learning wireless signal positioning model to be trained. The self-learning wireless signal positioning model to be trained includes: an input layer, a hidden layer, and an output layer;

[0132] The signal strength parameters received by the terminal base station are used as the input of the input layer, represented by R, where Ri represents the signal strength received by base station i from the terminal and is used as the input of the i-th neuron in the input layer, i = (1,..., i).

[0133] Vih represents the weight from the i-th neuron in the input layer to the h-th neuron in the hidden layer, where h = 1,..., H. αh is the input of the h-th neuron in the hidden layer, and the input formula of the hidden layer includes:

[0134]

[0135] The activation function of the hidden layer selects the Sigmod function:

[0136] Using γh to represent the threshold of the h-th neuron in the hidden layer, and bh to represent the output of the h-th neuron in the hidden layer, the output formula of the hidden layer includes:

[0137]

[0138] Using Whj to represent the weight from the h-th neuron in the hidden layer to the j-th neuron in the output layer, and using βj as the input of the j-th neuron in the output layer, the input formula of the output layer includes:

[0139]

[0140] Using X1 and X2 to represent the outputs of the output layer respectively, and θ j as the threshold of the j-th neuron in the output layer,

[0141] The activation function of the output layer selects the purelin function: F2(x) = x;

[0142] Then the training result is obtained as: Xj = F2(βj - θ j ) = βj - θj.

[0143] Specifically, in this embodiment, the self-learning signal positioning model to be trained further includes: obtaining the initial values of the weight parameters Vi h and Wh j as well as the threshold parameters γ h and θ j . It should be noted that the initialization of the weights mainly adopts the Xavier initialization method, and the initial values are uniformly distributed values with a variance of , where n represents the number of inputs.

[0144] In the above embodiment, the weight Vi h first takes a random number in [-1, 1], sets the number of input neurons and the number of output neurons to 3, and sets the derivative of the activation function SigMod near 0 to 0.25. Since the initial value is a uniformly distributed value of the variance of , the weight Vi h is uniformly distributed in the range [-4, 4].

[0145] The weight Wh j , sets the number of input neurons to 3, the number of output neurons to 2, and the derivative of the activation function purelin near 0 to 1. Since the initial value is a uniformly distributed value with a variance of , the weight Wh j is uniformly distributed in the range [-1.095, 1.095].

[0146] The threshold θ j, (where \(j = 1,\ldots,J\)), take its initial value as a random number between \([-1,1]\).

[0147] Threshold \(\gamma\) h(h＝1，…，h) The formula for obtaining the initial value is as follows:

[0148]

[0149] where \(M\) is a random number between \([-1,1]\).

[0150] Correction module 204: used to obtain a trained self - learning wireless signal positioning model if the training result does not exceed the training error range.

[0151] Specifically, in this embodiment, any conventional signal strength positioning algorithm is used to calculate the position of the steady - state signal data to obtain a training coordinate comparison set; in this application, a signal strength positioning algorithm can be selected to calculate the position of the steady - state signal data. By using the signal strength of the terminal to be detected received by the terminal base station, the distances from the terminal to the three nearest terminal base stations are calculated respectively using the wireless signal strength and distance formula, and then the trilateration method is used to calculate the position coordinates of the terminal to be located. That is, with these three terminal base stations as the centers and the distances from the terminal to be located to the terminal base stations as the radii, draw circles. The intersection of the three circles is the position where the terminal to be located is located. Among them, the intensity - distance calculation formula is as follows:

[0152]

[0153] \(d\) is the distance from the terminal to the base station, \(d_0\) is the reference distance, usually taken as 1 meter; \(P\) T is the transmission power; \(P\) L(d0) is the signal strength received by the base station when the terminal is at the reference distance; \(P\) L(d0) is the signal strength received by the base station when the terminal is at a distance \(d\) from the base station; \(\eta\) is the path loss exponent, usually taken as 2 - 4.

[0154] Furthermore, the correction module calculates whether the training error between the training coordinate result set and the training coordinate comparison set exceeds the training error range.

[0155] Specifically, in this embodiment, any training sample \(K\) in the training sample set is extracted, and the training coordinate comparison set measured by the terminal base station in a conventional manner The training result is The signal strength value of the training sample is, Then the error value is expressed by the least - squares method as the following formula:

[0156]

[0157] Further, in this embodiment, if the error exceeds the pre-set error range, the self-learning wireless signal positioning model to be trained is retrained until the training error is within the training error range.

[0158] Specifically, the correction module extracts the error value corresponding to the training error; uses the error value to perform a correction calculation according to the pre-set learning rate to obtain a correction value; according to E k And by setting a learning rate η between 0.01 - 0.1, the correction values of the weights and thresholds of each self-learning signal positioning model to be trained are calculated by backpropagation. The formulas for calculating the correction values of each weight sum and threshold are as follows:

[0159] ΔWhj is the correction value of Whj,

[0160] Δθj is the correction value of θj,

[0161] Specifically, in this embodiment, the activation function of the hidden layer is selected as the Sigmod function, then F1’(x) = F1(x)(1 - F1(x));

[0162] Among them, let

[0163] ΔVih is the correction value of Vih,

[0164] Δγh is the correction value of γh, Δγh = η * eh.

[0165] Further, the correction module uses the obtained correction values to correct the self-learning wireless signal positioning model to be trained, and obtains the corrected self-learning wireless signal positioning model; uses the corrected weights and thresholds to retrain the corrected self-learning wireless signal positioning model until the corresponding error value E k is within the training error range; stop correcting the self-learning wireless signal positioning model to be trained.

[0166] Further, if it does not exceed the pre-set error range, a trained self-learning wireless signal positioning model is obtained.

[0167] Positioning integration module 205: used to calculate the coordinate position of the terminal to be located by using the self-learning wireless signal positioning model, and output the terminal coordinates.

[0168] Specifically, in this embodiment, the positioning integration module uses any conventional signal strength positioning algorithm to calculate the position of the signal parameter set of the terminal to be located, and obtains the first positioning coordinate;

[0169] When the terminal to be located is in a steady state, output the first positioning coordinate as the terminal coordinate of the terminal to be located.

[0170] Further, the positioning integration module further includes positioning the terminal to be located in a non-steady state. By invoking the trained self-learning wireless signal positioning model to calculate the position of the wireless signal parameter set of the terminal to be located, a second positioning coordinate is obtained; weights corresponding to the first positioning coordinate and the second positioning coordinate are assigned to perform coordinate calculation.

[0171] Specifically, the positioning integration module determines whether the self-learning wireless signal positioning model in the current area to be recognized is in an available state. If it is in an available state, assign a low weight to the first positioning coordinate and a high weight to the second positioning coordinate, and use the weights to calculate the final coordinates of the first positioning coordinate and the second positioning coordinate to obtain a third positioning coordinate, and output the third positioning coordinate as the terminal coordinate of the terminal to be located;

[0172] If it is in an unavailable state, assign a high weight to the first positioning coordinate and a low weight to the second positioning coordinate, and use the weights to calculate the final coordinates of the first positioning coordinate and the second positioning coordinate to obtain a third positioning coordinate, and output the third positioning coordinate as the terminal coordinate of the terminal to be located.

[0173] Compared with the prior art, the embodiments of the present application mainly have the following beneficial effects: By deploying a certain number of terminal base stations in the area, the embodiments of the present application use the terminal base stations to determine whether the signal parameters of the terminal to be located are in a stable state, and use the signal parameters in the stable state to continuously train the self-learning signal positioning model to be trained. Through the above method, the cumbersome work of generating training samples through manual measurement is replaced, the deployment cost is reduced, and at the same time, the self-learning signal positioning model to be trained is further corrected and trained through correction values. The more self-learning training data accumulated according to the above method can greatly improve the positioning accuracy and better adapt to the laws of environmental changes and wireless signal fluctuations.

[0174] To solve the above technical problems, the embodiments of the present application also provide a computer device. For details, please refer to Figure 4 , Figure 4 which is the basic structural block diagram of the computer device in this embodiment.

[0175] The computer device 4 includes a memory 31, a processor 32, and a network interface 33 that are communicatively connected to each other via a system bus. It should be noted that only the computer device 3 with components 31-33 is shown in the figure, but it should be understood that it is not required to implement all the shown components, and more or fewer components can be implemented alternatively. Among them, those skilled in the art of the present technology can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.

[0176] The computer device can be a computing device such as a desktop computer, a notebook, a palm computer, and a cloud server. The computer device can perform human-computer interaction with the user through a keyboard, a mouse, a remote control, a touchpad, a voice control device, etc.

[0177] The memory 31 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (such as SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disc, etc. In some embodiments, the memory 31 can be an internal storage unit of the computer device 3, such as the hard disk or memory of the computer device 3. In other embodiments, the memory 31 can also be an external storage device of the computer device 3, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the computer device 3. Of course, the memory 31 can also include both the internal storage unit and the external storage device of the computer device 3. In this embodiment, the memory 31 is generally used to store the operating system and various application software installed on the computer device 3, such as the program code of the X method. In addition, the memory 31 can also be used to temporarily store various data that have been output or will be output.

[0178] In some embodiments, the processor 32 may be a Central Processing Unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips. The processor 32 is generally used to control the overall operation of the computer device 3. In this embodiment, the processor 32 is used to run the program code stored in the memory 31 or process data, such as running the program code of the X method.

[0179] The network interface 33 may include a wireless network interface or a wired network interface, and this network interface 33 is generally used to establish a communication connection between the computer device 3 and other electronic devices.

[0180] The present application also provides another implementation manner, that is, to provide a computer-readable storage medium storing the online platform transformation program, and the online platform transformation program can be executed by at least one processor, so that the at least one processor executes the steps of the self-learning wireless signal positioning method as described above.

[0181] Through the description of the above embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus the necessary general hardware online platform. Of course, it can also be implemented by hardware, but in many cases the former is a better implementation manner. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disc), and includes several instructions to enable a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in various embodiments of the present application.

[0182] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld devices or portable devices, tablet devices, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, small computers, large computers, distributed computing environments including any of the above systems or devices, and so on. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in a distributed computing environment where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0183] Obviously, the embodiments described above are only a part of the embodiments of this application, rather than all of them. The preferred embodiments of this application are shown in the accompanying drawings, but they do not limit the patent scope of this application. This application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosed content of this application more thorough and comprehensive. Although this application has been described in detail with reference to the foregoing embodiments, for those skilled in the art, they can still modify the technical solutions described in the foregoing specific embodiments, or perform equivalent replacements on some of the technical features. Any equivalent structure made by using the content of the specification and drawings of this application, directly or indirectly applied in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. A self-learning wireless signal positioning method, characterized in that The method includes: Obtaining a periodic signal parameter set of a to-be-located terminal within a to-be-identified area for parameter identification to obtain a state characterization parameter, where the state characterization parameter is used to characterize the motion state of the to-be-located terminal; Extracting the signal parameter set corresponding to when the state characterization parameter indicates that the to-be-located terminal has a steady-state characterization, and automatically and continuously generating a training signal parameter set that can be used for self-learning training by using the signal parameter set; Training a pre-constructed to-be-trained self-learning positioning model by using the training signal parameter set to obtain a training coordinate result set; If the training error calculated through the training coordinate result set does not exceed the training error range, a trained self-learning wireless signal positioning model is obtained; Calculating the position of the signal parameter set of the to-be-located terminal by using any conventional signal strength positioning algorithm to obtain a first positioning coordinate; When the to-be-located terminal has a steady-state characterization, outputting the first positioning coordinate as the terminal coordinate; And / or invoking the trained self-learning wireless signal positioning model to calculate the position of the wireless signal parameter set of the to-be-located terminal to obtain a second positioning coordinate; Assigning weights corresponding to the first positioning coordinate and the second positioning coordinate for coordinate calculation to obtain a third positioning coordinate; When the to-be-located terminal has a non-steady-state characterization, outputting the third positioning coordinate as the terminal coordinate.

2. The self-learning wireless signal positioning method according to claim 1, wherein Before obtaining the signal parameter set of the to-be-located terminal within the to-be-identified area for parameter identification, it further includes: Obtaining a position data set of multiple deployed receiving terminals to construct an identification area; Based on a preset area division rule, geometrically re-dividing the identification area by using the position data set to obtain multiple sub-areas; Performing signal strength identification on the signal parameters of the to-be-located terminal received in each sub-area to obtain a signal strength value set; Extracting the sub-areas corresponding to the signal strength values that meet the conditions in the signal strength value set to form a to-be-identified area.

3. The self-learning wireless signal positioning method according to claim 2, wherein The obtaining a periodic signal parameter set of the to-be-located terminal within the to-be-identified area for parameter identification to obtain a state characterization parameter, where the state characterization parameter is used to characterize the motion state of the to-be-located terminal, specifically includes: Judging whether the to-be-located terminal within the to-be-identified area is in a stationary state according to the motion state signal parameter in the signal parameter set; When the to-be-located terminal is in a stationary state, extracting the signal parameter set of the to-be-located terminal within the to-be-identified area for probability accumulation calculation to obtain a probability accumulation value; If the probability accumulation value exceeds a preset acceptable value, the state characterization parameter performs a steady-state characterization on the signal strength of the to-be-located terminal to form steady-state signal data; If the probability accumulation value does not exceed the preset acceptable value and / or when the to-be-located terminal is in a non-stationary state, the state characterization parameter performs a non-steady-state characterization on the signal strength of the to-be-located terminal to form non-steady-state signal data.

4. The self-learning wireless signal positioning method according to claim 3, wherein The if the training error calculated through the training coordinate result set does not exceed the training error range, a trained self-learning wireless signal positioning model is obtained, specifically includes: Calculate the position of the steady-state signal data using any conventional signal strength positioning algorithm to obtain a training coordinate comparison set; Calculate whether the training error between the training coordinate result set and the training coordinate comparison set exceeds the training error range; If it exceeds, retrain the to-be-trained self-learning wireless signal positioning model until the training error is within the training error range; If it does not exceed, obtain the trained self-learning wireless signal positioning model.

5. The self-learning wireless signal positioning method according to claim 4, characterized in that The step of retraining the to-be-trained self-learning wireless signal positioning model until the training error is within the training error range specifically includes: Extract the error value corresponding to the training error; Use the error value to perform correction calculation according to a preset learning rate to obtain a correction value; Use the correction value to correct the to-be-trained self-learning wireless signal positioning model to obtain the corrected to-be-trained self-learning wireless signal positioning model.

6. The self-learning wireless signal positioning method according to claim 5, wherein, After obtaining the corrected to-be-trained self-learning wireless signal positioning model, it further includes: Retrain the corrected to-be-trained self-learning wireless signal positioning model until the corresponding error value is within the training error range; Stop correcting the to-be-trained self-learning wireless signal positioning model.

7. A self-learning wireless signal positioning system, characterized in that, The system includes: A motion detection and reporting module, configured to measure the motion state of the positioning terminal and report it to the base station via wireless communication; An identification module, configured to obtain a signal parameter set of a to-be-positioned terminal in a to-be-identified area for parameter identification to obtain a state characterization parameter; A positioning training module, configured to extract a signal parameter set corresponding to when the state characterization parameter indicates that the to-be-positioned terminal has a steady-state characterization to train a pre-constructed to-be-trained self-learning positioning model to obtain a training coordinate result set; A correction module, configured to obtain the trained self-learning wireless signal positioning model if the training error calculated through the training coordinate result set does not exceed the training error range; A positioning integration module, configured to calculate the position of the signal parameter set of the to-be-positioned terminal using any conventional signal strength positioning algorithm to obtain a first positioning coordinate; is also used to output the first positioning coordinate as the terminal coordinate when the to-be-positioned terminal has a steady-state characterization; is also used to call the trained self-learning wireless signal positioning model to calculate the position of the wireless signal parameter set of the to-be-positioned terminal to obtain a second positioning coordinate; is also used to assign weights corresponding to the first positioning coordinate and the second positioning coordinate for coordinate calculation to obtain a third positioning coordinate; is also used to output the third positioning coordinate as the terminal coordinate when the to-be-positioned terminal has a non-steady-state characterization.

8. A computer device, including a memory and a processor, where computer-readable instructions are stored in the memory, and when the processor executes the computer-readable instructions, the steps of the self-learning wireless signal positioning method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium, characterized in that, Computer-readable instructions are stored on the computer-readable storage medium, and when the computer-readable instructions are executed by the processor, the steps of the self-learning wireless signal positioning method according to any one of claims 1 to 6 are implemented.

Images