Mobile signal terminal sequence analyzing and positioning method and system based on LSTM (Long Short Term Memory) model

Abstract

The invention discloses a mobile signal terminal sequence analyzing and positioning method and system based on an LSTM (Long Short Term Memory) model, and particularly relates to the field of mobile communication, and the method comprises the steps of S1, WiFi probe data acquisition, S2, data preprocessing, S3, time sequence conversion, S4, LSTM model construction, S5, positioning algorithm distance analysis, S6, equipment position analysis and S7, man-machine interaction. According to the method, the mapping relation between the RSSI value and the probe distance is established according to the SVM model, compared with a traditional logarithmic distance path loss model, the method can adjust the model according to the environment and has higher accuracy for distance judgment, meanwhile, the support vector of the method has a large influence on the classification or regression result of the model, and the accuracy of distance judgment is improved. Therefore, in the SVM training process, only the support vector needs to be considered, all training data points do not need to be considered, the calculation efficiency of the model is improved, and when the model is affected by noise or abnormal values, the model can be more stable in prediction of unknown samples through selection of the support vector.

Description

Technical Field

[0001] The present invention relates to the field of mobile communication technology, and more specifically, to a mobile signal terminal sequence analysis and positioning method and system based on an LSTM model. Background Art

[0002] The LSTM model is a special recurrent neural network (RNN) with powerful sequence data processing capabilities. It can capture long-term dependencies in sequence data and is suitable for processing time series data. In the mobile signal terminal sequence analysis and positioning method, the LSTM model can learn the historical data of the mobile signal terminal and predict future location information, thereby achieving accurate positioning.

[0003] Location service technology is an important application direction of the mobile signal terminal sequence analysis positioning method. With the popularization of smart phones and the development of mobile Internet, location services have become an indispensable part of people's daily life. The mobile signal terminal sequence analysis positioning method based on the LSTM model can provide more accurate and reliable positioning information for location services and improve user experience.

[0004] However, it still has some shortcomings in actual use. For example, the models used in the existing technology for positioning terminal devices are mostly logarithmic distance path loss models, which are a commonly used model to describe the relationship between signal propagation loss and distance. The versatility of this model also leads to its shortcomings of being not accurate enough in exploring the quantitative relationship between signal strength and distance. At the same time, the existing positioning system algorithm has a high computational complexity and requires a large amount of computing resources when faced with large data input. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a mobile signal terminal sequence analysis and positioning method and system based on an LSTM model, and solves the problems raised in the above-mentioned background technology through the following scheme.

[0006] To achieve the above object, the present invention provides the following technical solutions: a method and system for analyzing and locating a mobile signal terminal sequence based on an LSTM model, comprising: S1: WiFi probe data collection: collecting device signal feature data text through a WiFi probe; S2: Data preprocessing: performing a preprocessing operation on the collected device signal feature data text to obtain a first data text; S3: Time series conversion: Arrange the first data text in chronological order to form a time series data text, and the RSSI value at each time point is associated with the corresponding MAC address.

[0007] S4: LSTM model construction: Build an LSTM model and input the time series data text into the model to obtain the LSTM text output value; S5: Positioning algorithm distance analysis: used to establish the mapping relationship between RSSI value and position according to the SVM model and calculate the distance text output value of the LSTM model text output value; S6: device location analysis: determining the location of the device according to the distance text output value, the first data text, and the longitude and latitude information of the WiFi probe to obtain a time series location text output value; S7: Human-computer interaction: used to import the time series position text output value into the user information terminal.

[0008] Preferably, the device signal characteristic data text includes the device MAC address and signal strength. The device MAC address is a media access control address, also known as a physical address or hardware address. It is a unique identifier used by network devices when communicating at the network layer. RSSI is a measurement unit that represents the strength of a received electromagnetic wave signal.

[0009] Preferably, the preprocessing operation is as follows: first, verify the format and validity of the MAC address, remove duplicate MAC addresses and MAC addresses that do not comply with the IEEE 802 standard, then use an anomaly detection algorithm to perform noise filtering and outlier processing on the RSSI value, and finally use the MAC address as a unique identifier and associate it with the corresponding RSSI value to construct a first data text containing the MAC address and RSSI value as features.

[0010] Preferably, the LSTM model specifically includes: Forget gate formula: , in, Refers to the output of the forget gate at the current moment, Refers to the Sigmoid function, The weight matrix is ​​the input of the forget gate The linear transformation of The weight matrix represents the hidden state of the forget gate at the previous moment The linear transformation of Refers to the input at the current moment, Refers to the hidden state at the previous moment, Refers to the bias vector used for linear transformation of the forget gate; Input gate formula: , in, Refers to the output of the input gate at the current moment, Refers to the Sigmoid function, The weight matrix is ​​the input gate to the input The linear transformation of The weight matrix represents the hidden state of the input gate at the previous moment The linear transformation of Refers to the input at the current moment, Refers to the hidden state at the previous moment, Refers to the bias vector used for linear transformation of the input gate; Output gate formula: , in, Refers to the output of the output gate at the current moment, Refers to the Sigmoid function, The weight matrix is ​​the input gate to the input The linear transformation of The weight matrix represents the hidden state of the output gate at the previous moment The linear transformation of Refers to the input at the current moment, Refers to the hidden state at the previous moment. Refers to the bias vector used for linear transformation of the output gate; Candidate memory cell state formula: , in, It can also be expressed as refers to the state of the candidate memory unit at the current moment, refers to the hyperbolic tangent function, The weight matrix is ​​the candidate memory cell state input The linear transformation of The weight matrix represents the candidate memory unit state to the hidden state at the previous moment The linear transformation of Refers to the input at the current moment, Refers to the hidden state at the previous moment, Refers to the bias vector used for linear transformation of the candidate memory cell state; Memory unit state update formula: , in, Refers to the current state of the memory unit, Refers to the output of the forget gate at the current moment, Refers to the state of the memory unit at the previous moment, Refers to the output of the input gate at the current moment, Refers to the candidate memory unit state at the current moment, ⊙ refers to the element-by-element multiplication operation; Hidden state update formula: , in Refers to the hidden state at the current moment, Refers to the output of the output gate at the current moment, refers to the hyperbolic tangent function, Refers to the state of the memory unit at the current moment, and ⊙ refers to the element-by-element multiplication operation.

[0011] Preferably, the specific method for the SVM model to establish the mapping relationship between the RSSI value and the position is as follows: Reference points are selected in the area around the WiFi probe. These reference points should be evenly distributed to cover the entire area. At each reference point, a mobile smart terminal is used to receive RSSI values ​​from different wireless access points. The distance of each reference point from the probe and the corresponding RSSI value are recorded. The collected RSSI values ​​are filtered and denoised, missing values ​​or abnormal values ​​are filled or deleted, and the collected RSSI values ​​are vectorized to obtain a total data set and divide the total data set into a training set and a test set. The radial basis function is selected as the kernel function of the model, and the training set data is used to train the SVM model to obtain the optimal hyperplane and support vector. The parameters of the model are adjusted by cross-validation and other methods. The accuracy and recall rate are calculated using the test set data to evaluate the performance of the model. If the performance of the model is not ideal, the performance of the evaluation model is met by adjusting the kernel function, parameters or re-collecting data, that is, the model accuracy and recall rate are higher than the preset values.

[0012] Preferably, the method for parsing the location of the device is as follows: The device identity is identified according to the MAC address in the first data text, and the specific location information of the WiFi probe is determined by the longitude and latitude information of the WiFi probe. Finally, the location information of the device is determined by finding the intersection through the specific location information and distance information of the three WiFi probes that detect the device, and finally the time series position text output value is obtained according to the time series.

[0013] Preferably, a mobile signal terminal sequence analysis and positioning system based on the LSTM model includes: a system central processor module, a system operation database and a user information terminal, and also includes a WiFi probe data acquisition module, a data preprocessing module, a time series conversion module, an LSTM model construction module, a positioning algorithm distance analysis module, a device location analysis module and a human-computer interaction module; The system operation database includes all data texts of the mobile signal terminal sequence analysis and positioning system based on the LSTM model, and collects the information text output by each module in real time. The system central processor module is used to centrally control the information text instructions output by each module. The user information terminal is an information output device that receives the mobile signal terminal sequence analysis and positioning system based on the LSTM model; WiFi probe data collection module: collects device signal feature data text through WiFi probes, and outputs the device signal feature data text to the data preprocessing module; Data preprocessing module: performs preprocessing operation on the collected device signal feature data text to obtain a first data text, and outputs the first data text to the time series conversion module; Time series conversion module: arranges the first data text in chronological order to form a time series data text, the RSSI value at each time point is associated with the corresponding MAC address, and outputs the time series data text to the LSTM model construction module.

[0014] LSTM model building module: builds the LSTM model and inputs the time series data text into the model to obtain the LSTM text output value, and outputs the LSTM text output value to the positioning algorithm distance parsing module; Positioning algorithm distance parsing module: used to establish the mapping relationship between RSSI value and position according to the SVM model and calculate the distance text output value of the LSTM model text output value, and output the distance text output value to the device position parsing module; Device location parsing module: determines the location of the device according to the distance text output value, the first data text, and the longitude and latitude information of the WiFi probe to obtain a time series location text output value, and outputs the time series location text output value to the human-computer interaction module; Human-computer interaction module: used to import the time series position text output value into the user information terminal.

[0015] Technical effects and advantages of the present invention: The present invention establishes a mapping relationship between RSSI value and probe distance according to the SVM model. Compared with the traditional logarithmic distance path loss model, the method can adjust the model according to the environment, can be better applied to different scenarios, and has higher accuracy in distance judgment. At the same time, since the support vectors of the method are located near the decision boundary, they have a greater impact on the classification or regression results of the model. Therefore, in the training process of the SVM, only the support vectors need to be considered instead of all the training data points, which greatly improves the calculation efficiency of the model. At the same time, when the model is affected by noise or outliers, the selection of the support vectors can maintain the robustness of the model, making the model more stable in predicting unknown samples. The present invention establishes a mapping relationship between RSSI value and probe distance through the SVM model, and locates the device position according to multiple probes, thereby increasing the accuracy of positioning; BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic diagram of the overall structure of the present invention.

[0017] Figure 2 It is a schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to be used as limitations to the present application. As used in the specification of the present application, the singular expressions "one", "a", "said", "above", "the" and "this" are intended to also include plural expressions, unless there is a clear indication to the contrary in the context. It should also be understood that the term "and / or" used in the present application refers to and includes any or all possible combinations of one or more listed items.

[0020] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as suggesting or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, unless otherwise specified, "plurality" means two or more.

[0021] As attached Figure 1 A mobile signal terminal sequence analysis and positioning method based on an LSTM model is shown, comprising: S1: WiFi probe data collection: collect device signal feature data text through WiFi probe; The device signal characteristic data text includes the device MAC address and signal strength. The device MAC address (Media Access Control Address) is the media access control address, also known as the physical address or hardware address. It is a unique identifier used by network devices when communicating at the network layer. The MAC address is designed to uniquely identify each device in the physical network to ensure that the data frame (Frame) can be correctly sent to the destination device; RSSI is a measurement unit that represents the strength of the received electromagnetic wave signal, generally used in WiFi, Bluetooth and other wireless communication technologies; S2: Data preprocessing: performing a preprocessing operation on the collected device signal feature data text to obtain a first data text; The preprocessing operation is specifically as follows: first, verify the format and validity of the MAC address, remove duplicate MAC addresses and MAC addresses that do not comply with the IEEE 802 standard, then use an anomaly detection algorithm to perform noise filtering and outlier processing on the RSSI value to ensure the accuracy and reliability of the data, and finally use the MAC address as a unique identifier and associate it with the corresponding RSSI value to construct a first data text containing the MAC address and RSSI value as features; S3: Time series conversion: Arrange the first data text in chronological order to form a time series data text, and the RSSI value at each time point is associated with the corresponding MAC address.

[0022] S4: LSTM model construction: Build an LSTM model and input the time series data text into the model to obtain the LSTM text output value; The LSTM model specifically includes: Forget gate formula: , in, Refers to the output of the forget gate at the current moment, Refers to the Sigmoid function, The weight matrix is ​​the input of the forget gate The linear transformation of The weight matrix represents the hidden state of the forget gate at the previous moment The linear transformation of Refers to the input at the current moment, Refers to the hidden state at the previous moment, Refers to the bias vector used for linear transformation of the forget gate; Input gate formula: , in, Refers to the output of the input gate at the current moment, Refers to the Sigmoid function, The weight matrix is ​​the input gate to the input The linear transformation of The weight matrix represents the hidden state of the input gate at the previous moment The linear transformation of Refers to the input at the current moment, Refers to the hidden state at the previous moment, Refers to the bias vector used for linear transformation of the input gate; Output gate formula: , in, Refers to the output of the output gate at the current moment, Refers to the Sigmoid function, The weight matrix is ​​the input gate to the input The linear transformation of The weight matrix represents the hidden state of the output gate at the previous moment The linear transformation of Refers to the input at the current moment, Refers to the hidden state at the previous moment, Refers to the bias vector used for linear transformation of the output gate; Candidate memory cell state formula: , in, It can also be expressed as refers to the state of the candidate memory unit at the current moment, refers to the hyperbolic tangent function, The weight matrix is ​​the candidate memory cell state input The linear transformation of The weight matrix represents the candidate memory unit state to the hidden state at the previous moment The linear transformation of Refers to the input at the current moment, Refers to the hidden state at the previous moment, Refers to the bias vector used for linear transformation of the candidate memory cell state; Memory unit state update formula: , in, Refers to the current state of the memory unit, Refers to the output of the forget gate at the current moment, Refers to the state of the memory unit at the previous moment, Refers to the output of the input gate at the current moment, Refers to the candidate memory unit state at the current moment, ⊙ refers to the element-by-element multiplication operation; Hidden state update formula: , in Refers to the hidden state at the current moment, Refers to the output of the output gate at the current moment, refers to the hyperbolic tangent function, Refers to the current state of the memory unit, ⊙ refers to the element-by-element multiplication operation; S5: Positioning algorithm distance analysis: used to establish the mapping relationship between RSSI value and position according to the SVM model and calculate the distance text output value of the LSTM model text output value; What needs to be specifically explained in this embodiment is that the specific method for the SVM model to establish the mapping relationship between the RSSI value and the position is as follows: Select reference points in the area around the WiFi probe. These reference points should be evenly distributed to cover the entire area. At each reference point, use a mobile smart terminal to receive RSSI values ​​from different wireless access points, record the distance of each reference point from the probe and the corresponding RSSI value, filter and denoise the collected RSSI values, fill or delete missing values ​​or abnormal values, and vectorize the collected RSSI values ​​to obtain the total data set and divide the total data set into a training set and a test set. Select the radial basis function as the kernel function of the model, use the training set data to train the SVM model, obtain the optimal hyperplane and support vector, adjust the parameters of the model through cross-validation and other methods, use the test set data to calculate the accuracy and recall rate to evaluate the performance of the model. If the performance of the model is not ideal, adjust the kernel function, parameters or re-collect data to make the performance of the evaluation model meet the standard, that is, the model accuracy and recall rate are higher than the preset values; The optimal hyperplane specifically refers to the mapping relationship between the RSSI value and the probe distance, and the optimal hyperplane is used to distinguish the RSSI value ranges corresponding to different positions; The support vector specifically refers to the support vector corresponding to those RSSI value points that have a significant impact on the decision boundary in the SVM model of RSSI value and probe distance mapping; The preset values ​​refer to the warning values ​​of model accuracy and recall rate set according to different accuracy requirements of the model; S6: device location analysis: analyzing the location of the device according to the distance text output value, the first data text, and the longitude and latitude information of the WiFi probe to obtain a time series location text output value; What needs to be specifically explained in this embodiment is that the method for parsing the location of the device is as follows: Identify the device identity according to the MAC address in the first data text, and determine the specific location information of the WiFi probe through the longitude and latitude information of the WiFi probe, and finally find the intersection through the specific location information and distance information of the three WiFi probes that detect the device to determine the location information of the device, and finally output the time series location text value obtained according to the time series; S7: Human-computer interaction: used to import the time series position text output value into the user information terminal; What needs to be specifically explained in this embodiment is that the user information terminal is used to upload the time series position text output value of the same device to the user information terminal to help the user determine the device position at different times and the overall movement trajectory of the device.

[0023] As attached Figure 2A mobile signal terminal sequence analysis and positioning system based on an LSTM model is shown, comprising: a system central processor module, a system operation database and a user information terminal, and also comprising a WiFi probe data acquisition module, a data preprocessing module, a time series conversion module, an LSTM model construction module, a positioning algorithm distance analysis module, a device location analysis module and a human-computer interaction module; The system operation database includes all data texts of the mobile signal terminal sequence analysis and positioning system based on the LSTM model, and collects the information text output by each module in real time. The system central processor module is used to centrally control the information text instructions output by each module. The user information terminal is an information output device that receives the mobile signal terminal sequence analysis and positioning system based on the LSTM model; WiFi probe data collection module: collects device signal feature data text through WiFi probes, and outputs the device signal feature data text to the data preprocessing module; Data preprocessing module: performs preprocessing operation on the collected device signal feature data text to obtain a first data text, and outputs the first data text to the time series conversion module; Time series conversion module: arranges the first data text in chronological order to form a time series data text, the RSSI value at each time point is associated with the corresponding MAC address, and outputs the time series data text to the LSTM model construction module.

[0024] LSTM model building module: builds the LSTM model and inputs the time series data text into the model to obtain the LSTM text output value, and outputs the LSTM text output value to the positioning algorithm distance parsing module; Positioning algorithm distance parsing module: used to establish the mapping relationship between RSSI value and position according to the SVM model and calculate the distance text output value of the LSTM model text output value, and output the distance text output value to the device position parsing module; Device location parsing module: determines the location of the device according to the distance text output value, the first data text, and the longitude and latitude information of the WiFi probe to obtain a time series location text output value, and outputs the time series location text output value to the human-computer interaction module; Human-computer interaction module: used to import the time series position text output value into the user information terminal; Secondly: In the drawings of the embodiments disclosed in the present invention, only the structures related to the embodiments disclosed in the present invention are involved, and other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of the present invention can be combined with each other; Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A mobile signal terminal sequence analysis and positioning method based on LSTM model, characterized in that: include: S1: WiFi probe data collection: collect device signal feature data text through WiFi probe; S2: Data preprocessing: performing a preprocessing operation on the collected device signal feature data text to obtain a first data text; S3: Time series conversion: the first data text is arranged in time sequence to form a time series data text, and the RSSI value at each time point is associated with the corresponding MAC address; S4: LSTM model construction: Build an LSTM model and input the time series data text into the model to obtain the LSTM text output value; S5: Positioning algorithm distance analysis: used to establish the mapping relationship between RSSI value and position according to the SVM model and calculate the distance text output value of the LSTM model text output value; S6: device location analysis: determining the location of the device according to the distance text output value, the first data text, and the longitude and latitude information of the WiFi probe to obtain a time series location text output value; S7: Human-computer interaction: used to import the time series position text output value into the user information terminal.

2. According to a mobile signal terminal sequence analysis and positioning method based on LSTM model according to claim 1, it is characterized in that: The preprocessing operation is specifically as follows: first, verify the format and validity of the MAC address, remove duplicate MAC addresses and MAC addresses that do not comply with the IEEE 802 standard, then use an anomaly detection algorithm to perform noise filtering and outlier processing on the RSSI value, and finally use the MAC address as a unique identifier and associate it with the corresponding RSSI value to construct a first data text containing the MAC address and RSSI value as features.

3. According to claim 1, a mobile signal terminal sequence analysis and positioning method based on LSTM model is characterized in that: The device signal characteristic data text includes the device MAC address and signal strength. The device MAC address is the media access control address, also known as the physical address or hardware address. It is a unique identifier used by network devices when communicating at the network layer. RSSI is a measurement unit that represents the strength of the received electromagnetic wave signal.

4. According to a mobile signal terminal sequence analysis and positioning method based on LSTM model according to claim 1, it is characterized in that: The LSTM model specifically includes: Forget gate formula: ， in, Refers to the output of the forget gate at the current moment, Refers to the Sigmoid function, The weight matrix is ​​the input of the forget gate The linear transformation of The weight matrix represents the hidden state of the forget gate at the previous moment The linear transformation of Refers to the input at the current moment, Refers to the hidden state at the previous moment, Refers to the bias vector used for linear transformation of the forget gate; Input gate formula: ， in, Refers to the output of the input gate at the current moment, Refers to the Sigmoid function, The weight matrix is ​​the input gate to the input The linear transformation of The weight matrix represents the hidden state of the input gate at the previous moment The linear transformation of Refers to the input at the current moment, Refers to the hidden state at the previous moment. Refers to the bias vector used for linear transformation of the input gate; Output gate formula: ， in, Refers to the output of the output gate at the current moment, Refers to the Sigmoid function, The weight matrix is ​​the input gate to the input The linear transformation of The weight matrix represents the hidden state of the output gate at the previous moment The linear transformation of Refers to the input at the current moment, Refers to the hidden state at the previous moment, Refers to the bias vector used for linear transformation of the output gate; Candidate memory cell state formula: ， in, It can also be expressed as refers to the state of the candidate memory unit at the current moment, refers to the hyperbolic tangent function, The weight matrix is ​​the candidate memory cell state input The linear transformation of The weight matrix represents the candidate memory unit state to the hidden state at the previous moment The linear transformation of Refers to the input at the current moment, Refers to the hidden state at the previous moment. Refers to the bias vector used for linear transformation of the candidate memory cell state; Memory unit state update formula: ， in, Refers to the current state of the memory unit, Refers to the output of the forget gate at the current moment, Refers to the state of the memory unit at the previous moment, Refers to the output of the input gate at the current moment, Refers to the candidate memory unit state at the current moment, ⊙ refers to the element-by-element multiplication operation; Hidden state update formula: ， in Refers to the hidden state at the current moment, Refers to the output of the output gate at the current moment, refers to the hyperbolic tangent function, Refers to the state of the memory unit at the current moment, and ⊙ refers to the element-by-element multiplication operation.

5. The method for analyzing and locating a mobile signal terminal sequence based on an LSTM model according to claim 1, characterized in that: The specific method of establishing the mapping relationship between RSSI value and location by SVM model is as follows: Reference points are selected in the area around the WiFi probe. These reference points should be evenly distributed to cover the entire area. At each reference point, a mobile smart terminal is used to receive RSSI values ​​from different wireless access points. The distance of each reference point from the probe and the corresponding RSSI value are recorded. The collected RSSI values ​​are filtered and denoised, missing values ​​or abnormal values ​​are filled or deleted, and the collected RSSI values ​​are vectorized to obtain a total data set and divide the total data set into a training set and a test set. The radial basis function is selected as the kernel function of the model, and the training set data is used to train the SVM model to obtain the optimal hyperplane and support vector. The parameters of the model are adjusted by cross-validation and other methods. The accuracy and recall rate are calculated using the test set data to evaluate the performance of the model. If the performance of the model is not ideal, the performance of the evaluation model is met by adjusting the kernel function, parameters or re-collecting data, that is, the model accuracy and recall rate are higher than the preset values.

6. The method for analyzing and locating a mobile signal terminal sequence based on an LSTM model according to claim 1, characterized in that: The device location is parsed as follows: The device identity is identified according to the MAC address in the first data text, and the specific location information of the WiFi probe is determined by the longitude and latitude information of the WiFi probe. Finally, the location information of the device is determined by finding the intersection through the specific location information and distance information of the three WiFi probes that detect the device, and finally the time series position text output value is obtained according to the time series.

7. A mobile signal terminal sequence analysis and positioning system based on an LSTM model, used to implement a mobile signal terminal sequence analysis and positioning method based on an LSTM model as described in any one of claims 1 to 6, characterized in that: include: The system CPU module, system operation database and user information terminal also include WiFi probe data acquisition module, data preprocessing module, time series conversion module, LSTM model building module, positioning algorithm distance analysis module, device location analysis module and human-computer interaction module; The system operation database includes all data texts of the mobile signal terminal sequence analysis and positioning system based on the LSTM model, and collects the information text output by each module in real time. The system central processor module is used to centrally control the information text instructions output by each module. The user information terminal is an information output device that receives the mobile signal terminal sequence analysis and positioning system based on the LSTM model; WiFi probe data collection module: collects device signal feature data text through WiFi probes, and outputs the device signal feature data text to the data preprocessing module; Data preprocessing module: performs preprocessing operation on the collected device signal feature data text to obtain a first data text, and outputs the first data text to the time series conversion module; Time series conversion module: arranges the first data text in chronological order to form a time series data text, the RSSI value at each time point is associated with the corresponding MAC address, and outputs the time series data text to the LSTM model construction module; LSTM model building module: builds the LSTM model and inputs the time series data text into the model to obtain the LSTM text output value, and outputs the LSTM text output value to the positioning algorithm distance parsing module; Positioning algorithm distance parsing module: used to establish the mapping relationship between RSSI value and position according to the SVM model and calculate the distance text output value of the LSTM model text output value, and output the distance text output value to the device position parsing module; Device location parsing module: determines the location of the device according to the distance text output value, the first data text, and the longitude and latitude information of the WiFi probe to obtain a time series location text output value, and outputs the time series location text output value to the human-computer interaction module; Human-computer interaction module: used to import the time series position text output value into the user information terminal.

Images