Asset positioning monitoring method based on wi-fi and inertial navigation fusion and large language model

Abstract

The application belongs to the technical field of industrial asset protection and positioning, and discloses an asset positioning monitoring method based on Wi-Fi and inertial navigation fusion and a large language model, which comprises the following steps: acquiring multi-modal data and constructing a fingerprint library; through the fusion positioning technology of Wi-Fi and inertial navigation, combining the collected signal strength, the inertial navigation path and the feature data in the fingerprint library, the industrial asset is accurately positioned; the large language model is used to perform semantic fusion and analysis on the multi-modal data, identify potential risks, and generate a safety evaluation report. The asset positioning monitoring method based on Wi-Fi and inertial navigation fusion and the large language model can not only realize accurate tracking of the asset position, but also perceive the asset operation state, identify potential risks and provide real-time risk warning through the multi-modal data fusion technology and the abnormal detection algorithm.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of industrial asset protection and positioning, and particularly relates to an asset positioning monitoring method based on Wi-Fi and inertial navigation fusion and a large language model. BACKGROUND

[0002] In modern industrial environments, accurate asset positioning and protection are critical to ensuring production efficiency and safety. Traditional industrial asset positioning methods usually rely on a single technical means. These single technical methods each have some limitations and are difficult to cope with complex and dynamic industrial environments, especially in indoor or signal interference areas, where positioning accuracy often cannot meet actual needs. Wi-Fi and inertial navigation system combined with extended Kalman filter fusion can effectively compensate for their respective shortcomings and improve positioning accuracy and stability through multi-sensor data fusion. This method can improve positioning accuracy and enhance the system's ability to adapt to signal fluctuations and sensor errors.

[0003] Traditional industrial asset positioning methods mainly focus on obtaining asset location information, and the core goal is to provide basic location services for assets. However, these methods usually rely on a single technical means and have relatively single functions, only meeting the basic positioning needs of assets in static or regular dynamic environments. In complex industrial environments, such methods often cannot meet the high standard requirements of industrial asset protection and safety management. In addition, traditional positioning methods usually lack comprehensive perception and analysis capabilities for asset states and cannot timely identify potential abnormal conditions or provide effective risk warnings.

[0004] In the field of industrial asset protection, the positioning function has gradually shifted from "knowing the location" to "understanding the state" and "warning the risk". In modern industrial environments, the value of assets is not only reflected in their location data, but also in the monitoring of their operating state and the identification and prevention of risks. Therefore, it is urgent to develop a more intelligent industrial asset positioning monitoring method that can not only accurately track the location of assets, but also perceive the operating state of assets, identify potential risks, and provide real-time risk warnings through multi-modal data fusion technology and anomaly detection algorithms. SUMMARY

[0005] The purpose of the present application is to provide an asset positioning monitoring method based on Wi-Fi and inertial navigation fusion and a large language model to improve positioning accuracy and have anomaly detection and risk warning functions, achieving comprehensive intelligent protection of industrial assets.

[0006] To achieve the above-mentioned purpose, the present application provides an asset positioning monitoring method based on Wi-Fi and inertial navigation fusion and a large language model, comprising the following steps:

[0007] Step S1, acquiring multi-modal data to construct a fingerprint library; wherein the multi-modal data includes Wi-Fi signals, inertial navigation data, device logs, and image data;

[0008] Step S2, using a fusion positioning technology of Wi-Fi and inertial navigation, combined with the collected signal strength, inertial navigation path, and feature data in the fingerprint library, to accurately locate the industrial assets;

[0009] Step S3, using a large language model to perform semantic fusion and analysis on the multi-modal data, to identify potential risks and generate a safety assessment report containing safety ratings, key findings, and warning suggestions.

[0010] Preferably, in step S1, multi-modal data is acquired to construct a fingerprint library; wherein the multi-modal data includes Wi-Fi signals, inertial navigation data, device logs, and image data, and the specific process is as follows:

[0011] Step S11, using multi-modal data collection equipment to collect data D t from different sources as follows:

[0012]

[0013] wherein, represents Wi-Fi signal data, represents inertial navigation data, represents device logs and image data; the superscript t represents the sampling time;

[0014] Step S12, acquiring signal strength values D WiFi from multiple Wi-Fi access points in the environment as follows:

[0015] D WiFi = {rssi1, rssi2, …, rssi n};

[0016] wherein, D WiFi represents a set of signal strength values from n access points;

[0017] Step S13, acquiring inertial navigation data of the device through inertial sensors, specifically including linear acceleration D Linear , gravitational acceleration D Gravity , and rotation angle D Rotation as follows:

[0018] D Linear = {linear x , linear y , linear z};

[0019] D Gravity = {gravity x , gravity y , gravity z} ;

[0020] D Rotation = {rotation x , rotation y , rotation z , rotation w} ;

[0021] The set of inertial navigation data is defined as:

[0022] D IN = {D Linear , D Gravity , D Rotation} ;

[0023] Step S14, obtaining the operation log information from the industrial equipment, including timestamp, current state of the equipment, operation record, alarm information and maintenance history, as follows:

[0024] D Log = {L1, L2, L3, L4, L5} ;

[0025] Step S15, collecting real-time image data D Image through a camera; wherein D Image represents the collected image data.

[0026] Preferably, in step S2, the industrial asset is precisely positioned by the fusion positioning technology of Wi-Fi and inertial navigation, combined with the collected signal strength, inertial navigation path and feature data in the fingerprint library, and the specific process is as follows:

[0027] Step S21, pre-processing the collected multi-modal data including Wi-Fi signal strength and inertial navigation data;

[0028] Step S22, using inertial navigation technology to estimate the relative change of the equipment position;

[0029] Step S23, in the calculation process, the output of the inertial sensor is used to calculate the moving direction and step length of the equipment, and the position of the equipment is updated gradually;

[0030] Step S24, using the rotation vector quaternion integrated with accelerometer, magnetometer and gyroscope data to solve the heading;

[0031] Step S25, based on the previous position (x k , yk ), step length L k and direction ψ k , the current position (x k+1 ,y k+1 ) of the device is calculated;

[0032] Step S26, using real-time Wi-Fi signal strength data and pre-constructed fingerprint database, the position of the device is predicted using a convolutional neural network model;

[0033] The convolutional neural network model matches the currently scanned Wi-Fi signal with the data in the fingerprint library, thereby obtaining the estimated position of the device;

[0034] Step S27, in data fusion and positioning optimization, the Wi-Fi positioning and inertial navigation results are fused by extended Kalman filtering;

[0035] Inertial navigation provides the trend of relative position change, and Wi-Fi positioning provides absolute position reference; the extended Kalman filter uses a state vector to represent the current position and direction angle of the device, predicts the state through step length and direction angle, and corrects it in combination with the Wi-Fi positioning observation value.

[0036] Preferably, in step S22, the relative change of the device position is estimated using inertial navigation technology; first, based on walking step detection, step length estimation and direction calculation; then, every time a step is detected, the system calculates the relative displacement of the device, thereby obtaining the current relative position.

[0037] Preferably, in step S23, the step length is calculated based on the Weinberg nonlinear step length model, as follows:

[0038]

[0039] where L k is the step length of the kth step; and represent the maximum and minimum acceleration of the kth step, respectively; α represents the step length estimation parameter fitted by least squares method.

[0040] Preferably, in step S24, the moving direction calculation formula is as follows:

[0041]

[0042] where φ, θ, ψ represent roll, pitch, yaw, i.e. roll angle, pitch angle and yaw angle, respectively; q w , q x , q y , q z represent the quaternions provided by the inertial sensor.

[0043] Preferably, in step S25, the position updating formula is as follows:

[0044] X k+1 = x k + L k · sin(ψ k );

[0045] y k+1 = y k + L k · cos(ψ k );

[0046] where x k , y k are the device positions of the previous step; L k is the step size; ψ k is the direction angle of the current step.

[0047] Preferably, in step S3, the multi-modal data is semantically fused and analyzed using a large language model to identify potential risks and generate a security assessment report containing a security rating, key findings, and warning suggestions. The specific process is as follows:

[0048] Step S31, convert multi-modal data into structured information respectively to ensure efficient processing by the large language model;

[0049] Positioning data is arranged into time series position information and direction angle sequence to form a complete trajectory representation; event time, type, and status information are extracted from device logs; image data is extracted into high-dimensional feature vectors to represent visual information changes; after standardization processing, the above data is unified in format for further analysis;

[0050] Step S32, JSON encapsulation of the arranged multi-modal data;

[0051] Organize the positioning data, device logs, and image features into JSON data structures according to the predefined format, with each modality's data represented independently while retaining timestamp and other associated information;

[0052] Step S33, design prompt words to clearly define task objectives and analysis requirements for the large language model;

[0053] Combine JSON data to write task background, analysis objectives, and input data descriptions into prompt words to help the large language model understand the context and output high-quality evaluation results;

[0054] Step S34, call the large language model to perform semantic analysis and comprehensive processing on the input data;

[0055] The large language model maps the positioning data, device logs and image features to a shared semantic space by deeply encoding them through the built-in multi-modal feature fusion and semantic correlation module; by using the self-attention mechanism, the large language model can dynamically analyze the relationship between multi-modal data, automatically discover abnormal patterns and potential risks;

[0056] Step S35, generating a safety evaluation report based on the analysis result of the large language model;

[0057] The report is output in natural language form, which is simple and intuitive, and is convenient for users to quickly understand and apply;

[0058] Step S36, parsing the evaluation report returned by the large language model, and converting the result into structured content for further processing by the system;

[0059] According to the key findings and recommendations in the report, trigger the corresponding risk warning mechanism or take protective measures.

[0060] Preferably, in step S35, the report content includes three parts:

[0061] First, the safety rating, which evaluates the overall safety status of the industrial asset according to the quantitative analysis of the large language model on the risk;

[0062] Second, the key findings, which describe the potential problems or abnormalities identified in detail;

[0063] Third, the improvement suggestions, which propose specific optimization measures for the key findings.

[0064] Therefore, the asset positioning monitoring method based on Wi-Fi and inertial navigation fusion and large language model has the following beneficial effects:

[0065] The large language model is introduced innovatively, which realizes the in-depth understanding and dynamic monitoring of the running state of the industrial asset through its multi-modal data fusion and semantic analysis capability; with this technology, the application can efficiently identify potential equipment failures and abnormal states, and generate warning information in time, effectively breaking through the limitations of traditional single technology, and greatly improving the intelligent level of asset protection.

[0066] The technical solutions of the application will be further described in detail below with the help of the drawings and examples. DESCRIPTION OF DRAWINGS

[0067] Figure 1 is the system block diagram of the asset positioning monitoring method based on Wi-Fi and inertial navigation fusion and large language model of the application;

[0068] Figure 2 is the flowchart of the asset positioning monitoring method based on Wi-Fi and inertial navigation fusion and large language model of the application;

[0069] Figure 3 is a fusion positioning trajectory simulation diagram based on Wi-Fi and inertial navigation of the present application;

[0070] Figure 4 is a structured security assessment report obtained by using a large language model to analyze data in the present application;

[0071] Figure 5 is a large language model prompt word design example for industrial important asset protection analysis in the present application. DETAILED DESCRIPTION

[0072] The technical solutions of the present application are further described below through the drawings and examples.

[0073] As shown in Figure 1 , the asset positioning monitoring method based on Wi-Fi and inertial navigation fusion and large language model of the present application combines multi-modal data acquisition, Wi-Fi and inertial navigation fusion positioning, multi-modal semantic analysis, and abnormal detection and risk assessment modules, and realizes accurate positioning and intelligent protection of industrial assets.

[0074] As shown in Figure 2 , the asset positioning monitoring method based on Wi-Fi and inertial navigation fusion and large language model of the present application includes the following steps:

[0075] Step S1, acquire multi-modal data and construct a fingerprint library; wherein the multi-modal data includes Wi-Fi signals, inertial navigation data, device logs, and image data;

[0076] Step S2, use the fusion positioning technology of Wi-Fi and inertial navigation to accurately position the industrial assets by combining the collected signal strength, inertial navigation path, and feature data in the fingerprint library;

[0077] Step S3, use a large language model to perform semantic fusion and analysis on the multi-modal data, identify potential risks, and generate a security assessment report containing security ratings, key findings, and warning suggestions.

[0078] EMBODIMENT

[0079] Step S1, acquire multi-modal data and construct a fingerprint library; wherein the multi-modal data includes Wi-Fi signals, inertial navigation data, device logs, and image data.

[0080] Step S11, use multi-modal data acquisition equipment to collect data D t , as shown below:

[0081]

[0082] wherein, represents Wi-Fi signal data, represents inertial navigation data, represents device log and image data; the upper index t represents the sampling time.

[0083] Step S12, obtaining signal strength values D WiFi , as follows:

[0084] D WiFi = {rssi1, rssi2, …, rssi n} ;

[0085] wherein, D WiFi represents a set of signal strength values from n access points.

[0086] Step S13, obtaining inertial navigation data of the device through an inertial sensor, specifically including linear acceleration D Linear , gravity acceleration D Gravity and rotation angle D Rotation , as follows:

[0087] D Linear = {linear x , linear y , linear z} ;

[0088] D Gravity = {gravity x , gravity y , gravity z} ;

[0089] D Rotation = {rotation x , rotation y , rotation z , rotation w} ;

[0090] The set of the above inertial navigation data is defined as:

[0091] D IN = {D Linear , D Gravity , D Rotation}.

[0092] Step S14, obtaining running log information from the industrial device, including timestamp, current state of the device, operation record, alarm information and maintenance history, as follows:

[0093] DLog = {L1, L2, L3, L4, L5}.

[0094] Step S15, collecting real-time image data D through the camera Image ; wherein D Image represents the collected image data.

[0095] Step S2, using the fusion positioning technology of Wi-Fi and inertial navigation, combined with the collected signal strength, inertial navigation path and feature data in the fingerprint library, to accurately locate the industrial assets.

[0096] Step S21, pre-processing the collected multi-modal data including Wi-Fi signal strength and inertial navigation data.

[0097] Step S22, using inertial navigation technology to estimate the relative change of the device position.

[0098] First, based on the walking step detection, step length estimation and direction calculation.

[0099] Then, every time a step is detected, the system calculates the relative displacement of the device, and thus the current relative position.

[0100] Step S23, during the calculation process, the output of the inertial sensor is used to calculate the moving direction and step length of the device, and the position of the device is updated gradually.

[0101] Based on the Weinberg nonlinear step length model, the step length is calculated as follows:

[0102]

[0103] where L k is the step length of the kth step; and respectively represent the maximum and minimum acceleration of the kth step; α represents the step length estimation parameter fitted by the least squares method.

[0104] Step S24, using the rotation vector quaternion integrating accelerometer, magnetometer and gyroscope data for heading solution.

[0105] The moving direction calculation formula is as follows:

[0106]

[0107] where φ, θ, ψ represent roll, pitch and yaw, i.e. roll angle, pitch angle and yaw angle; q w , q x , q y , q zrepresents the quaternion provided by the inertial sensor.

[0108] Step S25, based on the previous position (x k ,y k ), step length L k and direction ψ k , the current position of the device (x k+1 ,y k+1 ) is calculated, so the position update formula is as follows:

[0109] x k+1 =x k +L k ·sin(ψ k );

[0110] y k+1 =y k +L k ·cos(ψ k );

[0111] where x k ,y k is the device position of the previous step; L k is the step length; ψ k is the direction angle of the current step.

[0112] Step S26, combine real-time Wi-Fi signal strength data and pre-constructed fingerprint database, use convolutional neural network model to predict the position of the device.

[0113] The convolutional neural network model matches the currently scanned Wi-Fi signal with the data in the fingerprint library, thereby obtaining the estimated position of the device.

[0114] Step S27, in data fusion and positioning optimization, fuse Wi-Fi positioning and inertial navigation results through extended Kalman filtering.

[0115] Inertial navigation provides relative position change trend, Wi-Fi positioning provides absolute position reference. Extended Kalman filtering uses state vector to represent the current position and direction angle of the device, predicts the state through step length and direction angle, and combines with Wi-Fi positioning observation value for correction.

[0116] Step S3, use large language model to perform semantic fusion and analysis on multi-modal data, identify potential risks, and generate a security assessment report containing security rating, key findings and warning suggestions.

[0117] Step S31, convert multi-modal data into structured information respectively, ensuring that it can be efficiently processed by the large language model.

[0118] The positioning data is arranged as a time series of position information and a sequence of direction angles to form a complete trajectory representation; the event time, type, and state information are extracted from the device log; and the image data is extracted as a high-dimensional feature vector to represent changes in visual information. After standardizing the above data, the data is uniformly formatted for further analysis.

[0119] Step S32, the arranged multi-modal data is JSON encapsulated.

[0120] The positioning data, device log, and image features are organized into a JSON data structure according to a predefined format, with each modality's data represented independently while retaining associated information such as timestamps.

[0121] Step S33, design the prompt word to clearly define the task target and the analysis requirements of the large language model.

[0122] In combination with the JSON data, the task background, analysis target, and input data description are written into the prompt word to enable the large language model to understand the context and output high-quality evaluation results.

[0123] Step S34, call the large language model to perform semantic analysis and comprehensive processing on the input data.

[0124] The large language model, through the built-in multi-modal feature fusion and semantic correlation module, deeply encodes the positioning data, device log, and image features and maps them to a shared semantic space. Using the self-attention mechanism, the large language model can dynamically analyze the relationships between multi-modal data and automatically discover abnormal patterns and potential risks.

[0125] Step S35, generate a safety assessment report based on the analysis results of the large language model.

[0126] The report content includes three parts: first, the safety rating, which evaluates the overall safety status of the industrial asset based on the large language model's quantitative analysis of risks; second, the key findings, which describe the identified potential problems or abnormalities in detail; and third, the improvement suggestions, which propose specific optimization measures for the key findings.

[0127] The report is output in natural language form, with concise and intuitive language to facilitate quick understanding and application by users.

[0128] Step S36, parse the evaluation report returned by the large language model and convert the results into structured content for further processing by the system.

[0129] According to the key findings and suggestions in the report, trigger the corresponding risk warning mechanism or take protective measures.

[0130] To verify the authenticity of the application, an experimental simulation was conducted, and the experimental data were based on real Wi-Fi fingerprints and inertial navigation databases, as well as self-built historical logs and image databases.

[0131] As Figure 3 shown, the fusion of Wi-Fi and inertial navigation data using extended Kalman filtering can significantly improve the positioning accuracy and meet the precise positioning requirements of industrial asset protection.

[0132] As Figure 4 shown, using a large language model to analyze data can generate a structured safety assessment report, indicating that a large language model can efficiently integrate and interpret complex industrial data, providing strong support for safety analysis.

[0133] As Figure 5 shown, through clear task descriptions and data requirements, clear guidance is provided for semantic fusion and comprehensive analysis of multi-modal data.

[0134] Therefore, the asset positioning monitoring method based on Wi-Fi and inertial navigation fusion and large language model can efficiently integrate and analyze multi-modal data, quickly identify potential safety hazards in industrial assets. By introducing the semantic analysis and anomaly detection capabilities of the large language model, this method significantly improves the monitoring and risk identification capabilities of industrial asset operating conditions, providing more comprehensive and intelligent support for safety analysis of important industrial assets.

[0135] Finally, it should be noted that: the above examples are only used to illustrate the technical solutions of the present application and not to limit it, although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand: it can still modify or replace the technical solutions of the present application, and these modifications or equivalent replacements cannot make the modified technical solutions deviate from the spirit and scope of the technical solutions of the present application.

Claims

1. A method for asset positioning monitoring based on Wi-Fi and inertial navigation fusion and large language model, characterized in that, Comprising the following steps: Step S1, acquiring multi-modal data and constructing a fingerprint library; wherein the multi-modal data includes Wi-Fi signals, inertial navigation data, device logs, and image data; Step S2, using a fusion positioning technology of Wi-Fi and inertial navigation, combined with the collected signal strength, inertial navigation path, and feature data in the fingerprint library, to accurately locate the industrial assets, the specific process is as follows: Step S21, pre-processing the collected multi-modal data including Wi-Fi signal strength and inertial navigation data; Step S22, using inertial navigation technology to estimate the relative change of the device position; Step S23, during the calculation process, using the output of the inertial sensor to calculate the moving direction and step length of the device, and gradually updating the position of the device; Step S24, using the rotation vector quaternion integrated with accelerometer, magnetometer, and gyroscope data to solve the heading; Step S25, based on previous position (x k ,y k ), step size L k and direction ψ k , calculate current position (x k+1 ,y k+1 ) of the device; Step S26, combining real-time Wi-Fi signal strength data and pre-constructed fingerprint database, using a convolutional neural network model to predict the position of the device; The convolutional neural network model matches the currently scanned Wi-Fi signal with the data in the fingerprint library, thereby obtaining the estimated position of the device; Step S27, in data fusion and positioning optimization, fusing Wi-Fi positioning and inertial navigation results through extended Kalman filtering; Inertial navigation provides relative position change trend, and Wi-Fi positioning provides absolute position reference; extended Kalman filtering uses state vector to represent the current position and direction angle of the device, predicts the state through step length and direction angle, and corrects it combined with Wi-Fi positioning observation value; Step S3, using a large language model to perform semantic fusion and analysis on multi-modal data, identify potential risks, and generate a safety assessment report containing safety rating, key findings, and warning suggestions, the specific process is as follows: Step S31, converting multi-modal data into structured information respectively, ensuring that it can be efficiently processed by the large language model; Positioning data is arranged as time series position information and direction angle sequence, forming a complete trajectory representation; event time, type, and state information are extracted from device logs; image data is extracted as a high-dimensional feature vector, representing changes in visual information; after standardization processing, the above data are unified in format for further analysis; Step S32, JSON packaging of the arranged multi-modal data; Organize the positioning data, device logs, and image features into JSON data structure according to predefined format, each modality's data is represented independently while retaining timestamp and other associated information; Step S33, design prompt words to clearly define task goals and analysis requirements of the large language model; Combine JSON data to write task background, analysis target, and input data description into prompt words, so that the large language model can understand the context and output high-quality evaluation results; Step S34, call the large language model to perform semantic analysis and comprehensive processing on the input data; The large language model encodes the positioning data, device logs, and image features through the built-in multi-modal feature fusion and semantic correlation module, mapping them to a shared semantic space. Using self-attention mechanisms, the large language model can dynamically analyze the relationships between multi-modal data, automatically discover abnormal patterns and potential risks. Step S35, generate a security assessment report based on the analysis results of the large language model; The report is output in natural language form, with concise and intuitive language, making it easy for users to quickly understand and apply; Step S36, parse the evaluation report returned by the large language model and convert the results into structured content for further processing by the system; According to the key findings and recommendations in the report, trigger the corresponding risk warning mechanism or take protective measures.

2. The asset positioning monitoring method based on Wi-Fi and inertial navigation fusion and large language model according to claim 1, characterized in that, In step S1, multi-modal data is obtained to build a fingerprint library; the multi-modal data includes Wi-Fi signals, inertial navigation data, device logs, and image data, and the specific process is as follows: Step S11, collecting data D from different sources using a multi-modal data acquisition device t As follows: wherein, denotes Wi-Fi signal data, denotes inertial navigation data, denotes device log and image data; the upper index t denotes the sampling time; Step S12, obtaining signal strength values D through a plurality of Wi-Fi access points in the environment WiFi As follows: D WiFi = {rssi1, rssi2,..., rssi n}; where D WiFi represents a set of signal strength values from n access points; Step S13, acquiring inertial navigation data of the device by the inertial sensor, specifically including linear acceleration D Linear , gravity acceleration D Gravity , and rotation angle D Rotation , as follows: D Linear = {linear x ,linear y ,linear z}; D Gravity = { gravity x , gravity y , gravity z}; D Rotation = {rotation x , rotation y , rotation z , rotation w}; The set of inertial navigation data is defined as follows: D IN = {D Linear , D Gravity , D Rotation}; Step S14, obtain the running log information from the industrial equipment, including timestamp, current state of the equipment, operation record, alarm information and maintenance history, as follows: D Log = {L1, L2, L3, L4, L5}; Step S15, collecting real-time image data D by camera Image ; wherein D Image represents the collected image data.

3. The asset positioning monitoring method based on Wi-Fi and inertial navigation fusion and large language model according to claim 1, characterized in that: In step S22, the relative change of the device position is estimated using inertial navigation technology; first, based on walking step detection, step length estimation and direction calculation; then, every time a step is detected, the system calculates the relative displacement of the device, thus obtaining the current relative position.

4. The asset positioning monitoring method based on Wi-Fi and inertial navigation fusion and large language model according to claim 1, characterized in that, In step S23, the step length is calculated based on the Weinberg nonlinear step length model, as follows: where L k is the step length of the kth step; and respectively represent the maximum and minimum acceleration of the kth step; and a represents the step length estimate fitted by the least square method.

5. The asset positioning monitoring method based on Wi-Fi and inertial navigation fusion and large language model according to claim 1, characterized in that, In step S24, the moving direction calculation formula is as follows: where φ, θ, ψ represent roll, pitch, yaw, i.e. roll angle, pitch angle and yaw angle, respectively; q w q x q y q z represents the quaternion provided by the inertial sensor.

6. The asset positioning monitoring method based on Wi-Fi and inertial navigation fusion and large language model according to claim 1, characterized in that, In step S25, the position update formula is as follows: x k+1 = x k + L k • sin(ψ k ) ; y k+1 = y k + L k cos (ψ k ); where x k ,y k are the device position of the previous step; L k is the step size; and ψ k is the direction angle of the current step.

7. The asset positioning monitoring method based on Wi-Fi and inertial navigation fusion and large language model according to claim 1, characterized in that, In step S35, the report content includes three parts: First, the safety rating, which evaluates the overall safety status of the industrial asset based on the large language model's quantitative analysis of risks; Second, the key findings, which describe the potential problems or abnormalities identified in detail; Third, improvement suggestions, which propose specific optimization measures for key findings.

Images