An industrial equipment multi-scenario positioning and intelligent warning system and method

Abstract

The present application discloses an industrial equipment multi-scenario positioning and intelligent warning system and method. The system includes: a trajectory analysis module that acquires radar signal data of industrial equipment, extracts signal intensity change features after denoising, and generates a human body dynamic trajectory when the matching degree of the features with a preset metal reflection model exceeds a threshold; an abnormality judgment module that acquires vibration sensing data at key positions, generates a vibration frequency spectrum by wavelet decomposition, and generates an abnormal vibration alarm message if the vibration frequency spectrum exceeds the range; an intrusion warning module that acquires node signal sources, obtains an initial positioning coordinate group through weighted fusion and inputs it into a three-dimensional map model, and obtains a final coordinate result through multi-source fusion. If the result falls within the boundary of a danger fence, an intrusion warning message is sent; a safety assessment module that performs time alignment and depth fusion on the human body dynamic trajectory, abnormal vibration alarm message, and intrusion warning message to generate a safety assessment report. The present invention realizes comprehensive safety monitoring and management of industrial equipment.

Description

Technical Field

[0001] The present invention belongs to the field of safety monitoring, and particularly relates to an industrial equipment multi-scenario positioning and intelligent warning system and method. Background Art

[0002] With the development of safety monitoring technology, related technologies of multi-scenario positioning and intelligent warning systems and methods have emerged. In the complex environment of industrial equipment areas, how to achieve accurate and reliable personnel positioning and safety monitoring is a major technical challenge. In terms of personnel monitoring, it is difficult to accurately obtain the dynamic trajectories of personnel within the equipment area, and it is impossible to detect in a timely manner dangerous behaviors such as personnel approaching dangerous areas or entering the interior of equipment. For the monitoring of the equipment's own state, relying on the method of manual regular inspections is inefficient and cannot detect potential faults in real time. For example, abnormal vibrations in key parts of the equipment are often difficult to capture in a timely manner. At the same time, various positioning technologies such as radar, UWB, and Bluetooth have their own limitations, and a single technology is difficult to meet the requirements of complex scenarios and cannot comprehensively evaluate the safety state of industrial equipment. Summary of the Invention

[0003] Based on this, it is necessary to provide an industrial equipment multi-scenario positioning and intelligent warning system and method that can improve the safety and reliability of industrial equipment operation and reduce accident risks for the above-mentioned technical problems.

[0004] In a first aspect, an embodiment of the present application provides an industrial equipment multi-scenario positioning and intelligent warning system, including:

[0005] A trajectory analysis module, configured to:

[0006] Obtain radar signal data in the industrial equipment area and denoise the radar signal; then extract features from the denoised radar signal data to obtain a signal intensity change feature vector;

[0007] If the matching degree between the signal intensity change feature vector and a preset metal reflection model exceeds a first threshold, generate a human body dynamic trajectory;

[0008] An abnormality judgment module, configured to:

[0009] Obtain vibration sensing data at key positions of industrial equipment; and perform extraction and screening on the vibration sensing data based on the wavelet decomposition algorithm to generate vibration spectra in each frequency band;

[0010] When the vibration spectrum exceeds a preset spectrum range, generate an abnormal vibration alarm message;

[0011] An intrusion warning module, configured to:

[0012] Obtain the UWB signal source and the Bluetooth MESH node signal source in the metal-dense area of the industrial device; and perform weighted fusion on the UWB signal source and the Bluetooth MESH node signal source to obtain an initial positioning coordinate group;

[0013] Input the initial positioning coordinate group into a preset three-dimensional structure map model and process it using a multi-source fusion algorithm to generate a final coordinate result;

[0014] If the final coordinate result falls within the boundary of a preset danger fence, generate an intrusion warning message;

[0015] A safety assessment module, used for:

[0016] Perform time alignment and depth fusion on the human body dynamic trajectory, abnormal vibration alarm information, and intrusion warning information to generate a safety status assessment report.

[0017] Preferably, if the matching degree between the signal strength change feature vector and a preset metal reflection model exceeds a first threshold, generate a human body dynamic trajectory, including:

[0018] Decompose the signal strength change feature vector to obtain attributes including time series change frequency and amplitude distribution;

[0019] Extract parameters based on the signal strength change feature vector decomposed by the preset metal reflection model to obtain metal reflection parameters corresponding to the dimension of the feature vector;

[0020] Use a dynamic threshold segmentation algorithm to judge the similarity coefficient between the metal reflection parameters and the signal strength change feature vector to obtain a feature similarity parameter;

[0021] Judge whether the feature similarity parameter exceeds the upper limit of a preset dynamic threshold interval;

[0022] In response to the feature similarity parameter exceeding the upper limit of the preset dynamic threshold interval, identify the human body dynamic trajectory and construct a spatial coordinate sequence of the human body dynamic trajectory to generate the human body dynamic trajectory corresponding to the signal strength change feature vector.

[0023] Preferably, the feature similarity parameter is represented by the following formula:

[0024] (1);

[0025] Where, represents the feature similarity parameter, represents the dimension of the signal strength change feature vector, represents the metal reflection parameter, represents the component of the signal strength change feature vector, Represents a dynamic threshold parameter.

[0026] Preferably, a weighted fusion is performed on the UWB signal source and the Bluetooth MESH node signal source to obtain an initial positioning coordinate group, including:

[0027] Obtain the phase difference sequence of the UWB signal source and the Bluetooth MESH node signal source in the metal-dense area;

[0028] Match the phase difference sequence with a preset multipath interference template library to generate a compensation offset;

[0029] Superimpose the compensation offset on the signal attenuation model to obtain a corrected propagation loss distribution map;

[0030] Based on the corrected propagation loss distribution map and the dynamic weight coefficient, construct a spatial probability density field;

[0031] Extract the maximum point clusters based on the spatial probability density field to obtain an original positioning coordinate group;

[0032] Perform confidence accumulation calculation on the coordinate points in the original positioning coordinate group to eliminate the coordinate points with accumulated values lower than the second threshold, and obtain an initial positioning coordinate group.

[0033] Preferably, input the initial positioning coordinate group into a preset three-dimensional structure map model and process it using a multi-source fusion algorithm to generate a final coordinate result, including:

[0034] Extract the discrete coordinate data of the initial positioning coordinate group;

[0035] Map the discrete coordinate data to a preset three-dimensional structure map model to generate a three-dimensional positioning point set;

[0036] Construct a state vector matrix based on the three-dimensional positioning point set;

[0037] Input the state vector matrix into a Kalman filter model to obtain the covariance matrix of the positioning point variance;

[0038] Judge whether the value of the covariance matrix exceeds the third threshold;

[0039] In response to the value of the covariance matrix exceeding the third threshold, trigger an RFID node activation signal to obtain positioning compensation data;

[0040] Perform spatial coordinate fusion on the positioning compensation data and the three-dimensional positioning point set to obtain an optimized positioning point set;

[0041] Based on the optimized positioning point set and the multi-source fusion algorithm, generate a final coordinate result.

[0042] Preferably, the covariance matrix is represented by the following formula:

[0043] = (2);

[0044] where, represents the covariance matrix of the current time state, represents the state vector matrix of the time state, represents the process noise covariance matrix.

[0045] Preferably, generating the final coordinate result based on the optimized positioning point set and the multi-source fusion algorithm includes:

[0046] Conduct parameter construction on the optimized positioning point set and the data of each signal source to obtain a positioning input set; the positioning point set includes multiple reference coordinates and confidence parameter;

[0047] Adjust the weight distribution coefficient of the reference coordinates based on the gradient descent algorithm to obtain the weighted fusion positioning coordinates;

[0048] Input the fusion positioning coordinates into the coordinate calibration model, and calculate the distance in combination with the positioning point set to obtain a residual matrix;

[0049] Iteratively correct the confidence parameter and adjust the fusion positioning coordinates according to the residual matrix until the modulus value of the residual matrix reaches the convergence condition, and use the fusion positioning coordinates when the convergence condition is met as the final coordinate result.

[0050] Preferably, performing time alignment and depth fusion on the human body dynamic trajectory, abnormal vibration alarm information, and intrusion warning information to generate a safety status evaluation report includes:

[0051] Obtain the original timestamps of the human body dynamic trajectory and the abnormal vibration alarm information;

[0052] Eliminate the device clock offset of the original timestamps through the timestamp alignment algorithm to obtain an aligned time series;

[0053] Extract the trajectory acceleration frequency domain features and abnormal vibration waveform features of the original timestamps according to the aligned time series to obtain key feature parameters;

[0054] Input the key feature parameters into a preset abnormal correlation analysis model to obtain an abnormal correlation coding vector;

[0055] Fuse the abnormal correlation coding vector with the priority parameter of the intrusion warning information to obtain a safety status evaluation parameter;

[0056] Input the safety status evaluation parameters into a pre-trained safety evaluation model to obtain a safety status evaluation report.

[0057] In a second aspect, an embodiment of the present application provides an industrial equipment multi-scenario positioning and intelligent warning method, including the following steps:

[0058] Obtain radar signal data in the industrial equipment area and denoise the radar signal; then extract features from the denoised radar signal data to obtain a signal intensity change feature vector; if the matching degree between the signal intensity change feature vector and a preset metal reflection model exceeds a first threshold, generate a human body dynamic trajectory;

[0059] Obtain vibration sensing data at key positions of the industrial equipment; and extract and screen the vibration sensing data based on the wavelet decomposition algorithm to generate vibration spectra in each frequency band; when the vibration spectrum exceeds a preset spectrum range, generate an abnormal vibration alarm message;

[0060] Obtain the UWB signal source and the Bluetooth MESH node signal source in the metal-dense area of the industrial equipment; perform weighted fusion on the UWB signal source and the Bluetooth MESH node signal source to obtain a group of initial positioning coordinates; input the group of initial positioning coordinates into a preset three-dimensional structure map model and process it using a multi-source fusion algorithm to generate a final coordinate result; if the final coordinate result falls within the boundary of a preset danger fence, generate an intrusion warning message;

[0061] Perform time alignment and deep fusion on the human body dynamic trajectory, abnormal vibration alarm message, and intrusion warning message to generate a safety status evaluation report.

[0062] Compared with the prior art, the present invention provides an industrial equipment multi-scenario positioning and intelligent warning system and method. The trajectory analysis module acquires radar signal data in the industrial equipment area, denoises it, and extracts the signal intensity change characteristics. When the matching degree of the characteristics with the preset metal reflection model exceeds the threshold, a human body dynamic trajectory is generated. The abnormal judgment module acquires vibration sensing data at key positions of the industrial equipment, uses the wavelet decomposition algorithm for extraction and screening to generate vibration spectra in each frequency band. If the vibration spectrum exceeds the preset range, an abnormal vibration alarm message is generated. The intrusion warning module acquires UWB signal sources and Bluetooth MESH node signal sources in the metal-dense area, obtains the initial positioning coordinate group through weighted fusion, inputs it into the preset three-dimensional structure map model, and uses the multi-source fusion algorithm to obtain the final coordinate result. If the result falls within the boundary of the preset dangerous fence, an intrusion warning message is issued. The safety assessment module performs time alignment and deep fusion on the above-generated human body dynamic trajectory, abnormal vibration alarm message, and intrusion warning message to generate a safety status assessment report. This system and method can quickly capture abnormalities, gain the initiative for equipment maintenance, and reduce the equipment failure downtime; accurately identify the intrusion behavior of the dangerous area by positioning the dynamic trajectory of personnel, effectively prevent personnel from approaching the dangerous area, and reduce the risk of safety accidents; thereby improving the overall operation safety and stability of industrial equipment, and realizing the all-round safety monitoring and efficient management of industrial equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] By referring to the following drawings, the exemplary embodiments of the present invention can be more fully understood. The drawings are used to provide a further understanding of the embodiments of the present application, and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present invention and do not constitute a limitation to the present invention. In the drawings, the same reference numerals generally represent the same components or steps.

[0064] Figure 1 FIG. [X] is a structural block diagram of an industrial equipment multi-scenario positioning and intelligent warning system provided by an exemplary embodiment of the present application;

[0065] Figure 2 FIG. [Y] is a flowchart of an industrial equipment multi-scenario positioning and intelligent warning method provided by an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0066] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the drawings. Although the exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be fully conveyed to those skilled in the art.

[0067] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on the present invention. In addition, the terms "first", "second", "third" are only used for descriptive purposes and should not be construed as indicating or implying relative importance.

[0068] In addition, the technical features involved in different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0069] First, the implementation environment of the embodiments of the present application is described. Exemplarily, the implementation environment includes a data monitoring device, a positioning module, a data processing device, and an alarm device.

[0070] In an industrial equipment multi-scenario positioning and intelligent alarm system, the data monitoring device is responsible for collecting the operation status data of industrial equipment, connecting to the data processing device through a wired or wireless communication link, and transmitting the real-time data to the latter for analysis and processing; the positioning module uses technologies such as satellite positioning, Bluetooth, Wi-Fi, or ultra-wideband to determine the location information of industrial equipment, and also transmits the location data to the data processing device through the communication link; after receiving the data transmitted by the data monitoring device and the positioning module, the data processing device performs in-depth analysis according to preset algorithms and rules; once an abnormal situation is detected, the data processing device immediately sends an instruction to the alarm device through the communication link to trigger audible and visual alarms, SMS alarms, voice alarms, etc., to ensure the efficient and stable operation of the system.

[0071] The data monitoring device integrates various types of sensors, such as temperature sensors, pressure sensors, rotational speed sensors, vibration sensors, and liquid level sensors, etc. The sensors are accurately deployed at the key parts of the mixing equipment and can collect various state parameters during the operation of the equipment in real time and accurately. Through advanced data acquisition technology, the sensors convert physical signals into electrical signals or digital signals, and then perform preprocessing and packaging through the data acquisition and transmission unit, and then transmit the data stably and efficiently to the data processing device through wired communication or wireless communication.

[0072] For industrial equipment used in outdoor open environments, such as concrete mixers on construction sites, a global positioning system module is usually adopted for the positioning module. The module calculates the geographical location of the equipment accurately by receiving signals transmitted by multiple satellites and using the principle of triangulation. Its positioning accuracy can reach the level of several meters, and it can track the driving trajectory and working position of the equipment in real time. The position information obtained by the positioning module is transmitted to the data processing device through the communication link, enabling the system to clearly master the spatial position of the mixing equipment in different scenarios.

[0073] The data processing device usually consists of a high-performance industrial computer, a server or a cloud computing platform. The data processing device receives the device operation status data collected by the data monitoring device and the position information feedback by the positioning module through the communication link. After receiving the data, on the one hand, through data mining and machine learning technologies, trend analysis, fault prediction and performance evaluation are carried out on the device operation status data to identify abnormal patterns that may occur during the device operation; on the other hand, combined with the position information, the operation area and moving path of the industrial equipment are monitored and analyzed to determine whether the device is operating within the specified range. Once the analysis result shows that the device operation is abnormal, the data processing device quickly generates an alarm instruction and sends it to the alarm device through the communication link.

[0074] The alarm device is the terminal for the multi-scenario positioning and intelligent alarm system of industrial equipment to transmit device abnormal information to the operator. When the data processing device detects an abnormal situation in the operation status or position of the industrial equipment, it immediately sends an instruction to the alarm device. The alarm devices cooperate with each other to form an all-round and multi-level alarm system to ensure that the system can notify relevant personnel in a timely and effective manner to handle the device abnormal situation and reduce the losses caused by device failures.

[0075] Combined with the above implementation environment, the application scenarios of the embodiments of the present application are described.

[0076] An industrial equipment multi-scenario positioning and intelligent alarm system and method provided by the embodiments of the present application integrate multiple key devices such as data monitoring, positioning, data processing and alarm, and realize comprehensive and accurate management of industrial equipment. The data monitoring device is deployed at key parts of the equipment, and the operation status data is collected by various sensors. The positioning module uses different positioning technologies to clarify the position of the equipment. Both transmit the data to the data processing device. The latter uses complex algorithms and programs to deeply analyze the device operation status and position data. Once an abnormality is identified, it immediately sends instructions to the sound and light alarm, the short message alarm module, the voice alarm device, etc., to build an all-round intelligent alarm system covering multiple scenarios. Exemplarily, an industrial equipment multi-scenario positioning and intelligent alarm system and method provided by the embodiments of the present application can be applied to at least one of the following scenarios including but not limited to the following scenarios. Only the centralized scenario of the mixing equipment is taken as an example for illustration.

[0077] First, the multi-scenario positioning and intelligent warning system and method for the mixing equipment are applied to the highway construction scenario. During highway construction, the temperature sensor in the data monitoring equipment continuously monitors the temperature of the asphalt mixing tank to ensure that the asphalt heating temperature precisely meets the construction standards; the pressure sensor continuously monitors the pressure exerted on the materials during the mixing process to ensure the full mixing of asphalt and stone materials. The positioning module uses GPS technology to perform real-time positioning on the transport vehicles loaded with asphalt, enabling the construction command center to keep track of the vehicle positions and driving trajectories in real time. The data processing equipment aggregates the data transmitted by the data monitoring equipment and the positioning module to predict possible equipment failures. Once an abnormality occurs in the equipment, such as too high temperature, excessive pressure fluctuations, or the transport vehicle deviating from the predetermined route, the warning equipment immediately responds to remind the on-site personnel to handle it in a timely manner.

[0078] Second, the multi-scenario positioning and intelligent warning system and method for the mixing equipment are applied to the beverage production scenario. During beverage production, various mixing equipment is used to prepare different types of beverages. The rotation speed sensor in the data monitoring equipment strictly regulates the rotation speed of the mixing paddle blades to ensure the uniform mixing of various ingredients in the beverage formula; the liquid level sensor precisely monitors the liquid levels of the raw material tank and the finished product tank to achieve automated batching and canning control. Inside the beverage production workshop, the positioning module locates the mobile mixing equipment. The data processing equipment conducts in-depth analysis on the received data and, based on the quality standards and technological processes of beverage production, determines whether the equipment is operating normally to ensure the accuracy of the beverage formula and the batch stability. If an abnormality is detected, such as abnormal mixing speed or excessive liquid level deviation, the warning equipment is quickly activated to promptly adjust the production process and maintain the brand reputation and production efficiency of the beverage production enterprise.

[0079] Third, the multi-scenario positioning and intelligent warning system and method for the mixing equipment are applied to the pharmaceutical scenario. The vibration sensor in the data monitoring equipment monitors the vibration of the mixing equipment during operation in real time to avoid affecting the quality of drug production due to equipment failures; the pressure sensor and temperature sensor precisely monitor the pressure and temperature inside the reaction kettle to prevent the drug synthesis reaction from getting out of control and ensure the stability and safety of the drug production process. In the complex environment of the pharmaceutical factory area, the positioning module uses ultra-wideband technology to provide high-precision positioning for the mixing equipment to ensure the safe operation of the equipment within the specified clean area. The data processing equipment uses machine learning algorithms to predict equipment failures and abnormal reaction processes. Once an abnormality is discovered, the warning equipment quickly takes action to minimize the risk of pharmaceutical production accidents and ensure the safe and stable operation of drug quality and production facilities.

[0080] Refer to Figure 1 , this embodiment discloses an industrial equipment multi-scenario positioning and intelligent warning system 10, including:

[0081] A trajectory analysis module 101, used for:

[0082] Obtain the radar signal data of the industrial equipment area and denoise the radar signal; then extract features from the denoised radar signal data to obtain the signal strength change feature vector;

[0083] If the matching degree between the signal strength change feature vector and the preset metal reflection model exceeds the first threshold, generate the human body dynamic trajectory;

[0084] Specifically, the industrial equipment in this embodiment is a stirring equipment, and taking this as an example, this module accurately obtains the radar signal data of the industrial equipment area through a specific signal acquisition device. Then, use an advanced signal denoising algorithm to denoise the radar signal data. After denoising, a professional feature extraction algorithm is used to deeply mine the data, so as to obtain the signal strength change feature vector that can reflect the signal change characteristics. The module calculates the matching degree between the obtained signal strength change feature vector and the preset metal reflection model. When the matching degree exceeds the preset threshold, it indicates that the currently detected signal features are highly similar to the metal reflection signal features caused by human activities, and generate the corresponding human body dynamic trajectory.

[0085] The anomaly judgment module 102 is used for:

[0086] Obtain the vibration sensing data of the key positions of the industrial equipment; and based on the wavelet decomposition algorithm, extract and screen the vibration sensing data to generate the vibration spectrum of each frequency band;

[0087] When the vibration spectrum exceeds the preset spectrum range, generate the abnormal vibration alarm information;

[0088] Specifically, the key positions are usually the stirring equipment, which are the parts where the force is relatively complex during operation and the vibration change has a greater impact on the overall performance and stability of the equipment, such as the connection between the stirring shaft and the main body of the equipment, near the transmission components, etc. By installing high-precision vibration sensors at the positions, the original data reflecting the vibration state of the equipment can be collected in real time and accurately. After obtaining the vibration sensing data, the anomaly judgment module extracts and screens it based on the wavelet decomposition algorithm. In this way, the characteristic information of each frequency component can be extracted from the original vibration signal, and then the vibration spectrum of each frequency band can be generated to clearly show the vibration energy distribution of the equipment at different frequencies.

[0089] The intrusion warning module 103 is used for:

[0090] Obtain the UWB signal source and the Bluetooth MESH node signal source in the metal-dense area of the industrial equipment; and perform weighted fusion on the UWB signal source and the Bluetooth MESH node signal source to obtain the initial positioning coordinate group;

[0091] Input the initial positioning coordinate group into a preset three-dimensional structure map model and process it using a multi-source fusion algorithm to generate the final coordinate result;

[0092] If the final coordinate result falls within the boundary of the preset danger fence, an intrusion warning message is generated;

[0093] Specifically, this module first uses a dedicated signal receiving device to obtain the UWB signal source and Bluetooth MESH node signal source in the metal-dense area of the mixing equipment. According to factors such as the stability and accuracy of UWB and Bluetooth MESH signals in different environments, different weights are assigned to them. Through a weighted fusion algorithm, the signals are comprehensively processed to obtain the initial positioning coordinate group. Then, the intrusion warning module inputs the initial positioning coordinate group into a preset three-dimensional structure map model. At the same time, using a multi-source fusion algorithm, combined with information such as terrain and buildings in the map model, the initial positioning coordinate group is deeply processed and optimized, and finally an accurate final coordinate result is generated to achieve precise positioning of the target object.

[0094] Safety assessment module 104, for:

[0095] Perform time alignment and deep fusion on the human body's dynamic trajectory, abnormal vibration alarm information, and intrusion warning information to generate a safety status assessment report.

[0096] First, use an advanced timestamp alignment algorithm to precisely process the original timestamps attached to the information. By eliminating error factors such as device clock offsets, data from different sources are accurately aligned in the time dimension to ensure the correct correspondence of each piece of information on the time axis. After completing the time alignment, the safety assessment module conducts deep fusion on the information. For the human body's dynamic trajectory, extract the frequency domain characteristics of the trajectory acceleration to analyze the changing pattern of the person's movement state in the mixing equipment area; for the abnormal vibration alarm information, extract the waveform characteristics of the abnormal vibration to accurately judge the type and severity of the equipment's abnormal vibration. At the same time, considering the priority parameters of the intrusion warning information, take into account the differences in the potential hazards of intrusion behaviors in different areas. Organically fuse the information after feature extraction and priority processing to construct a feature vector that comprehensively reflects the safety status of the mixing equipment.

[0097] Subsequently, the security assessment module inputs the fused feature vectors into a pre-trained assessment model. This assessment model is constructed based on a large amount of historical data and professional machine learning algorithms. The large amount of historical data includes a vast amount of relevant historical data such as personnel activity trajectories, equipment vibration data, and area intrusion events collected during the long-term operation of industrial equipment. These data cover various normal and abnormal working conditions. Using professional machine learning algorithms such as decision trees and neural networks, the historical data is used as training samples, allowing the algorithms to automatically learn the complex patterns, rules, and feature relationships in the data. During the training process, the algorithms continuously adjust their own parameters to optimize the model performance, enabling it to accurately capture the internal connections between personnel activities, equipment operating states, and area intrusion situations. After repeated training and verification, an assessment model that can accurately evaluate the security state of industrial equipment is finally constructed to ensure that it can generate a reliable security state assessment report based on the input fused feature vectors.

[0098] The above industrial equipment multi-scenario positioning and intelligent warning system can quickly capture anomalies, gain the upper hand for equipment maintenance, and reduce equipment failure downtime; accurately identify the dynamic trajectories of personnel and locate intrusion behaviors in dangerous areas, effectively preventing personnel from approaching dangerous areas and reducing the risk of safety accidents; thereby enhancing the overall operation safety and stability of the mixing equipment and achieving all-round safety monitoring and efficient management of the mixing equipment.

[0099] In a preferred embodiment, if the matching degree between the signal strength change feature vector and the preset metal reflection model exceeds the first threshold, a human body dynamic trajectory is generated, including:

[0100] Decompose the signal strength change feature vector to obtain the time-series change frequency and amplitude distribution attributes;

[0101] Extract parameters from the decomposed signal strength change feature vector based on the preset metal reflection model to obtain metal reflection parameters corresponding to the dimension of the feature vector;

[0102] Specifically, the preset metal reflection model is a mathematical model constructed based on a large amount of experimental data and theoretical analysis, used to describe the characteristics of radar signals when reflected by metal objects. Considering the characteristics of radar signals such as frequency, amplitude, and phase, as well as the influence of factors such as the shape, material, and position of metal objects on the reflected signals, it provides a standard reference for judging whether there is a metal reflection caused by human activities in radar signals.

[0103] Preferably, metal reflection parameters corresponding to the dimension of the signal strength change feature vector are extracted. The parameters include attributes reflecting the time-series change frequency and amplitude distribution of the metal reflection signal, used for comparative analysis with the actually obtained radar signal feature vector to judge whether there is a change in the reflection trajectory caused by human activities.

[0104] The similarity coefficient of the metal reflection parameter and the signal intensity change eigenvector is judged by using a dynamic threshold segmentation algorithm to obtain a feature similarity parameter;

[0105] Specifically, the dynamic threshold parameter determines the range of the dynamic threshold interval. The dynamic threshold parameter will be adjusted in real time according to factors such as the system operation status and environmental changes, and the upper and lower limits of the dynamic threshold interval are determined based on this parameter. When judging the feature similarity parameter, if it exceeds the upper limit of the preset dynamic threshold interval, a corresponding action is triggered (such as identifying the human body's dynamic trajectory).

[0106] Judge whether the feature similarity parameter exceeds the upper limit of the preset dynamic threshold interval;

[0107] In response to the feature similarity parameter exceeding the upper limit of the preset dynamic threshold interval, identify the human body's dynamic trajectory, construct a spatial coordinate sequence of the human body's dynamic trajectory, and generate the human body's dynamic trajectory corresponding to the signal intensity change eigenvector.

[0108] Specifically, first extract the signal intensity change eigenvector containing the time series change frequency and amplitude distribution attributes from the denoised radar signal data. Then, based on the preset metal reflection model, extract the metal reflection parameters corresponding to the dimension of the eigenvector. Subsequently, use the dynamic threshold segmentation algorithm to judge the similarity coefficient between the metal reflection parameter and the eigenvector to obtain the feature similarity parameter. If this parameter exceeds the upper limit of the preset dynamic threshold interval, it is determined that there is a change in the reflection trajectory, and the human body's dynamic trajectory recognition is started. Finally, construct a spatial coordinate sequence according to the human body's dynamic trajectory recognition result to obtain the human body's dynamic trajectory corresponding to the eigenvector.

[0109] By finely processing the radar signal data in multiple steps, noise interference can be effectively excluded, and the signal feature changes caused by human activities can be accurately captured. Using the preset metal reflection model and dynamic threshold algorithm can improve the accuracy and reliability of recognition and reduce the misjudgment rate. The constructed spatial coordinate sequence of the human body's dynamic trajectory provides intuitive and accurate personnel activity path information for safety management, helps to detect abnormal behaviors in a timely manner, and ensures the safety of personnel and the normal operation of equipment in the mixing equipment area.

[0110] In a preferred embodiment, the feature similarity parameter is represented by the following formula:

[0111] (1);

[0112] Wherein, represents the feature similarity parameter, represents the dimension of the signal intensity change eigenvector, represents the metal reflection parameter, represents the component of the signal intensity change eigenvector, Represents the dynamic threshold parameter.

[0113] On the one hand, through the feature extraction and analysis of the denoised radar signal, the signal changes caused by human activities can be effectively captured, improving the accuracy and reliability of human dynamic trajectory recognition and providing accurate data support for subsequent safety monitoring and early warning. On the other hand, by using the preset metal reflection model and dynamic threshold segmentation algorithm, it is possible to quickly and accurately judge whether there is human activity and the change of the activity trajectory, enhancing the real-time response ability and intelligent level of the system. At the same time, the constructed human dynamic trajectory spatial coordinate sequence can intuitively display the activity path of personnel in the mixing equipment area, helping managers to detect abnormal behaviors in a timely manner, take corresponding safety measures, and ensure the safe operation of the mixing equipment and the life safety of personnel.

[0114] In a preferred embodiment, the UWB signal source and the Bluetooth MESH node signal source are weighted and fused to obtain an initial positioning coordinate group, including:

[0115] Obtain the phase difference sequence of the UWB signal source and the Bluetooth MESH node signal source in the metal-dense area;

[0116] Match the phase difference sequence with the preset multipath interference template library to generate a compensation offset;

[0117] Preferably, the multipath interference template library is constructed based on a large amount of test data of the UWB and Bluetooth MESH node signal sources in complex environments such as metal-dense areas. The phase difference sequence data of signal propagation is collected under different scenarios, different signal propagation distances, and different interference conditions, covering various possible multipath interference situations. Using algorithms such as clustering analysis and pattern recognition, the collected data is classified and feature-extracted to find the typical phase difference sequence patterns corresponding to different types of multipath interference. Organizing these typical patterns into a library forms the preset multipath interference template library.

[0118] Superimpose the compensation offset on the signal attenuation model to obtain a corrected propagation loss distribution map;

[0119] Specifically, the signal attenuation model is a mathematical model used to describe the law of signal intensity weakening with changes in transmission distance, environmental factors, etc. in the process of signal transmission. In this industrial equipment positioning scenario, it comprehensively considers the influence of factors such as spatial loss, obstruction by obstacles, and multipath effect on the signal intensity during signal propagation. Through the quantitative analysis of these factors, a functional relationship that can reflect the gradual decrease of signal intensity from the transmitter to the receiver is constructed.

[0120] Based on the corrected propagation loss distribution map and the dynamic weight coefficient, construct a spatial probability density field;

[0121] Furthermore, the dynamic weight coefficient is a weight value that is adjusted in real time according to the real-time characteristics of different signal sources, such as the stability, accuracy, etc. of UWB and Bluetooth MESH node signal sources in different environments. In practical applications, it is set by monitoring and analyzing the performance of signal sources in complex environments. Signal sources with good stability and high accuracy are given higher weights, while those with lower stability and accuracy have lower weights. For example, in a metal-dense area, if the UWB signal is less interfered with and has high positioning accuracy, its weight will be relatively increased; while if the Bluetooth MESH node signal is highly interfered with, its weight will be correspondingly decreased.

[0122] Based on the spatial probability density field, the extraction of the maximum point clusters is carried out to obtain the original positioning coordinate group;

[0123] The confidence accumulation calculation is performed on the coordinate points in the original positioning coordinate group to eliminate the coordinate points with accumulated values lower than the second threshold, and the initial positioning coordinate group is obtained.

[0124] Specifically, the second threshold is a standard for screening coordinate points when performing the confidence accumulation calculation on the original positioning coordinate group. The setting of this second threshold usually depends on a large amount of experimental data and actual application scenarios. Generally speaking, it is comprehensively determined according to factors such as the allowable error range of the positioning system, the stability of the signal, and the degree of environmental interference. If the positioning requires high precision, stable signal, and low interference, the second threshold can be set higher; on the contrary, in the case of lower positioning accuracy requirements and easily interfered signals, the second threshold can be appropriately reduced.

[0125] In this embodiment, first, a professional signal receiving device is used to obtain the phase difference sequence of UWB and Bluetooth MESH node signal sources in this area. This sequence contains the phase change information during the signal propagation process. Then, the obtained phase difference sequence is matched with a pre-established multipath interference template library, and the real-time obtained phase difference sequence is matched with the patterns in the template library to quickly determine the type of multipath interference suffered by the current signal. Through the matching, a compensation offset for compensating the signal deviation is generated. Subsequently, the generated compensation offset is superimposed on traditional signal attenuation models, such as the free space propagation model, the logarithmic distance path loss model, the Okumura-Hata model, etc. The model is corrected to obtain a corrected propagation loss distribution map that is more in line with the actual situation, and this map can more accurately reflect the propagation loss situation of the signal in the metal-dense area.

[0126] Construct a spatial probability density field based on the corrected propagation loss distribution map combined with dynamic weight coefficients. The dynamic weight coefficients will be adjusted in real time according to factors such as the stability and accuracy of different signal sources, so that the spatial probability density field can more accurately reflect the possible position distribution of the target in space. Then, perform a maximum point clustering extraction operation on the constructed spatial probability density field, determine the regions with larger probability density as the possible target positions, and thus obtain the initial positioning coordinate group. Finally, in order to improve the accuracy and reliability of positioning, perform a confidence accumulation calculation on each coordinate point in the initial positioning coordinate group, set a preset threshold, and eliminate the coordinate points with cumulative values lower than the threshold to obtain the updated initial positioning coordinate group.

[0127] In this embodiment, by utilizing the complementary characteristics of UWB and Bluetooth MESH node signal sources, more comprehensive positioning information can be obtained, enhancing the adaptability of the positioning system in complex metal environments. Compensate for multipath interference and correct the signal attenuation model, effectively reducing the impact of signal interference and propagation loss on positioning accuracy and improving the accuracy of positioning. The introduction of dynamic weight coefficients and the construction of the spatial probability density field make the positioning results more scientific and reasonable, and can more accurately reflect the actual position of the target. And performing a confidence accumulation calculation and screening on the coordinate points further removes possible mispositioning points, improves the reliability of the positioning results, and provides a solid foundation for subsequent safety monitoring, personnel management, etc.

[0128] In a preferred embodiment, input the initial positioning coordinate group into a preset three-dimensional structure map model and process it using a multi-source fusion algorithm to generate the final coordinate result, including:

[0129] Extract the discrete coordinate data of the initial positioning coordinate group;

[0130] Map the discrete coordinate data to the preset three-dimensional structure map model to generate a three-dimensional positioning point set;

[0131] Specifically, the three-dimensional structure map model contains the three-dimensional coordinate information of the terrain, buildings, obstacles, etc. in the area where the device is located, as well as the spatial relationship between them, presenting the positions of objects in space.

[0132] The three-dimensional positioning point set is a set of points with spatial position information formed after mapping the discrete coordinate data to the preset three-dimensional structure map model. These points correspond to the possible positions of the target to be located in the three-dimensional map, and each point contains accurate three-dimensional coordinates (such as X, Y, Z axis coordinates) to determine the specific orientation of the target in space. At the same time, the points in the point set may also carry relevant attribute information, such as the confidence parameter of the positioning, used to characterize the reliability of the positioning point.

[0133] Construct a state vector matrix based on the three-dimensional positioning point set;

[0134] Input the state vector matrix into the Kalman filter model to obtain the covariance matrix of the positioning point variance;

[0135] Determine whether the value of the covariance matrix exceeds the third threshold;

[0136] Further, the third threshold is set by conducting multiple positioning tests on the initial positioning coordinate group under different environmental conditions based on a large amount of experimental data, and recording the values of the covariance matrix in each test. Analyze the distribution of these data, and combine the system positioning accuracy requirements and the acceptable error range to set the threshold. If the system has extremely high requirements for positioning accuracy and tries to minimize the positioning error, the threshold will be set lower, making it easier to meet the conditions for triggering the RFID node to obtain positioning compensation data; conversely, if the accuracy requirements are relatively low, to avoid wasting system resources due to excessive compensation, the threshold will be set higher.

[0137] In response to the value of the covariance matrix exceeding the third threshold, trigger the RFID node activation signal to obtain positioning compensation data;

[0138] Specifically, the RFID reader sends a radio frequency signal to the surrounding RFID tags to activate the tags. The activated tags return the stored unique identification information and data related to positioning, such as the pre-stored location information of the tag or the relative position information with surrounding reference points, etc. After the system receives these returned data, it parses and processes them, associates the tag location information with the three-dimensional positioning point set, and calculates the positioning compensation data based on the relative position relationship between the tag and the target. These data include position correction values, direction deviation adjustment values, etc., which are used for spatial coordinate fusion of the three-dimensional positioning point set to optimize the positioning result and improve the positioning accuracy.

[0139] Perform spatial coordinate fusion on the positioning compensation data and the three-dimensional positioning point set to obtain an optimized positioning point set;

[0140] Generate the final coordinate result based on the optimized positioning point set and the multi-source fusion algorithm.

[0141] Specifically, the multi-source fusion algorithm includes the weighted average method, the Kalman filter algorithm, the particle filter algorithm, etc. The weighted average method assigns different weights according to the reliability, accuracy, etc. of each signal source, and performs weighted calculation and fusion on the data; the Kalman filter algorithm performs optimal estimation and fusion on the signal with noise through two steps of prediction and update, using the system state equation and the observation equation; the particle filter algorithm is based on Monte Carlo simulation, represents the system state through a large number of particles, adjusts the particle weights according to the observation data, and then performs resampling and other operations to achieve fusion. In practical applications, the data of signal sources such as UWB and Bluetooth MESH are first preprocessed to remove noise and outliers; then, a suitable algorithm model is selected according to the algorithm characteristics, and relevant data is input for calculation; finally, the fusion result is analyzed and evaluated. If the accuracy requirement is not met, the algorithm parameters are adjusted or re-fused, so as to obtain accurate and reliable positioning results.

[0142] In a preferred embodiment, the covariance matrix is represented by the following formula:

[0143] = (2);

[0144] Wherein, represents the covariance matrix of the current state, represents the state vector matrix of the state at time, and

[0145] In this embodiment, by gradually optimizing the initial positioning coordinate group, the accuracy and reliability of the positioning data are effectively improved. Combining the positioning data with the three-dimensional structure map model endows the positioning result with actual spatial significance, which is convenient for the operator to intuitively understand the target position. The application of the Kalman filter model and the mechanism of triggering the RFID node to obtain positioning compensation data according to the covariance matrix significantly improve the ability of the positioning system to cope with complex environments and data fluctuations, and reduce the positioning error. The application of the multi-source fusion algorithm makes full use of the advantages of multiple data sources, comprehensively improves the accuracy and stability of positioning, provides accurate and reliable position information support for the safety monitoring, personnel management, etc. of the mixing equipment area, and effectively guarantees the efficient operation of the entire system.

[0146] In a preferred embodiment, based on the optimized positioning point set and the multi-source fusion algorithm, the final coordinate result is generated, including:

[0147] Construct parameters for the optimized positioning point set and the data of each signal source to obtain a positioning input set; the positioning point set includes multiple reference coordinates and confidence parameters;

[0148] Specifically, each signal source data includes the UWB signal source and the Bluetooth MESH node signal source in the metal-dense area. The UWB signal source data contains information related to the phase change of signal propagation, such as the phase difference sequence. Similarly, the Bluetooth MESH node signal source data also contains similar phase change information and its own characteristic data. Integrate the optimized positioning point set with the UWB and Bluetooth MESH node signal source data, extract the key parameters related to positioning from the signal source data, combine the reference coordinates and confidence parameters in the optimized positioning point set, and organize and arrange these data according to specific rules to form a positioning input set.

[0149] Adjust the weight distribution coefficient of the reference coordinates based on the gradient descent algorithm to obtain the weighted fusion positioning coordinates;

[0150] Specifically, the initial weights are generally randomly assigned or set according to experience, so that each reference coordinate has an initial contribution ratio in the fusion. The weight distribution rule aims to minimize the objective function, usually guided by the positioning error. The reference coordinates with high positioning accuracy and good stability will be assigned higher weights.

[0151] The gradient descent algorithm first determines the objective function, such as calculating the sum of squared errors between the fusion positioning coordinates and the true coordinates as the objective function; then calculates the gradient of the objective function with respect to the weight distribution coefficient, which reflects the rate of change of the objective function when the weight changes slightly; then adjusts the weight distribution coefficient in the opposite direction of the gradient, and the step size is determined according to the learning rate. The learning rate is a hyperparameter set by humans, which controls the amplitude of each weight adjustment to avoid adjusting too fast and missing the optimal solution or adjusting too slowly resulting in a too slow convergence rate; continuously repeat the above process until the objective function converges, that is, the change is very small or no longer changes, to obtain the optimized weights, and use these weights to perform weighted summation on the reference coordinates to obtain the weighted fusion positioning coordinates.

[0152] Input the fusion positioning coordinates into the coordinate calibration model, and calculate the distance in combination with the positioning point set to obtain the residual matrix;

[0153] Specifically, the coordinate calibration model corrects the initial positioning result by fusing multi-source positioning data and a preset map model. The core mechanism is to quantify the positioning error by calculating the residual matrix: map the fusion positioning coordinates to the preset positioning point set in the three-dimensional structure map model, calculate the Euclidean distance between the two to form an observation distance matrix; at the same time, construct a theoretical distance matrix based on the true position relationship of the reference points in the map model. The residual matrix is the difference matrix between the observation distance and the theoretical distance, which reflects the deviation degree between the positioning result and the real environment.

[0154] Iteratively correct the confidence parameter and adjust the fusion positioning coordinates according to the residual matrix until the modulus value of the residual matrix reaches the convergence condition, and take the fusion positioning coordinates when the convergence condition is met as the final coordinate result.

[0155] Further, the convergence condition refers to one or more criteria set during the iterative correction process for determining whether the residual matrix has reached stability and is small enough. Generally, the convergence condition can be that the modulus value of the residual matrix is less than a preset threshold, which is determined according to specific positioning accuracy requirements and actual application scenarios.

[0156] Specifically, to generate the final coordinate result, first, the optimized positioning point set and the data of each signal source are constructed into a positioning input set containing multiple reference coordinates and confidence parameters. Then, the gradient descent algorithm is used to adjust the reference coordinate weight distribution coefficients to obtain the weighted fusion positioning coordinates. Subsequently, the fusion positioning coordinates are input into the coordinate calibration model, and the residual matrix is calculated by combining the positioning point set to calculate the distance. Finally, the confidence parameter and the fusion positioning coordinates are iteratively corrected based on the residual matrix until the modulus value of the residual matrix converges, and the final coordinate result that satisfies the multi-source constraint relationship is obtained. This process effectively improves the positioning accuracy and comprehensively utilizes the advantages of multi-source data.

[0157] In a preferred embodiment, time alignment and deep fusion are performed on the human body dynamic trajectory, abnormal vibration alarm information, and intrusion warning information to generate a safety status assessment report, including:

[0158] Obtain the original timestamps of the human body dynamic trajectory and abnormal vibration alarm information;

[0159] Eliminate the device clock offset of the original timestamps through the timestamp alignment algorithm to obtain the aligned time series;

[0160] Specifically, eliminating the device clock offset is to reduce or remove the differences in clock timing between different devices, ensure that each device is synchronized in time, and avoid errors or inconsistencies in data transmission, processing, etc. caused by clock inconsistencies. Its processing process generally includes the following steps: First, select a reference clock source with high accuracy and stability as the time reference standard for the entire system. Then, through time synchronization protocols such as the Network Time Protocol (NTP), compare the clocks of each device with the reference clock source to calculate the clock offset. Next, according to the calculated offset, adjust the clock of the device. Software algorithms can be used to fine-tune the frequency of the device clock or directly set the time value of the clock to make the device clock consistent with the reference clock. During the operation of the system, clock synchronization checks and adjustments are also performed regularly to adapt to the possible drift of the device clock and the influence of external environmental factors on the clock, so as to continuously maintain the clock synchronization between devices and ensure the normal operation of the system.

[0161] Extract the trajectory acceleration frequency domain features and abnormal vibration waveform features from the original timestamps according to the aligned time series to obtain key feature parameters;

[0162] Input the key feature parameters into a preset abnormal association analysis model to obtain an abnormal association coding vector;

[0163] Specifically, the abnormal association analysis model is a data analysis model used to discover abnormal patterns and their interrelationships in data. It is usually based on technologies such as machine learning and data mining. By learning and analyzing a large amount of historical data and real-time data, a normal association pattern between data is constructed. Then, when new data is input, the model will identify abnormal data points based on the established pattern and analyze the association relationships between these abnormal data to determine whether there are potential abnormal events or trends.

[0164] Fuse the abnormal association coding vector with the priority parameter of the intrusion warning information to obtain a security status evaluation parameter;

[0165] Furthermore, the priority parameter is used to quantify the influence weights of different factors on positioning accuracy and security evaluation. It usually includes signal quality parameters, device credibility parameters, environmental adaptability parameters, and security risk levels. The setting of the parameters is usually achieved through multi-objective optimization algorithms, such as the Analytic Hierarchy Process (AHP) or the entropy weight method. It is trained based on historical data in the offline stage and dynamically adjusted according to real-time environment feedback in the online stage to ensure the positioning reliability and security response speed of the system in complex industrial scenarios.

[0166] Input the security status evaluation parameter into a pre-trained security evaluation model to obtain a security status evaluation report.

[0167] Specifically, the security evaluation model is a model constructed based on machine learning or deep learning algorithms. By learning and analyzing a large amount of data related to the security status, such as various security status evaluation parameters and corresponding security status labels (such as safe, unsafe, at risk, etc.), it mines the features and laws in the data, so as to be able to evaluate and judge the security status of the system or object according to the input security status evaluation parameter.

[0168] In this embodiment, first obtain the original timestamps of the human body dynamic trajectory data and the abnormal vibration alarm information. Then, use the timestamp alignment algorithm to eliminate the device clock offset and obtain the aligned time series. Subsequently, based on the aligned time series, extract the trajectory acceleration frequency domain features and abnormal vibration waveform features to obtain the key feature parameters. After that, input the key feature parameters into a preset abnormal association analysis model to obtain an abnormal association coding vector. Then, fuse this vector with the priority parameter of the intrusion warning information to obtain a security status evaluation parameter. Finally, input the evaluation parameter into the trained evaluation model to generate a security status evaluation report.

[0169] In this embodiment, the timestamp alignment algorithm is used to solve the problem of device clock offset, ensuring the consistency of data in the time dimension and providing an accurate basis for subsequent analysis. Feature extraction is performed on multi-source data, and correlation analysis and fusion are carried out, comprehensively considering various factors such as personnel activities, equipment vibration, and area intrusion, making the evaluation results more comprehensive and accurate. The trained evaluation model can efficiently process evaluation parameters and generate a scientific safety status evaluation report, providing a strong decision-making basis for ensuring the safe and stable operation of the mixing equipment.

[0170] In another embodiment, the present application also provides an industrial equipment multi-scenario positioning and intelligent warning method, including the following steps:

[0171] S1: Obtain the radar signal data in the industrial equipment area and denoise the radar signal; then perform feature extraction on the denoised radar signal data to obtain a signal intensity change feature vector; if the matching degree between the signal intensity change feature vector and the preset metal reflection model exceeds the first threshold, generate a human body dynamic trajectory;

[0172] S2: Obtain the vibration sensing data at the key positions of the industrial equipment; and perform extraction and screening on the vibration sensing data based on the wavelet decomposition algorithm to generate vibration spectra in each frequency band; when the vibration spectrum exceeds the preset spectrum range, generate an abnormal vibration alarm message;

[0173] S3: Obtain the UWB signal source and the Bluetooth MESH node signal source in the metal-dense area of the industrial equipment; perform weighted fusion on the UWB signal source and the Bluetooth MESH node signal source to obtain an initial positioning coordinate group; input the initial positioning coordinate group into the preset three-dimensional structure map model and use the multi-source fusion algorithm for processing to generate a final coordinate result; if the final coordinate result falls within the preset dangerous fence boundary, generate an intrusion warning message;

[0174] S4: Perform time alignment and deep fusion on the human body dynamic trajectory, abnormal vibration alarm message, and intrusion warning message to generate a safety status evaluation report.

[0175] The solution for solving the problem provided by the system in the embodiment of the present application is similar to the solution described in the above system. Therefore, the specific limitations in the following embodiments can refer to the limitations on an indoor positioning and temperature anomaly collaborative warning system in the above text, and will not be repeated here.

[0176] The above multi-scenario positioning and intelligent warning method for industrial equipment first obtains radar signal data in the area of the mixing equipment through a specific device, denoises it, and then extracts features to obtain the signal intensity change features. Once the matching degree of these features with the preset metal reflection model exceeds the threshold, a human body dynamic trajectory is generated. Then, vibration sensing data at key positions of the mixing equipment is obtained, and the wavelet decomposition algorithm is used to extract and screen the data to generate vibration spectra in each frequency band. When the vibration spectrum exceeds the preset spectrum range, an abnormal vibration alarm message is generated. Subsequently, the UWB signal source and the Bluetooth MESH node signal source in the metal-dense area are obtained, and the signal sources are weighted and fused to obtain a group of initial positioning coordinates. The group of initial positioning coordinates is input into the preset three-dimensional structure map model, and the multi-source fusion algorithm is used to generate the final coordinate result. If the final coordinate result falls within the boundary of the preset danger fence, an intrusion warning message is generated. Finally, the generated human body dynamic trajectory, abnormal vibration alarm message, and intrusion warning message are time-aligned and deeply fused to obtain a comprehensive and accurate safety status assessment report. This method can quickly capture abnormalities, gain the initiative for equipment maintenance, and reduce the equipment failure downtime; accurately identify the intrusion behavior of the dangerous area by positioning the dynamic trajectory of personnel, effectively prevent personnel from approaching the dangerous area, and reduce the risk of safety accidents; and further improve the overall operation safety and stability of the mixing equipment, realizing the full-range safety monitoring and efficient management of the mixing equipment.

[0177] It should be understood that although the steps in the flowcharts involved in the above embodiments are shown in sequence according to the arrows, these steps do not necessarily have to be executed in the order indicated by the arrows. Unless clearly stated in this article, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, at least some of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages. These steps or stages do not necessarily have to be executed at the same time, but can be executed at different times. The execution order of these steps or stages does not necessarily have to be sequential, but can be executed alternately or in turn with at least some of the steps or stages in other steps or other steps.

[0178] It should be noted that the flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, as well as the combination of blocks in the block diagram and / or flowchart, may be implemented by a dedicated hardware-based system that performs the specified functions or actions, or may be implemented by a combination of dedicated hardware and computer instructions.

[0179] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0180] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For another example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the couplings, direct couplings, or communication connections shown or discussed among each other may be through some communication interfaces. The indirect couplings or communication connections of the devices or units may be in electrical, mechanical, or other forms.

[0181] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0182] In addition, the functional units in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0183] When the above-mentioned functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The foregoing storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs.

[0184] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of this application, rather than to limit them; although this application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of various embodiments of this application, and they should all be covered by the scope of the claims and the description of this application.

Claims

1. An industrial equipment multi-scenario positioning and intelligent warning system, characterized in that, Including: A trajectory analysis module, configured to: Obtain radar signal data of an industrial equipment area and denoise the radar signal; then extract features from the denoised radar signal data to obtain a signal intensity change feature vector; If the matching degree between the signal intensity change feature vector and a preset metal reflection model exceeds a first threshold, generate a human body dynamic trajectory; the metal reflection model is used to describe the characteristics of radar signals when reflected by metal objects; An abnormality judgment module, configured to: Obtain vibration sensing data at key positions of industrial equipment; and perform extraction and screening on the vibration sensing data based on a wavelet decomposition algorithm to generate vibration spectra of each frequency band; When the vibration spectrum exceeds a preset spectrum range, generate an abnormal vibration alarm message; An intrusion warning module, configured to: Obtain UWB signal sources and Bluetooth MESH node signal sources in the metal-dense area of the industrial equipment; and perform weighted fusion on the UWB signal sources and Bluetooth MESH node signal sources to obtain an initial positioning coordinate group; Input the initial positioning coordinate group into a preset three-dimensional structure map model and process it using a multi-source fusion algorithm to generate a final coordinate result; If the final coordinate result falls within the boundary of a preset dangerous fence, generate an intrusion warning message; A safety assessment module, configured to: Perform time alignment and depth fusion on the human body dynamic trajectory, abnormal vibration alarm message, and intrusion warning message to generate a safety status assessment report; Among them, if the matching degree between the signal intensity change feature vector and a preset metal reflection model exceeds a first threshold, generating a human body dynamic trajectory includes: Decompose the signal intensity change feature vector to obtain time series change frequencies and amplitude distribution attributes; Extract parameters based on the signal intensity change feature vector decomposed by the preset metal reflection model to obtain metal reflection parameters corresponding to the dimension of the feature vector; Use a dynamic threshold segmentation algorithm to judge the similarity coefficient between the metal reflection parameters and the signal intensity change feature vector to obtain a feature similarity parameter; Judge whether the feature similarity parameter exceeds the upper limit of a preset dynamic threshold interval; In response to the feature similarity parameter exceeding the upper limit of the preset dynamic threshold interval, identify the human body dynamic trajectory and construct a spatial coordinate sequence of the human body dynamic trajectory to generate the human body dynamic trajectory corresponding to the signal intensity change feature vector.

2. The system according to claim 1, wherein The feature similarity parameter is represented by the following formula: （1）； Among them, represents the feature similarity parameter, represents the dimension of the signal intensity change feature vector, represents the metal reflection parameter, represents the component of the signal intensity change feature vector, represents the dynamic threshold parameter.

3. The system according to claim 1, wherein Performing weighted fusion on UWB signal sources and Bluetooth MESH node signal sources to obtain an initial positioning coordinate group includes: Obtain the phase difference sequence of the UWB signal sources and Bluetooth MESH node signal sources in the metal-dense area; Match the phase difference sequence with a preset multipath interference template library to generate a compensation offset; Superimpose the compensation offset on a signal attenuation model to obtain a corrected propagation loss distribution map; Construct a spatial probability density field based on the corrected propagation loss distribution map and a dynamic weight coefficient; Extract a cluster of maximum value points based on the spatial probability density field to obtain an original positioning coordinate group; Perform confidence accumulation calculation on the coordinate points in the original positioning coordinate group to eliminate the coordinate points with cumulative values lower than the second threshold, and obtain the initial positioning coordinate group.

4. The system according to claim 1, characterized in that, Input the initial positioning coordinate group into a preset three-dimensional structure map model and process it using a multi-source fusion algorithm to generate a final coordinate result, including: Extract the discrete coordinate data of the initial positioning coordinate group; Map the discrete coordinate data to the preset three-dimensional structure map model to generate a three-dimensional positioning point set; Construct a state vector matrix based on the three-dimensional positioning point set; Input the state vector matrix into a Kalman filter model to obtain the covariance matrix of the positioning point variance; Judge whether the value of the covariance matrix exceeds the third threshold; In response to the value of the covariance matrix exceeding the third threshold, trigger an RFID node activation signal to obtain positioning compensation data; Perform spatial coordinate fusion on the positioning compensation data and the three-dimensional positioning point set to obtain an optimized positioning point set; Generate a final coordinate result based on the optimized positioning point set and the multi-source fusion algorithm.

5. The system according to claim 4, wherein The covariance matrix is represented by the following formula: = （2）； Among them, represents the covariance matrix of the state at the current moment, represents the state vector matrix of the state at the moment, and represents the process noise covariance matrix.

6. The system according to claim 5, characterized in that, The generating of the final coordinate result based on the optimized positioning point set and the multi-source fusion algorithm includes: Conduct parameter construction on the optimized positioning point set and the data of each signal source to obtain a positioning input set; the positioning point set contains multiple reference coordinates and confidence parameters; Adjust the weight distribution coefficient of the reference coordinates based on the gradient descent algorithm to obtain the weighted fusion positioning coordinates; Input the fusion positioning coordinates into a coordinate calibration model, and calculate the distance in combination with the positioning point set to obtain a residual matrix; Iteratively correct the confidence parameters and adjust the fusion positioning coordinates according to the residual matrix until the modulus value of the residual matrix reaches the convergence condition, and use the fusion positioning coordinates when the convergence condition is satisfied as the final coordinate result.

7. The system according to claim 1, wherein The time alignment and depth fusion of the human body dynamic trajectory, abnormal vibration alarm information, and intrusion warning information to generate a security status assessment report includes: Obtain the original timestamps of the human body dynamic trajectory and the abnormal vibration alarm information; Eliminate the device clock offset of the original timestamps through a timestamp alignment algorithm to obtain an aligned time series; Extract the trajectory acceleration frequency domain features and abnormal vibration waveform features of the original timestamps according to the aligned time series to obtain key feature parameters; Input the key feature parameters into a preset abnormal correlation analysis model to obtain an abnormal correlation coding vector; Fuse the abnormal correlation coding vector with the priority parameters of the intrusion warning information to obtain a security status assessment parameter; Input the security status assessment parameter into a pre-trained security assessment model to obtain a security status assessment report.

8. A multi-scenario positioning and intelligent warning method for industrial equipment, characterized in that, Include the following steps: Obtain the radar signal data of the industrial equipment area and denoise the radar signal; then extract features from the denoised radar signal data to obtain a signal intensity change feature vector; if the matching degree of the signal intensity change feature vector and a preset metal reflection model exceeds the first threshold, generate a human body dynamic trajectory; the metal reflection model is used to describe the characteristics of radar signals when reflected by metal objects; Obtain vibration sensing data at key positions of industrial equipment; and extract and screen the vibration sensing data based on the wavelet decomposition algorithm to generate vibration spectra in each frequency band; when the vibration spectra exceed the preset spectrum range, generate abnormal vibration alarm information; Obtain the UWB signal source and Bluetooth MESH node signal source in the metal-dense area of the industrial equipment; and perform weighted fusion on the UWB signal source and Bluetooth MESH node signal source to obtain an initial positioning coordinate group; Input the initial positioning coordinate group into a preset three-dimensional structure map model and process it using a multi-source fusion algorithm to generate a final coordinate result; If the final coordinate result falls within the boundary of a preset danger fence, generate intrusion warning information; Perform time alignment and depth fusion on the human body dynamic trajectory, abnormal vibration alarm information, and intrusion warning information to generate a safety status assessment report; Among them, if the matching degree between the signal intensity change feature vector and a preset metal reflection model exceeds a first threshold, generate a human body dynamic trajectory, including: Decompose the signal intensity change feature vector to obtain attributes including time series change frequency and amplitude distribution; Extract parameters based on the signal intensity change feature vector decomposed by a preset metal reflection model to obtain metal reflection parameters corresponding to the dimension of the feature vector; Use a dynamic threshold segmentation algorithm to judge the similarity coefficient between the metal reflection parameters and the signal intensity change feature vector to obtain a feature similarity parameter; Judge whether the feature similarity parameter exceeds the upper limit of a preset dynamic threshold interval; In response to the feature similarity parameter exceeding the upper limit of the preset dynamic threshold interval, identify the human body dynamic trajectory and construct a spatial coordinate sequence of the human body dynamic trajectory to generate the human body dynamic trajectory corresponding to the signal intensity change feature vector.

Images