Mine personnel positioning method based on ultra wide band and millimeter wave radar fusion

Abstract

The invention belongs to the technical field of intelligent mines, particularly relates to a mine personnel positioning method based on ultra-wideband and millimeter-wave radar fusion, and aims to improve the positioning precision of mine operation personnel. Comprising the steps that a portable ultra wide band positioning terminal is arranged for personnel, the personnel are preliminarily positioned through the high time resolution characteristic of ultra wide band signals, millimeter wave radar equipment is used for synchronously conducting three-dimensional space detection on the surrounding environment of the personnel, and dynamic environment information including obstacles, topographic changes and personnel movement is obtained; the central processing unit adopts a multi-source data fusion algorithm based on the combination of Kalman filtering and a deep neural network to perform joint modeling on the acquired preliminary positioning data and dynamic environment information, and dynamically corrects errors in the preliminary positioning data by analyzing complementarity and environment correlation of the two types of data, so that the positioning accuracy is improved; and the obtained high-precision positioning result is output to a monitoring platform, so that accurate tracking and dynamic management of the personnel are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent mine, in particular to a mine personnel positioning method based on ultra-wideband and millimeter wave radar fusion. BACKGROUND

[0002] The mine operation environment has the characteristics of narrow space, complex structure, dense obstacles and strong sealing, which brings great challenges to personnel positioning. The traditional satellite positioning technology is completely invalid in underground space, while the positioning methods based on wireless communication such as Wi-Fi and Bluetooth are limited by signal coverage range and stability, and are easily affected by multi-path effect and shielding caused by walls, metal equipment and other factors, so the positioning accuracy is difficult to meet the needs of safety production and emergency rescue.

[0003] Although ultra-wideband positioning is widely used in indoor positioning due to its high time resolution and anti-interference ability, it is still affected by non-line-of-sight, multi-path reflection, obstacle shielding and other factors in the complex environment of mine, and has problems such as large precision fluctuation and insufficient robustness. Although millimeter wave radar has three-dimensional space perception and certain penetration ability, it can provide the possibility for positioning error compensation, but the deep fusion of the two for high-precision positioning in mine without infrastructure is not mature. SUMMARY

[0004] The purpose of the present application is to provide a mine personnel positioning method based on ultra-wideband and millimeter wave radar fusion, aiming to improve the positioning accuracy of mine workers.

[0005] In order to achieve the above purpose, the technical scheme adopted by the present application is as follows: the present application provides a mine personnel positioning method based on ultra-wideband and millimeter wave radar fusion, the method comprising S1: providing a portable ultra-wideband positioning terminal for personnel, preliminarily positioning the personnel through the high time resolution characteristic of ultra-wideband signal, obtaining preliminary positioning data and transmitting it to the center processing unit as the basic data source for fusion positioning; S2: using millimeter wave radar equipment to synchronously detect the three-dimensional space of the personnel's surrounding environment, obtaining dynamic environment information including obstacles, terrain changes and personnel movement, and transmitting it to the center processing unit to provide environmental feature information for data fusion; S3: the center processing unit adopts a multi-source data fusion algorithm based on Kalman filtering and deep neural network to jointly model the preliminary positioning data obtained in step S1 and the dynamic environment information obtained in step S2, dynamically corrects the errors in the preliminary positioning data by analyzing the complementarity and environmental correlation of the two types of data, and obtains high-precision positioning results; S4: outputting the high-precision positioning results obtained in step S3 to the monitoring platform to realize accurate tracking and dynamic management of personnel.

[0006] In step S1, the preliminary positioning of the personnel by the high time resolution characteristic of the ultra-wideband signal includes: through the ultra-wideband wireless signal communication between the portable ultra-wideband positioning terminal equipped by the personnel and the preset plurality of ultra-wideband anchors, the time difference of arrival of the signal is obtained ; the distance from the ultra-wideband positioning terminal to each ultra-wideband anchor is estimated according to the time difference of arrival , a positioning equation set is established and solved to obtain the preliminary position estimation of the personnel; the positioning equation set is as follows: Among them, the ultra-wideband anchor refers to a device fixedly arranged in the mine, used for receiving the signal transmitted by the ultra-wideband positioning terminal and assisting in positioning; ; The time of arrival of the signal received by the ultra-wideband anchor is denoted as ; c is the speed of light, is the error introduced by the noise environment interference; ; ; are the three-dimensional coordinates of the position of the portable ultra-wideband positioning terminal and the preset plurality of ultra-wideband anchors respectively, denotes the initial time of the ultra-wideband signal transmitted by the portable ultra-wideband positioning terminal.

[0007] The positioning equation set is solved by the least square method to obtain the preliminary position estimation of the personnel ; wherein, .

[0008] In step S2, the three-dimensional space detection of the environment around the personnel is performed synchronously by using the millimeter wave radar device, which includes: the millimeter wave radar device transmits a frequency-modulated continuous wave and receives the reflection signal of the target point in the space, and the distance of the target point is calculated according to the round-trip time delay of the echo signal; the horizontal angle and the pitch angle of the target point are obtained in combination with the angle resolution capability of the radar, the three-dimensional coordinates of the target point are calculated, and a point cloud set is formed to describe the structure and dynamic change of the environment around the personnel.

[0009] In step S3, a state vector of the personnel at a certain moment is defined, the state vector includes the three-dimensional position coordinates and the velocity component of the personnel, the state vector is predicted and updated by using the Kalman filter to obtain the state value after the Kalman filter processing; the deep neural network is introduced to learn and correct the residual error of the state value after the Kalman filter processing to obtain the high-precision positioning result.

[0010] When the state is updated by using the Kalman filter, the observation vector used includes the distance from the ultra-wideband positioning terminal to each ultra-wideband anchor and the three-dimensional point cloud coordinates obtained by the millimeter wave radar device; wherein, the observation vector refers to a multi-source observation data set used for correcting the predicted state of the Kalman filter. ​​

[0011] The observation vector is constructed by communication data of the ultra-wideband positioning terminal and the ultra-wideband anchor point, and three-dimensional space detection data of the millimeter wave radar device.

[0012] The deep neural network is a gated recurrent unit neural network.

[0013] In step S4, the motion trajectory of the personnel is generated according to the high-precision positioning results at continuous time points; and it is judged whether the personnel enters a high-risk area in combination with preset dangerous area information, and an alarm is issued when the personnel enters the high-risk area.

[0014] Step S4 further includes generating an optimal path for emergency rescue according to the high-precision positioning results of the personnel and the position of the accident point.

[0015] Compared with the prior art, the application has the following beneficial effects: 1. The mine personnel positioning method based on ultra-wideband and millimeter wave radar fusion provided by the application can effectively overcome the interference of factors such as multipath, shielding, non-line-of-sight and the like in the complex mine environment on single ultra-wideband positioning by dynamically correcting errors by using the algorithm combining Kalman filtering and the gated recurrent unit neural network, and realize high-precision, stable and reliable personnel positioning.

[0016] 2. The positioning can be completed only by the portable ultra-wideband positioning terminal and the millimeter wave radar device worn by the personnel, without the need of additional ground base stations, beacons and other fixed facilities, which greatly reduces the deployment and maintenance costs, and enhances the adaptability and scalability in the mine structure variable and space limited scenarios.

[0017] 3. The output high-precision positioning result can generate a personnel motion trajectory, realize risk warning in combination with dangerous areas, and also provide optimal path planning for emergency rescue, deeply link mine safety monitoring, dynamic management and emergency response and other businesses, and significantly improve the safety guarantee level and management efficiency of mine operation. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 FIG. 1 is a schematic diagram of a mine personnel positioning method based on ultra-wideband and millimeter wave radar fusion provided by an embodiment of the application; Figure 2 FIG. 2 is a framework diagram of the mine personnel positioning method based on ultra-wideband and millimeter wave radar fusion provided by an embodiment of the application. DETAILED DESCRIPTION

[0019] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by a person of ordinary skill in the art without creative effort are within the scope of the present application.

[0020] In the embodiments of the present application, the terms “comprising”, “containing” or any other variants thereof are intended to cover the non-exclusive inclusion, so that the process, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or includes the elements inherent to such process, article or device. Without more limitation, the element defined by the sentence “comprising a…” does not exclude the presence of another identical element in the process, article or device comprising the element.

[0021] In the embodiments of the present application, the words such as “exemplary” or “for example” are used to represent an example, illustration or description. Any embodiment or design scheme described as “exemplary” or “for example” in the embodiments of the present application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. Rather, the words such as “exemplary” or “for example” are intended to present the relevant concept in a specific manner.

[0022] The embodiments of the present application provide a mine personnel positioning method based on fusion of ultra-wideband and millimeter wave radar, and exemplary Figure 1 and Figure 2 The method comprises the following steps: S1: A portable ultra-wideband positioning terminal is provided for personnel, a preliminary positioning of the personnel is performed through the high time resolution characteristic of the ultra-wideband signal, preliminary positioning data are obtained and transmitted to a central processing unit as a basic data source for fusion positioning.

[0023] The ultra-wideband (UWB) positioning terminal refers to a portable device capable of transmitting and receiving ultra-wideband wireless signals for realizing personnel position information acquisition.

[0024] In step S1, the preliminary positioning of the personnel through the high time resolution characteristic of the ultra-wideband signal comprises: the portable ultra-wideband positioning terminal provided for the personnel communicates with a plurality of preset ultra-wideband anchor points through ultra-wideband wireless signals, time difference of arrival (TDOA) of the signals is obtained , the distances of the ultra-wideband positioning terminal to each ultra-wideband anchor point are estimated according to the TDOA , a positioning equation set is established and solved to obtain a preliminary position estimate of the personnel; the positioning equation set is as follows: The ultra-wideband anchor point refers to a device fixedly arranged in the mine and used for receiving a signal transmitted by an ultra-wideband positioning terminal and assisting in positioning. ; The arrival time of the signal received by the ultra-wideband anchor point is denoted as T. ; c is the speed of light, The error introduced by noise environmental interference is denoted as e. ; The initial time at which the portable ultra-wideband positioning terminal transmits the ultra-wideband signal is denoted as t. ; The three-dimensional coordinates of the portable ultra-wideband positioning terminal and the preset plurality of ultra-wideband anchor points are denoted as (x, y, z) and (x', y', z'), respectively. The three-dimensional coordinates of the portable ultra-wideband positioning terminal and the preset plurality of ultra-wideband anchor points are denoted as (x, y, z) and (x', y', z'), respectively. The initial time at which the portable ultra-wideband positioning terminal transmits the ultra-wideband signal is denoted as t.

[0025] It should be understood that, in actual mine environment deployment, the number N of ultra-wideband anchor points needs to be adaptively set according to the length, width and complexity of the roadway, and generally should be not less than 3 to meet the basic requirements of three-dimensional positioning. For complex areas with multiple bends and multiple branches, the anchor point density needs to be appropriately increased to improve the reliability of the preliminary positioning data.

[0026] As a possible implementation manner, the least square method is used to solve the positioning equation set to obtain the preliminary position estimate of the personnel ; wherein, When the preliminary position estimate of the personnel is obtained, the distance measurement data, the time stamp original observation information of each anchor point are transmitted to the central processing unit together, to provide data support for environment perception and multi-source fusion correction.

[0027] For example, in the data transmission process, an anti-interference wireless transmission protocol is used to ensure that the preliminary positioning data can be completely and real-timely transmitted to the central processing unit under the strong electromagnetic interference environment of the mine, the transmission frequency is kept synchronous with the detection frequency of the millimeter wave radar, and is generally set to 10-20 Hz to ensure the time consistency of the two types of data.

[0028] S2: The three-dimensional space around the personnel is synchronously detected by using the millimeter wave radar device to obtain dynamic environment information including obstacles, terrain changes and personnel movement, and the information is transmitted to the central processing unit to provide environment feature information for data fusion.

[0029] The millimeter wave radar device refers to a device capable of transmitting and receiving millimeter wave signals and used for detecting the three-dimensional space environment and dynamic targets around the mine.

[0030] In step S2, the three-dimensional space detection of the surrounding environment of the person by the millimeter wave radar device in synchronization includes: the millimeter wave radar device transmits a frequency-modulated continuous wave and receives a reflection signal of a target point in space, and the distance of the target point is calculated according to the round-trip time delay of the echo signal; the horizontal angle and the pitch angle of the target point are obtained in combination with the angle resolution capability of the radar, the three-dimensional coordinates of the target point are calculated, and a point cloud set is formed to describe the structure and dynamic changes of the surrounding environment of the person.

[0031] More specifically, on the basis of obtaining the preliminary position estimation of the person , the three-dimensional space detection and dynamic perception of the surrounding environment of the person by the millimeter wave radar device. The millimeter wave radar is installed at a key node of the device mine or carried by the person, and the coordinates are , the millimeter wave radar transmits a frequency-modulated continuous wave with a frequency of , scans different directions in space, and the reflection signal of a target point in space is received by the millimeter wave radar. According to the measurement principle of the millimeter wave radar, the distance is calculated by the round-trip time delay of the echo signal: wherein is the speed of light, represents all echo points detected in one scan of the millimeter wave radar.

[0032] In combination with the angle resolution capability of the millimeter wave radar, suppose that the horizontal angle and the pitch angle of a certain echo point are , then the three-dimensional coordinates of the th point cloud in space are: The point cloud set obtained by each scan of the millimeter wave radar records the three-dimensional environmental structure and dynamic changes of the surrounding of the person. By comparing the point clouds at adjacent time points, the movement of the person, the dynamic changes of obstacles, and the spatial changes can be analyzed.

[0033] It should be understood that the detection range of the millimeter wave radar needs to cover a space area of 5-10 meters around the person to ensure that the obstacle information that may affect the ultra-wideband positioning can be completely captured. As a possible implementation manner, the collected point cloud data is preprocessed, including removing noise points, splicing continuous frame point clouds, and separating static obstacles such as walls, equipment, and temporary accumulations and dynamic targets such as persons and mobile devices from the point cloud set through ground extraction, obstacle clustering, and other algorithms, to provide targeted environmental features for subsequent error correction.

[0034] S3: The center processing unit adopts a multi-source data fusion algorithm based on Kalman filtering combined with a deep neural network to jointly model the preliminary positioning data obtained in step S1 and the dynamic environment information obtained in step S2, dynamically correct the errors in the preliminary positioning data by analyzing the complementarity and environment correlation of the two types of data, and obtain a high-precision positioning result.

[0035] In step S3, a state vector of a person at a certain time is defined, the state vector including three-dimensional position coordinates and velocity components of the person, Kalman filtering is used to predict and update the state vector to obtain a state value after Kalman filtering processing; a deep neural network is introduced to learn and correct the residual error of the state value after Kalman filtering processing to obtain the high-precision positioning result.

[0036] For example, the state vector of the person at time k is ; wherein, is a three-dimensional position coordinate, is a velocity component, the state of the person is predicted using classical Kalman filtering, and the state transition equation is: ; wherein, is a state prediction value at time k, is a state transition matrix, is process noise, is a process noise covariance matrix; the state covariance prediction is: wherein, is a prediction error covariance, is an error covariance after updating in the previous step.

[0037] When the state is updated using Kalman filtering, the observation vector includes the distance from the ultra-wideband positioning terminal to each ultra-wideband anchor point and the three-dimensional point cloud coordinates obtained by the millimeter wave radar device; wherein the observation vector refers to a multi-source observation data set used to correct the predicted state of Kalman filtering. The observation vector is jointly constructed through the communication data between the ultra-wideband positioning terminal and the ultra-wideband anchor points and the three-dimensional space detection data of the millimeter wave radar device.

[0038] For example, multi-source observation and measurement updating is performed, and the fused observation vector is defined as wherein, is the distance from the person to the i-th anchor point measured by the UWB at time k, is the j-th three-dimensional point cloud coordinate obtained by the millimeter wave radar; the measurement model is wherein, To measure the matrix, the spatial geometric relationship of the UWB distance observation and the millimeter wave radar point cloud observation is derived, The observation noise is; the Kalman gain matrix is calculated as: Where R is the measurement noise covariance matrix; the state is updated using the observation: And the state covariance is updated In the formula, I represents a unit matrix, and the dimension of the unit matrix I is consistent with the dimension of the personnel state vector .

[0039] As a possible implementation manner, the deep neural network is a gated recurrent unit neural network. The gated recurrent unit neural network is introduced to correct the residual error of the state prediction value, and a correction network is defined as , the input is the current Kalman filter prediction state and the observation vector , and the output is a nonlinear residual correction value : The final fusion positioning result is Wherein is the high-precision positioning output after Kalman filtering and neural network correction.

[0040] The gated recurrent unit parameters are trained end to end by collecting historical data, and the high-precision position is manually labeled as , and the positioning error loss is minimized: The training data set of the gated recurrent unit neural network covers ultra-wideband positioning data, millimeter wave radar point cloud data and corresponding high-precision true value positions under different scenarios (such as straight lanes, curved lanes, and equipment-intensive areas) in the mine and different personnel motion states (still, walking, and walking with tools). The generalization ability of the model is monitored in real time through the validation set to avoid overfitting.

[0041] S4: Output the high-precision positioning result obtained in step S3 to the monitoring platform to realize accurate tracking and dynamic management of personnel.

[0042] In step S4, the motion trajectory of the personnel is generated according to the high-precision positioning result at consecutive time points; in combination with the preset dangerous area information, it is judged whether the personnel enters a high-risk area, and an alarm is issued when the personnel enters the high-risk area.

[0043] Exemplarily, the final fusion positioning output is This indicates that personnel are at all times Three-dimensional spatial coordinates; positioning results at consecutive time points. The movement trajectory of the workers: The precise tracking and monitoring system can display the movement trajectory of personnel in real time. Combining mine map with digital twin mine model The system dynamically maps the position, movement path, and dwell area of ​​personnel in the underground space; based on the trajectory differential, the personnel's velocity and acceleration can be calculated in real time: in, This represents the sampling time interval.

[0044] Using trajectory points and danger zones Based on spatial relationships, the formula for determining whether personnel have entered a high-risk area is as follows: when , When the preset safe distance threshold is reached, the system will automatically issue an alarm and record the event.

[0045] Step S4 also includes: generating the optimal emergency rescue route based on the high-precision positioning results of personnel and the location of the accident point.

[0046] As one possible approach, linking real-time location results with the mine management platform can enable personnel distribution statistics: counting the number of personnel in each work area at any given time. in, As an index function, determine the first Is the person in the first partition Inside; Location results It can provide real-time optimal route recommendations for emergency rescue, assuming the accident point is... Shortest rescue route Dijkstra's algorithm can be used to... The format is automatically generated.

[0047] The application innovatively integrates ultra-wideband positioning and millimeter wave radar environmental perception technology, and constructs a multi-source data dynamic error compensation and adaptive correction mechanism. Although the ultra-wideband positioning is easily affected by multipath, non-line-of-sight and shielding in a complex environment, leading to large positioning error, the application utilizes the high-resolution perception ability of millimeter wave radar for three-dimensional space obstacles and dynamic environment, combines Kalman filtering and deep neural network intelligent fusion algorithm, and realizes dynamic correction and high-robustness accurate positioning of personnel position. Compared with existing single ultra-wideband, Wi-Fi or Bluetooth positioning technology, the application can effectively resist multipath and shielding interference in extreme environments such as mines and tunnels, achieve centimeter-level continuous, high-precision and low-delay personnel positioning, and significantly improve the accuracy, stability and applicability of the system.

[0048] The overall design of the positioning system of the application does not need to rely on traditional ground base stations, beacons or other fixed infrastructure, and only needs the portable ultra-wideband positioning terminal and millimeter wave radar equipment worn by the operating personnel to realize high-precision personnel positioning and real-time environmental perception. Compared with the existing technology which needs to lay and maintain a large number of base stations, beacons, reference points and other infrastructure, the application greatly reduces the deployment and maintenance cost, and the system can be flexibly arranged according to the actual needs on site, greatly enhances the adaptability and scalability in the scenes of mine structure change, space limitation and complex environment, and is especially suitable for mines and underground spaces with strong mobility, temporary operation or limited infrastructure layout.

[0049] The high-precision, continuous personnel three-dimensional motion trajectory output by the application can not only be directly integrated into the mine safety monitoring and intelligent management platform, but also can be deeply linked with many types of business systems such as operation process, emergency rescue and risk control. The system can real-time identify personnel distribution and dynamic state, automatically identify abnormal behavior and dangerous area intrusion, and can provide accurate and real-time position information support for emergency rescue path planning and personnel evacuation. Compared with the existing passive monitoring and low-resolution personnel positioning method, the application significantly improves the safety guarantee capability, emergency response efficiency and management intelligent level of mine operation, lays a solid technical foundation for the construction of smart mine and the digital transformation of mine, and has a broad application prospect.

[0050] In the description of the present application, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0051] The above is only a specific implementation of the application, but the protection scope of the application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical scope disclosed by the application, which should be covered within the protection scope of the application. Therefore, the protection scope of the application should be subject to the protection scope of the claims.

Claims

1. A method for locating personnel in mines based on the fusion of ultra-wideband and millimeter-wave radar, characterized in that, include: S1: Equip personnel with portable ultra-wideband positioning terminals to perform preliminary positioning of personnel using the high time resolution characteristics of ultra-wideband signals, acquire preliminary positioning data and transmit it to the central processing unit as the basic data source for fusion positioning. S2: Simultaneously use millimeter-wave radar equipment to perform three-dimensional spatial detection of the surrounding environment of personnel, acquire dynamic environmental information including obstacles, terrain changes and personnel movement, and transmit it to the central processing unit to provide environmental feature information for data fusion; S3: The central processing unit adopts a multi-source data fusion algorithm based on Kalman filtering and deep neural network to jointly model the preliminary positioning data obtained in step S1 and the dynamic environmental information obtained in step S2. By analyzing the complementarity of the two types of data and the correlation with the environment, the error in the preliminary positioning data is dynamically corrected to obtain a high-precision positioning result. S4: Output the high-precision positioning results obtained in step S3 to the monitoring platform to achieve accurate tracking and dynamic management of personnel.

2. The method for locating personnel in mines based on the fusion of ultra-wideband and millimeter-wave radar according to claim 1, characterized in that, In step S1, preliminary positioning of personnel using the high time resolution characteristics of ultra-wideband signals includes: communicating with multiple preset ultra-wideband anchor points via ultra-wideband wireless signals using a portable ultra-wideband positioning terminal equipped with the personnel, and obtaining the signal arrival time difference. The distance from the ultra-wideband positioning terminal to each ultra-wideband anchor point is estimated based on the time difference of arrival. A system of positioning equations was established and solved to obtain a preliminary estimate of the personnel's location. The system of positioning equations is as follows: Among them, ultra-wideband anchor points refer to devices that are fixedly installed in the mine to receive signals transmitted by ultra-wideband positioning terminals and assist in positioning. ; This refers to the arrival time of the signal received at the ultra-wideband anchor point; c is the speed of light. Errors introduced by noise interference; ; ( ); and These are the three-dimensional coordinates of the portable ultra-wideband positioning terminal and the positions of multiple preset ultra-wideband anchor points, respectively. This refers to the initial time when a portable ultra-wideband positioning terminal transmits an ultra-wideband signal.

3. The method for locating personnel in mines based on the fusion of ultra-wideband and millimeter-wave radar according to claim 2, characterized in that, The least squares method was used to solve the positioning equations to obtain a preliminary estimate of the personnel's location. ;in, .

4. The method for locating personnel in mines based on the fusion of ultra-wideband and millimeter-wave radar according to claim 1, characterized in that, In step S2, the three-dimensional spatial detection of the surrounding environment of personnel using millimeter-wave radar equipment includes: the millimeter-wave radar equipment transmits frequency-modulated continuous waves and receives the reflected signals of target points in space, calculates the distance of the target points based on the round-trip time delay of the echo signals; obtains the horizontal angle and elevation angle of the target points by combining the radar's angle resolution capability, calculates the three-dimensional coordinates of the target points, and forms a point cloud set to describe the structure and dynamic changes of the surrounding environment of personnel.

5. A method for locating personnel in mines based on the fusion of ultra-wideband and millimeter-wave radar according to claim 1, characterized in that, In step S3, a state vector of a person at a certain moment is defined. This state vector includes the person's three-dimensional position coordinates and velocity components. The state vector is predicted and updated using Kalman filtering to obtain the state value after Kalman filtering. A deep neural network is introduced to perform residual learning correction on the state value after Kalman filtering to obtain the high-precision positioning result.

6. A method for locating personnel in mines based on the fusion of ultra-wideband and millimeter-wave radar according to claim 5, characterized in that, When using Kalman filtering for state updates, the observation vectors used include the distance from the ultra-wideband positioning terminal to each ultra-wideband anchor point and the three-dimensional point cloud coordinates obtained by the millimeter-wave radar equipment; where the observation vector refers to the set of multi-source observation data used to correct the Kalman filter predicted state.

7. A method for locating personnel in mines based on the fusion of ultra-wideband and millimeter-wave radar according to claim 6, characterized in that, The observation vector is jointly constructed using communication data between the ultra-wideband positioning terminal and the ultra-wideband anchor point, and three-dimensional spatial detection data from the millimeter-wave radar equipment.

8. A method for locating personnel in mines based on the fusion of ultra-wideband and millimeter-wave radar according to claim 5, characterized in that, Deep neural networks are gated recurrent unit neural networks.

9. A method for locating personnel in mines based on the fusion of ultra-wideband and millimeter-wave radar according to claim 1, characterized in that, In step S4, the movement trajectory of the personnel is generated based on the high-precision positioning results at continuous times; combined with the preset danger zone information, it is determined whether the personnel have entered the high-risk zone, and an alarm is issued when the personnel enter the high-risk zone.

10. A method for locating personnel in mines based on the fusion of ultra-wideband and millimeter-wave radar according to claim 9, characterized in that, Step S4 also includes: generating the optimal emergency rescue route based on the high-precision positioning results of personnel and the location of the accident point.

Images