Automatic driving mine vehicle dynamic positioning switching control method and system

Abstract

The invention relates to a dynamic positioning switching control method for an automatic driving mine vehicle, and the method comprises the following steps: carrying out the control of the dynamic positioning switching of the automatic driving mine vehicle based on the dust concentration change rate data and the visibility change rate data of a historical set time period, and the first dust concentration change rate data and the first visibility change rate data of a current region; predicting second dust concentration change rate data and second visibility change rate data in a future set time period; and determining a risk level of the current area and a trigger condition corresponding to the risk level based on a preset area threshold table, and comparing the second dust concentration change rate data and the second visibility change rate data with the trigger condition. According to the method, the pre-switching process can be triggered before the sudden change of the dust concentration occurs, precious buffer time is gained for a system, mode conversion from traditional passive response to active pre-judgment and adjustment is achieved, the positioning interruption and out-of-control risks of a vehicle in the scene of the sudden change of the dust are reduced, and the active safety capacity of driving operation is improved.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving technology for intelligent mining vehicles, specifically to a control method and system for dynamic positioning switching of autonomous mining vehicles. Background Technology

[0002] The positioning system is the core of autonomous driving vehicles in mines, enabling them to perceive the environment and navigate autonomously. In dusty environments, the performance of optical sensors such as lidar degrades drastically. Therefore, it is usually necessary to introduce heterogeneous sensors such as millimeter-wave radar and UWB to form a redundant positioning system, which switches to the backup system when the main sensor fails, in order to ensure the basic safety of the vehicle.

[0003] Currently, the industry commonly uses a threshold-triggered direct switching control method. This method monitors the signal quality indicators (such as point cloud density and confidence level) of the primary positioning system (e.g., LiDAR) in real time. When these indicators fall below a preset threshold, the system directly cuts off the output of the primary positioning system and immediately switches to the backup positioning system. However, this method requires waiting for the sensor signal to deteriorate to the threshold before initiating the switch, making it impossible to predict and prepare for short-term dust deterioration trends. This results in an inevitable response delay when dust concentration changes abruptly, and vehicles face the risk of positioning interruption during the delay period. Furthermore, due to the primary and backup positioning... The working principle and output characteristics of the positioning system are fundamentally different. This direct and discontinuous "hard switching" will cause a significant step jump in the final positioning output. This jump is transmitted to the vehicle control system and will cause the vehicle to shake or jerk. In narrow mine tunnels, it can easily lead to scraping or even loss of control, seriously threatening the safety and stability of operations. Therefore, there is an urgent need for a dynamic positioning switching control method that can achieve forward-looking triggering and smooth, imperceptible switching, so as to realize the transformation from passive response to active prediction, avoid the risk of positioning failure in advance, and eliminate the output jump during the main and backup positioning switching process to ensure the stability of vehicle control. Summary of the Invention

[0004] This application provides a control method and system for dynamic positioning switching of autonomous mining vehicles to solve the above problems.

[0005] In a first aspect, embodiments of this application provide a control method for dynamic positioning switching of an autonomous mining vehicle, comprising the following steps: Based on historical dust concentration change rate data and visibility change rate data for a set time period, as well as the first dust concentration change rate data and the first visibility change rate data for the current area, predict the second dust concentration change rate data and the second visibility change rate data for a set time period in the future. Based on the pre-switching window control, the main positioning mode and the backup positioning mode are operated in parallel. The first dust concentration data and the first visibility data at the current moment are obtained. Based on the first dust concentration data and the first visibility data, the first weight parameter of the main positioning mode and the second weight parameter of the backup positioning mode are calculated. Based on the first weight parameter and the second weight parameter, the first output data of the main positioning mode and the second output data of the backup positioning mode are weighted, fused, smoothed and weighted constrained. The system monitors the output difference between the first output data and the second output data, as well as the residual external parameters of the backup positioning mode. When the output difference and the residual external parameters meet the preset switching conditions, the system performs a switch from the primary positioning mode to the backup positioning mode at the second time point. If the preset switching conditions are not met, the system maintains the operation of the primary positioning mode and controls the vehicle to travel at the preset final speed limit.

[0006] In conjunction with the first aspect, in one implementation, the preset area threshold table includes the risk level of the current area, triggering conditions, and a basic speed limit. The risk level of the current area includes low-risk areas, medium-risk areas, and high-risk areas, wherein: The triggering conditions for the high-risk area include: the second dust concentration change rate data is greater than or equal to the first set benchmark threshold, and the second visibility change rate data is less than or equal to the second set benchmark threshold; the basic speed limit for the high-risk area is the first set speed value. The triggering conditions for the medium-risk area include: the second dust concentration change rate data is greater than or equal to the third set benchmark threshold, and the second visibility change rate data is less than or equal to the fourth set benchmark threshold; the basic speed limit for the medium-risk area is the second set speed value. The triggering conditions for the low-risk area include: the second dust concentration change rate data is greater than or equal to the fifth set benchmark threshold, and the second visibility change rate data is less than or equal to the sixth set benchmark threshold; the basic speed limit for the low-risk area is the third set speed value.

[0007] In conjunction with the first aspect, in one implementation, the first weighting parameter is calculated using the following formula: ; in, This is the first dust concentration data. As the basic weight for dust concentration, This is the dust concentration sensitivity coefficient. As the basic weight for visibility, Visibility sensitivity coefficient This is the primary visibility data. Ideal visibility for mine tunnel design The lower limit of the weight preset for the primary positioning mode. The preset weight limit for the primary positioning mode; The second weighting parameter is calculated using the following formula: ; in, This is the first weight parameter.

[0008] In conjunction with the first aspect, in one implementation, the smoothing process is specifically implemented using an exponential smoothing algorithm, wherein the smoothing coefficient α of the exponential smoothing algorithm ranges from 0.1 to 0.3. The weight constraint processing includes constraining the gradient of the change of the first weight parameter and the second weight parameter within the pre-switching window, such that the absolute value of the change in weight between adjacent calculation cycles does not exceed 0.15.

[0009] In conjunction with the first aspect, in one implementation, the preset switching condition is: within a set number of sampling frames, the output difference between the first output data and the second output data is less than a first set difference value, and the residual extrinsic parameter of the backup positioning mode is less than a second set difference value.

[0010] In conjunction with the first aspect, in one implementation, the step of generating a trigger signal and initiating the pre-switching window at the first time point also includes: The comprehensive risk value is calculated based on the first dust concentration data, the first visibility data, the second dust concentration change rate data, and the second visibility change rate data. The comprehensive risk value is weighted and adjusted to obtain the final risk assessment value; The warning level for the current area is determined based on the final risk assessment value. The warning levels include high-risk warning level, medium-risk warning level and low-risk warning level. Based on the aforementioned warning level, the corresponding risk strategy will be implemented.

[0011] In conjunction with the first aspect, in one implementation, determining the warning level of the current area based on the final risk assessment value, and executing the corresponding risk strategy based on the warning level, specifically includes: If the final risk assessment value is greater than or equal to the first assessment threshold and less than the second assessment threshold, it is determined to be a low-risk warning level, and a low-risk strategy is executed. The low-risk strategy includes: limiting the vehicle speed to the final speed limit determined based on the risk level and low-risk warning level of the current area, maintaining the parallel operation of the main positioning mode and the backup positioning mode, and performing a first-level safety anchoring on the vehicle. The first-level safety anchoring includes limiting the output power of the drive system according to a set ratio and triggering the vehicle parking brake system to enter a preload state. If the final risk assessment value is greater than or equal to the second assessment threshold and less than the third assessment threshold, it is determined to be a medium-risk warning level, and a medium-risk strategy is implemented. The medium-risk strategy includes: limiting the vehicle speed to the final speed limit determined based on the risk level of the current area and the medium-risk warning level, forcibly maintaining the parallel operation of the primary positioning mode and the backup positioning mode and prohibiting mode switching, and implementing a second-level safety anchor for the vehicle, which includes cutting off the driving power and implementing the parking brake. If the final risk assessment value is greater than or equal to the third assessment threshold, it is determined to be a high-risk warning level and a high-risk strategy is executed. The high-risk strategy includes: immediately implementing a third-level safety anchor for the vehicle and / or controlling the vehicle to the final speed limit determined based on the risk level and high-risk warning level of the current area, and evacuating along a preset emergency path. The third-level safety anchor includes cutting off all unsafe power and locking all wheel brakes. If the risk level is already high, and the final risk assessment value is greater than or equal to the second assessment threshold, and less than or equal to the difference between the third assessment threshold and the set delay amount, then the risk level will be downgraded to medium risk and a medium risk strategy will be implemented. If the risk level is already at the medium risk warning level, and the final risk assessment value is greater than or equal to the first assessment threshold, and less than or equal to the difference between the second assessment threshold and the set delay amount, then the risk level will be downgraded to the low risk warning level, and a low risk strategy will be implemented.

[0012] In conjunction with the first aspect, in one implementation, the final speed limit is determined by the following steps: Determine the visibility scaling factor based on the first visibility data; The level of suppression coefficient is determined based on the current warning level of the region; The corresponding basic speed limit is determined based on the current risk level of the area. Based on the basic speed limit, visibility scaling factor, and level suppression factor, calculate the preliminary speed limit value; The initial speed limit is compared with the preset minimum safe speed, and the maximum value is taken as the final speed limit.

[0013] In conjunction with the first aspect, in one implementation, a self-learning optimization step is also included, which includes autonomous iterative optimization of the parameters in the first weight parameter calculation formula, and / or the parameters and / or smoothing coefficients involved in calculating the comprehensive risk value, based on the performance data accumulated during operation used to characterize positioning error, control smoothness, and system energy consumption.

[0014] Secondly, embodiments of this application provide a system for a control method of dynamic positioning switching based on autonomous driving mining vehicles, comprising: The data acquisition module is configured to acquire dust concentration change rate data and visibility change rate data for a set historical time period, first dust concentration change rate data and first visibility change rate data for the current area, and first dust concentration data and first visibility data for the current moment. The prediction module is configured to predict the second dust concentration change rate data and the second visibility change rate data for a future set time period; The processing module is configured to determine the risk level of the current area and the corresponding triggering conditions based on a preset area threshold table; compare the second dust concentration change rate data and the second visibility change rate data with the triggering conditions; if either change rate data reaches or exceeds the triggering condition corresponding to the risk level of the current area, a trigger signal is generated at the first time point and a pre-switching window is initiated; control the main positioning mode and the backup positioning mode to run in parallel based on the pre-switching window; simultaneously calculate the first weight parameter of the main positioning mode and the second weight parameter of the backup positioning mode based on the first dust concentration data and the first visibility data; perform weighted fusion, smoothing, and weight constraint processing on the first output data of the main positioning mode and the second output data of the backup positioning mode based on the first weight parameter and the second weight parameter; monitor the output difference between the first output data and the second output data, as well as the extrinsic parameter residual of the backup positioning mode; when the output difference and the extrinsic parameter residual meet the preset switching conditions, execute the switch from the main positioning mode to the backup positioning mode at the second time point; if the preset switching conditions are not met, maintain the operation of the main positioning mode and control the vehicle to travel at the preset final speed limit.

[0015] The beneficial effects of the technical solutions provided in this application include: 1. By predicting the dust concentration change rate and visibility change data for a set future time period, and determining the risk level of the current area based on a preset regional threshold table, as well as the triggering conditions corresponding to the risk level, this method can trigger the pre-switching process before a sudden change in dust concentration occurs, thus gaining valuable buffer time for the system. This achieves a shift from traditional passive response to proactive prediction and adjustment, significantly reducing the risk of positioning interruption and loss of control faced by vehicles in scenarios with sudden changes in dust concentration, and improving the proactive safety capability of driving operations from the source.

[0016] 2. By introducing a pre-switching window, during the parallel operation of the primary and backup positioning modes, a comprehensive approach combining dynamic weight calculation, weighted fusion, output smoothing, and weight gradient constraints is adopted to generate continuous and stable fused positioning output. This approach fundamentally avoids the step jumps caused by traditional hard switching, effectively suppresses vehicle driving vibration and jerking, and ensures the smoothness and safety of the operation.

[0017] 3. By constructing an early warning level assessment method and a graded response mechanism for the final risk assessment value, the system achieves precise and intelligent emergency control. It also introduces a set delay amount to avoid state jitter, enabling the system to accurately pre-set differentiated response strategies for low, medium, and high risk warning levels. From low-speed warning and forced parallel operation to emergency anchoring and autonomous evacuation, a progressively increasing safety response ladder is formed. This ensures rapid safety response in high-risk warning scenarios while avoiding excessive intervention in low-risk warning scenarios, thereby significantly improving the continuity of mine vehicle operations and operational efficiency while ensuring the bottom line of safety. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram illustrating the main steps of the present invention; Figure 2 This is a flowchart illustrating the key timing sequence when the pre-switching window is started in this invention. Detailed Implementation

[0020] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0021] Example 1: Please see Figure 1 and Figure 2 This application provides a control method for dynamic positioning switching of an autonomous mining vehicle, comprising the following steps: S1. Based on historical dust concentration change rate data and visibility change rate data for a set time period, as well as the first dust concentration change rate data and the first visibility change rate data for the current area, predict the second dust concentration change rate data and the second visibility change rate data for a set time period in the future. In the specific implementation process, to ensure the reliability of the system's basic data, necessary system initialization is required, which includes: Time synchronization and sensor calibration: High-precision time synchronization protocols (such as PTP / NTP) are adopted to ensure the consistency of data timestamps. At the same time, the external parameters of key sensors (such as IMU, radar, and camera) need to be initially calibrated and periodically calibrated online to provide an accurate data foundation for the core positioning fusion algorithm.

[0022] The dust concentration change rate data and visibility change rate data are directly acquired by the vehicle-mounted dust concentration sensor and vision sensor (such as an industrial camera). The dust concentration sensor and vision sensor synchronously collect real-time dust concentration raw signals and high-definition image raw data in the mine roadway at a preset fixed frequency. For the acquisition of visibility information, the system analyzes the collected image raw data in real time through the built-in image processing algorithm. The algorithm establishes a mapping relationship between key indicators such as edge sharpness, overall contrast and texture feature attenuation of the image and the physical visibility value, thereby indirectly and quantitatively estimating the visibility of the current environment. The dust concentration and visibility change rate data are obtained by the system through time series difference calculation of continuous concentration values ​​and visibility values.

[0023] The second dust concentration change rate data and the second visibility change rate data are predicted for a future set time period. The prediction is made by using a short time series prediction model. Based on the historical data of dust concentration and visibility change rate in the past set time period (preferably 30 seconds in this embodiment), the model predicts the change trend in the future set time period (preferably 10 seconds in this embodiment), and outputs the second dust concentration change rate data and the second visibility change rate data.

[0024] Short-term series prediction models can be Kalman filters or lightweight long short-term memory networks. The lightweight long short-term memory network model needs to be pre-trained on a dataset containing a large number of typical working conditions in the mining area (such as normal ventilation, post-blasting, and transportation dust) to learn the complex evolution of the underground dust environment. To ensure the long-term reliability of the prediction, an online Kalman filter can be introduced to correct and smooth the initial prediction results of the lightweight long short-term memory network model in real time, thereby effectively combating interference such as underground ventilation disturbances and sudden dust from equipment, and improving the prediction accuracy.

[0025] S2. Determine the risk level of the current area and the triggering conditions corresponding to the risk level based on the preset area threshold table. Compare the triggering conditions with the second dust concentration change rate data and the second visibility change rate data. If any change rate data reaches or exceeds the triggering conditions corresponding to the risk level of the current area, generate a trigger signal at the first time point and start the pre-switching window. S201: The regional threshold table is a data structure (such as a database or configuration file) pre-installed in the system. It establishes a mapping relationship between different roadway locations and safety parameters in the mine's electronic map. The system obtains the vehicle's absolute position or roadway number in real time through an onboard precise positioning system (such as UWB or a fusion positioning system), and then queries the regional threshold table to automatically determine the risk level of the current area and the corresponding triggering conditions and basic speed limits. Specifically: ① The triggering conditions for high-risk areas include: the second dust concentration change rate data is greater than or equal to the first set benchmark threshold, and the second visibility change rate data is less than or equal to the second set benchmark threshold; the basic speed limit for the high-risk area is the first set speed value; High-risk areas mainly include narrow, poorly ventilated main transport roadways or areas near the tunneling face. In this embodiment, the preset first threshold value is preferably 2.0. The second set threshold is preferably 0.3 The preferred first speed setting is 1.2 km / h.

[0026] ② The triggering conditions for medium-risk areas include: the second dust concentration change rate data is greater than or equal to the third set benchmark threshold, and the second visibility change rate data is less than or equal to the fourth set benchmark threshold; the basic speed limit for the medium-risk area is the second set speed value; Medium-risk areas mainly include side roads connecting to main alleyways or areas with good ventilation. In this embodiment, the preset third threshold is preferably 2.0. The fourth set threshold is preferably... 0.2 The second set speed is preferably 25 km / h.

[0027] ③ The triggering conditions for low-risk areas include: the second dust concentration change rate data is greater than or equal to the fifth set benchmark threshold, and the second visibility change rate data is less than or equal to the sixth set benchmark threshold; the basic speed limit for the low-risk area is the third set speed value.

[0028] Low-risk areas mainly include spacious, well-ventilated main air intake passages or temporary vehicle parking areas. In this embodiment, the preset fifth threshold is preferably 0.8. The sixth set threshold is preferably... 0.15 The third set speed value is preferably 30 km / h.

[0029] Based on a preset regional threshold table, the system determines the risk level of the current area and the corresponding triggering conditions. The system will obtain the location of the vehicle in the tunnel in real time through the vehicle-mounted precise positioning module (such as UWB or fusion positioning system), and automatically determine the risk level of the current area and the corresponding triggering conditions based on the above regional threshold table. It should be noted that the trigger thresholds corresponding to each risk level in the regional threshold table are determined by engineering calibration methods based on a large amount of real vehicle test data in the mining area and system response time analysis. The calibration principle is to ensure safety while taking into account operational efficiency. Specifically, the system uses historical test data to determine the environmental parameter boundaries when the performance of the primary positioning system reaches a critical point of degradation. Based on this, different safety margins are assigned to areas with different risk levels. High-risk areas have shorter fault tolerance times, so more sensitive thresholds are set to allow for longer pre-switching times. Low-risk areas have larger fault tolerance spaces, so the thresholds can be relatively lenient to avoid unnecessary frequent switching. All thresholds are engineering optimal solutions that have been verified through extensive testing and can reliably trigger the pre-switching process before the performance of the primary positioning system deteriorates significantly.

[0030] S202: Based on the predicted second dust concentration change rate data and second visibility change rate data in step S1, compare them with the triggering conditions corresponding to the risk level of the current area. If either predicted change rate data does not meet the triggering conditions corresponding to the risk level of the current area in step S201, then generate a trigger signal at the first time point T0 and start the pre-switching window. In this embodiment, the holding duration of the pre-switching window is preferably 3 seconds, and the trigger signal includes a flag signal indicating that the pre-switching state has been activated and a risk level signal for this trigger.

[0031] S3. Based on the pre-switching window control, the main positioning mode and the backup positioning mode are operated in parallel. The first dust concentration data and the first visibility data at the current moment are obtained. Based on the first dust concentration data and the first visibility data, the first weight parameter of the main positioning mode and the second weight parameter of the backup positioning mode are calculated. Based on the first weight parameter and the second weight parameter, the first output data of the main positioning mode and the second output data of the backup positioning mode are weighted, fused, smoothed and weighted constrained. The first dust concentration data and the first visibility data are directly acquired by the vehicle-mounted dust concentration sensor and vision sensor (such as an industrial camera), and are absolute values ​​when in use.

[0032] S301: After the pre-switching window is opened, the system will start and run the primary positioning mode and the backup positioning mode simultaneously. The primary positioning mode and the backup positioning mode will output their positioning results at the same time, namely the first output data and the second output data.

[0033] The main positioning mode is a high-precision positioning algorithm based on lidar and high-precision maps on mining vehicles. This mode can provide centimeter-level high-precision positioning results in environments with good visibility and low dust concentration, making it the preferred solution for normal vehicle operation. However, its core sensor (lidar) is extremely sensitive to dust. When the dust concentration increases, the laser beam will be severely scattered and attenuated, resulting in sparse point cloud data or even failure, and a sharp drop in positioning accuracy. The backup positioning mode serves as a safety redundancy in case the primary positioning fails. It is mainly based on the dust-resistant positioning algorithm of millimeter-wave radar / visual odometry and high-precision map on mining vehicles. Millimeter-wave radar has a longer wavelength and has a strong ability to penetrate dust and haze. Visual odometry calculates displacement by analyzing changes in continuous image features. Its performance slowly decays as visibility decreases. Therefore, this combination scheme shows stronger robustness in high dust and low visibility environments, but its absolute accuracy in ideal environments is usually lower than that of the lidar scheme.

[0034] It should be noted that this method is specifically designed for the specific scenario in a mining environment where the performance of the main positioning source (LiDAR) fails intermittently and suddenly due to sudden changes in dust concentration. When the mining vehicle is already equipped with the aforementioned dual positioning system, the method utilizes the anti-interference capability of the backup positioning mode on the mining vehicle to achieve complementary advantages of the two heterogeneous positioning schemes. This ensures safety under extreme conditions while maximizing the optimal performance of the system under normal operating conditions.

[0035] S302: ① The first weighting parameter is calculated using the following formula: ; in, This is the first dust concentration data; As the basic weight for dust concentration, As the basic weight for visibility, a fixed coefficient calibrated through experiments is typically used to balance the effects of the two factors, which are a + b = 1.0. This is the dust concentration sensitivity coefficient, used to adjust the rate at which the weight decays with respect to dust concentration. The larger the value, the faster the primary positioning weight decreases as dust increases. This is the visibility sensitivity coefficient, used to adjust the weights' response sensitivity to changes in visibility. and Similarly, the results were obtained through real vehicle testing and experimental calibration. This is the primary visibility data; The ideal visibility for mine tunnel design is determined based on the specific mine tunnel's safe driving standards and pre-calibrated engineering parameters for sensor performance. The method is as follows: Under baseline conditions of good ventilation and no active dust generation in the mine tunnels, through extensive field measurements, a typical visibility value that meets the requirements for safe vehicle operation and stable sensor (such as industrial camera) perception is statistically derived and set as the ideal visibility for mine tunnel design. .

[0036] The lower limit of the weight preset for the primary positioning mode. The weight limit preset by the main positioning mode, and meets the following requirements. <1, which determines the safety boundary of the weighted output. and The specific values ​​were determined through extensive historical operational data and simulation testing to ensure the redundancy, security, and robustness of the positioning fusion system. The second weighting parameter is calculated using the following formula: ; in, This is the first weight parameter.

[0037] In addition, to prevent frequent oscillations of weight values ​​near the critical point of environmental parameters, this method also sets a hysteresis constant H (preferably 0.02 in this embodiment). In actual calculation, when the weight needs to change from a low value to a high value, the threshold for triggering the change will be more stringent than the threshold for changing from a high value to a low value, forming a hysteresis interval, thereby effectively suppressing frequent jitter.

[0038] ② The smoothing process is implemented using the exponential smoothing algorithm EMA. The smoothing coefficient α of the exponential smoothing algorithm ranges from 0.1 to 0.3. It can be dynamically adjusted between 0.1 and 0.3 according to the real-time load or risk level of the system. The smaller the value of α, the stronger the smoothing effect, but the response delay will also increase accordingly. ③ Weight constraint processing includes constraining the gradient changes of the first and second weight parameters within the pre-switching window. The absolute value of the weight change between adjacent calculation cycles does not exceed 0.15, fundamentally preventing weight jumps caused by sudden environmental changes. Specifically: In scenarios involving sudden changes in mine dust, the predicted second dust concentration change rate and second visibility change rate data may experience short-term, drastic fluctuations. Directly transmitting these fluctuations to the weight calculation would cause a jump in the fusion weights, resulting in a step change in the final positioning output, ultimately manifesting as vehicle shaking. Therefore, this invention introduces a weight gradient constraint mechanism, which forcibly limits the absolute value of the change in the first weight parameter within adjacent calculation cycles to no more than 0.15 within the pre-switching window. This value has achieved an optimal engineering balance between response speed and smoothness through extensive simulations and real-vehicle testing. Thus, even if there are instantaneous disturbances in the upstream environmental prediction, this mechanism can ensure that the change in the fusion weights follows a smooth trajectory, thereby effectively suppressing vehicle shaking and jerking, and ensuring the smoothness and safety of operations.

[0039] S4. Monitor the output difference between the first output data and the second output data, as well as the residual external parameters of the backup positioning mode. When the output difference and the residual external parameters meet the preset switching conditions, switch from the main positioning mode to the backup positioning mode at the second time point. If the preset switching conditions are not met, maintain the operation of the main positioning mode and control the vehicle to travel at the preset final speed limit.

[0040] S401: Specifically, within the pre-switching window, the system continuously monitors the output data of the primary positioning mode and the backup positioning mode, using them as a joint criterion for the switching decision; The preset switching conditions are: within a set number of sampling frames, the output difference between the first output data and the second output data is less than the first set difference value, and the residual of the external parameters of the backup positioning mode is less than the second set difference value; Specifically, in this embodiment, the output frequency of the primary positioning mode and the backup positioning mode is 20Hz, and the number of continuous sampling frames is preferably 5 frames. ① Output difference monitoring: The Euclidean distance between the first output data of the main positioning mode and the second output data of the backup positioning mode is calculated in real time and recorded as the output difference; In this embodiment, the first set difference is 0.2m. If the output difference between the first output data and the second output data is less than 0.2m, it means that the consistency between the primary positioning mode and the backup positioning mode is high at the current moment. ②External parameter residual monitoring: Real-time calculation of the external parameter residuals between the sensors and inertial measurement units used in the backup positioning mode (millimeter-wave radar / visual odometry and high-precision map); The extrinsic parameter residual refers to the error in the spatial transformation parameters (i.e., extrinsic parameters) between different sensors (such as millimeter-wave radar and inertial measurement unit, IMU) during actual operation in the standby positioning mode. This residual can be calculated by comparing the pose differences derived from the data of each sensor, or its uncertainty can be estimated in real time through online calibration algorithms (such as graph-optimized hand-eye calibration algorithms). In this embodiment, the second set difference value is 0.05 rad / m. If the extrinsic parameter residual of the standby positioning mode is less than 0.05 rad / m, it indicates that the spatial calibration relationship between the sensors in the standby positioning mode remains stable, and the consistency of multi-source data fusion is high. Therefore, it can be determined that the standby positioning mode is in a healthy state, its output positioning data is reliable, and the conditions for switching to this mode are met.

[0041] If the above two conditions are met for 5 consecutive frames at a sampling frequency of 20Hz, the system will issue a command at the second time point T1 to switch the positioning output source from the main positioning mode to the backup positioning mode.

[0042] S402: If the above joint criteria are not met by the end of the pre-switch window, it indicates that the current environment is too harsh or the backup system is unreliable, and a forced switchover poses a risk. Therefore, a safety fallback strategy is implemented, which includes: Maintain operational status: Continue operating in primary positioning mode; Activate speed reduction control: Controls the vehicle to travel at the preset final speed limit; This speed limit is a safety value, significantly lower than the normal operating speed, dynamically calculated based on the current warning level of the area. It reduces vehicle speed to provide a greater safety margin, offering basic safety assurance in situations where vehicle positioning performance may be unstable, until the environment improves or the system returns to normal. Specifically: The final speed limit is determined through the following steps: A. Determine the visibility scaling factor based on the first visibility data; B. Determine the level of suppression coefficient based on the established warning level; C. Determine the corresponding basic speed limit based on the current risk level of the area; D. Calculate the preliminary upper speed limit based on the base speed limit, visibility scaling factor, and level suppression factor; E. Compare the initial speed limit with the preset minimum safe speed, and take the larger value as the final speed limit.

[0043] It is expressed by the following formula: ; in, The preset global minimum safe speed, The base speed limit is obtained by querying the regional threshold table based on the current region's risk level. Based on first visibility data The calculated visibility scaling factor, This is the level suppression coefficient determined based on the identified warning level; ; in, The visibility offset defines a base value for the scaling factor when visibility is close to 0, ensuring that the function output is not lower than... ; The visibility sensitivity gain is the magnitude by which the control coefficient increases as visibility improves; This is a visibility reference value, representing the typical visibility of the mine tunnel under baseline conditions of good ventilation and no active dust generation. It is used to compare the first visibility data. Perform normalization processing; This is the primary visibility data; This is a non-linear adjustment factor used to finely adjust the response curve of the coefficient during visibility changes; This serves as the lower limit for the scaling factor, ensuring that a minimum speed control capability is maintained even under the lowest visibility conditions. In a specific calibration example, based on the operational data of a large mine, this example... The preferred value is 0.25. The preferred value is 0.75. Preferably 20 The preferred value is 0.6. The preferred value is 0.25; all the above parameters were calibrated through regression analysis of a large amount of historical mining data and safety operation experience testing of drivers in simulators, and their values ​​make the function It can accurately reflect the quantitative relationship between visibility and safe vehicle speed in a specific mining environment.

[0044] Grade Suppression Coefficient The system determines the warning level based on real-time assessments, and pre-sets corresponding warning levels within the system. The table shows that the higher the warning level, the smaller the inhibition factor, and the more strictly the speed limit is restricted. Specifically: If it is determined to be a low-risk warning level It is 0.85; If it is determined to be at the medium-risk warning level It is 0.65; If it is determined to be at a high-risk warning level It is 0.45; In addition, this embodiment also includes optional modification items. To address corrections for curves and slopes, specifically, it includes curvature correction terms and slope correction terms. The coefficient is derived from the theoretical safety upper limit based on the vehicle dynamics model (such as lateral acceleration constraints, uphill and downhill braking distances), and then calibrated through a large amount of simulation and real vehicle test data. It is fitted into a function or lookup table that is easy to implement in engineering, and its value range is (0,1]. The final speed limit can be described as: .

[0045] S403: Based on the warning level mentioned in step S402, this method also includes a warning level determination step, which specifically includes: A. Calculate the comprehensive risk value based on the first dust concentration data, the first visibility data, the second dust concentration change rate data, and the second visibility change rate data; )+ )+ )+ ); in, This is the second data point on the rate of change in dust concentration; The set benchmark threshold is the dust concentration change rate corresponding to the risk level of the current area determined by the area threshold table in step S201. This parameter serves as a buffer zone for dust changes. It is used to normalize the dust change rate exceeding a threshold, thus smoothly quantifying the severity of the change. In this embodiment, its preferred value range is 0.5 or 1.0. In the calculation of the comprehensive risk value, when the rate of change of dust exceeds the threshold... hour, If the value is positive, it is 0 otherwise. The numerical value of this calculation result directly reflects the relative degree to which the dust concentration rise rate exceeds the safety limit.

[0046] This parameter serves as a buffer zone for visibility changes and is used to normalize the rate of visibility decline exceeding a threshold. In this embodiment, its preferred value range is 0.05 to 0.1. In the calculation of the comprehensive risk value, when the rate of decline in visibility exceeds a threshold... hour, If the value is positive, it is 0 otherwise. The numerical value of this calculation result directly reflects the relative degree to which the rate of visual field deterioration exceeds the safety limit.

[0047] This is the second visibility change rate data. The baseline threshold is set for the visibility change rate corresponding to the current area's risk level based on the defects in the area threshold table in step S201. This is the concentration weighting coefficient, set based on experimental calibration or experience (e.g., 0.4~0.6), its purpose being to balance the effect of this factor on concentration. The impact; This reflects the risk level of the first dust concentration data; if the current concentration is close to or exceeds the safety benchmark... Increase; This is a visibility weighting coefficient, set based on experiments or experience. + =1, Used to control the proportion of static visual impact, this item reflects how much the current visibility has decreased compared to the ideal value; This is the first dust concentration data. This is the primary visibility data; To ensure a safe dust concentration according to mining area safety standards, this embodiment preferably uses 100 mg / m³. The maximum dust concentration that the equipment can withstand or that is specified by the standard is preferably 300 mg / m³ in this embodiment; The ideal visibility for mine tunnel design is derived from mine tunnel design standards, and in this embodiment, it is preferably 20 m. B. Weight the comprehensive risk value to obtain the final risk assessment value; ; in, For comprehensive risks; The confidence weight for the current prediction is dynamically adjusted based on the system's assessment of the reliability of the current prediction: its value is determined by conducting a quality assessment of the second dust concentration change rate data and the second visibility change rate data for the predicted future time period. Specifically... When the predicted data sequences of the second dust concentration change rate and the second visibility change rate for a given future time period are continuous, without abnormal jumps, and conform to the expected physical laws, the higher value should be selected. A value of 0.7 biases the decision towards the latest risk assessment results. However, when the second dust concentration change rate data and the second visibility change rate data for the predicted future time period experience instantaneous spike noise or signal interruption, A value of 0.3 allows decisions to rely more on smoothed historical trends, thus effectively filtering out unreliable instantaneous fluctuations. This is a smoothed value of the composite risk value from recent historical data, which is derived from... The historical sequence is calculated using exponential smoothing or moving average, representing a short-term baseline or trend of risk.

[0048] C. Determine the warning level for the current area based on the final risk assessment value. The warning levels include high-risk warning level, medium-risk warning level and low-risk warning level. Specifically, it includes upgrade conditions and downgrade adjustments: Upgrade requirements include: ①If the final risk assessment value If the value is greater than or equal to the first judgment threshold and less than the second judgment threshold, it is determined to be a low-risk warning level and a low-risk strategy is implemented. In this embodiment, the first judgment threshold is preferably 0.6, and the second judgment threshold is preferably 1.2, that is, when At that time, it was determined to be a low-risk warning level; ②If the final risk assessment value If the value is greater than or equal to the second judgment threshold and less than the third judgment threshold, it is determined to be a medium-risk warning level, and a medium-risk strategy is implemented. In this embodiment, the third judgment threshold is preferably 2, that is, when At that time, it was determined to be at a medium-risk warning level; ③If the final risk assessment value If the value is greater than or equal to the third judgment threshold, it is judged as a high-risk warning level, and a high-risk strategy is implemented. That is when At that time, it was determined to be at a high-risk warning level; Downgrade conditions include: ④ If it has already reached a high-risk warning level, and the final risk assessment value is... If the value is greater than or equal to the second judgment threshold and less than or equal to the difference between the third judgment threshold and the set delay amount, the level will be downgraded to a medium-risk warning level and a medium-risk strategy will be implemented. In this embodiment, the delay is set to 0.2, that is, when If so, it will be downgraded to a medium-risk warning level; ⑤ If it has already reached the medium-risk warning level, and the final risk assessment value is... If the value is greater than or equal to the first judgment threshold and less than or equal to the difference between the second judgment threshold and the set delay amount, the risk level will be downgraded to a low-risk warning level and a low-risk strategy will be implemented. That is when If so, it will be downgraded to a low-risk warning level; This method introduces a set delay amount into the degradation conditions, which enables a rapid and decisive response to high-risk warning levels. It also avoids control jitter at the risk boundary through the hysteresis mechanism, which significantly improves the stability and safety of autonomous mining vehicles operating in complex and harsh environments. The set delay amount is determined based on the analysis of the frequency of risk level switching in historical operating data, and is usually between 0.1 and 0.3. In this embodiment, it is preferably 0.2. It should be noted that the first judgment threshold, the second judgment threshold, the third judgment threshold, and the set delay amount are all calibrable configuration parameters. Their specific values ​​can be pre-calibrated and adjusted according to the roadway structure, ventilation conditions, and operational requirements of different mining areas. The values ​​(0.6, 1.2, 2, 0.2) given in this embodiment are only a preferred example.

[0049] D. Implement corresponding risk strategies based on the warning level; Low-risk strategies include: The vehicle speed is limited to the final speed limit determined based on the risk level and low-risk warning level of the current area, as shown in step S402. Maintain the parallel operation of the primary positioning mode and the backup positioning mode; The vehicle is subjected to Level 1 safety anchoring, which includes limiting the output power or torque of the drive system (such as an electric motor or engine) to 30% of the normal level, effectively preventing the vehicle from accelerating unexpectedly and counteracting the rollback that may be caused by slight inclines. The vehicle's parking brake system is triggered to enter a preload state. In this state, a slight, non-locking contact pressure is established between the brake pads and the brake disc, which shortens the response time of the braking system from the usual 50-100 milliseconds to less than 20 milliseconds, preparing for the instantaneous establishment of maximum braking force in an emergency, while not generating significant driving resistance. Medium-risk strategies include: Vehicle speed will be limited to the final speed limit determined based on the current risk level and medium-risk warning level of the area. The system forces the parallel operation of the primary positioning mode and the backup positioning mode and prohibits mode switching. Regardless of whether the switching conditions are met, the system maintains the fusion output of dual-mode positioning data. This can avoid erroneous switching or positioning data jumps that may occur due to the critical performance of the sensor in the current environment, and puts the stability of the positioning system in the highest priority. And to perform a second level of safety anchoring on the vehicle, which includes immediately executing a full parking brake command to make the brake pads fully engage with the brake discs and lock the wheels. For heavy mining vehicles, auxiliary stability systems such as differential locks or hydraulic outriggers can be activated at the same time to ensure that the vehicle does not move in roadways with a slope of ≤ 5°. The anchoring command can be executed automatically by the vehicle controller, or in a manned mode, it can be executed automatically if the driver is not canceled within a specified time (e.g., 100ms) after being alerted by an audible and visual alarm.

[0050] High-risk strategies include: immediately implementing Level 3 safety anchoring and / or controlling the vehicle to the final speed limit determined based on the current risk level and high-risk warning level of the area, and evacuating along a preset emergency path. Level 3 safety anchoring includes cutting off all non-essential power except for safety systems (such as positioning, communication, and environmental monitoring), locking all wheel brakes, and activating both mechanical and electronic braking systems. For specific models of mining vehicles, physical outriggers or track locking mechanisms can be further deployed to form rigid support, ensuring that the vehicle remains absolutely stationary even under conditions with a slope of ≤10°. Emergency evacuation, as an alternative or follow-up plan, controls vehicles to evacuate along a pre-set emergency route at an extremely low final speed limit (e.g., ≤3 km / h) determined based on high-risk areas. This strategy is only activated when "anchoring in place and waiting for the risk to pass" may cause greater secondary hazards (such as detecting excessive gas concentration or early warning of roadway structural instability). Throughout the evacuation process, the system continuously monitors the environment at the highest frequency (≥ 50 Hz). Once the positioning fails or the visibility suddenly drops to zero, the evacuation will be immediately interrupted and rigid anchoring will be re-executed. Within the set stabilization time (preferably 10 seconds in this embodiment), the final risk assessment value is... The process ends when the data remains below the first judgment threshold, and both the second dust concentration change rate data and the second visibility change rate data within the dynamically predicted next 10 seconds are within 50% of the trigger conditions corresponding to the risk level of the current area in the area threshold table.

[0051] It should be noted that the specific thresholds, parameters, durations, and control strategies (such as speed limit ratios and anchoring level definitions) in the embodiments of the present invention are all exemplary. In practical applications, they can be flexibly adjusted and calibrated according to different vehicle models, operating scenarios, and safety standards.

[0052] S404: To achieve long-term adaptive capability of the system and reduce manual maintenance costs, the present invention also includes a self-learning optimization step. This step continuously collects performance data during vehicle operation, establishes a closed-loop optimization circuit, and performs autonomous iterative optimization of key algorithm parameters. The self-learning optimization step includes autonomous iterative optimization of parameters in the first weight parameter calculation formula, and / or parameters and / or smoothing coefficients involved in calculating the comprehensive risk value, based on the performance data accumulated during operation used to characterize positioning error and control smoothness.

[0053] Specifically: Performance data acquisition: The system continuously accumulates data to characterize performance during operation, mainly including: Positioning error: including location error (Deviation from high-precision reference trajectory) and heading error ; Controlling smoothness: through accelerometer To quantify the smoothness of vehicle driving; System energy consumption: The total energy consumption generated by performing positioning, calculation, and control actions. .

[0054] Optimization includes online and offline updates. Online updates employ gradient descent or reinforcement learning algorithms, guided by the reward function, to optimize the sensitivity system in the first weight parameter calculation formula. and The smoothing coefficient α in the exponential smoothing algorithm is fine-tuned; offline updates employ batch gradient descent or supervised learning algorithms, adjusting the weight coefficients involved in calculating the comprehensive risk value based on historical datasets. and The basic weight of dust concentration in the formula for calculating the first weight parameter and visibility base weight The weight coefficients in the reward function are refitted. Specifically, online updates include: Online learning is conducted at a relatively high frequency (preferably 10 seconds in this embodiment) on 20% grayscale example vehicles to fine-tune parameters that need to adapt quickly to environmental changes.

[0055] Optimization parameters: Sensitivity system in the first weighting parameter calculation formula in step S302 and And the smoothing coefficient α in the exponential smoothing algorithm.

[0056] Reward Function and Optimization Mechanism: The optimization direction is defined by a reward function. The system adjusts parameters by seeking to maximize the reward. The reward function comprehensively considers key performance indicators. ; in, The positioning error weighting coefficient is used because when dust or visibility deteriorates in the mining area, positioning accuracy directly affects driving safety (such as obstacle avoidance and path tracking). ∈[0.3,0.5]; This is the heading error weighting coefficient: Heading deviation may cause the vehicle to deviate from the planned trajectory, which is especially risky in narrow mine tunnels. ∈[0.2,0.4]; For the jerk weighting coefficient: j This reflects the smoothness of the ride; excessive pressure can lead to vehicle bumps, loose cargo, or equipment damage. ∈[0.1,0.2]; Energy consumption weighting coefficient: Optimize energy consumption under the premise of safety and stability. ∈[0.05,0.1]; The initial values ​​of the above four weighting coefficients need to be verified through simulation tests: simulations should be conducted in typical mining scenarios (such as low-risk areas, medium-risk areas, and high-risk areas, with different dust concentration data and visibility data) to ensure that the weight combination can guide the system to prioritize reducing safety-related errors. In one update scheme of this embodiment, if the optimized performance (evaluated by the reward function R) deteriorates by more than 5% compared with the stable baseline version within a 5-minute rolling window, the system automatically abandons the optimization result and rolls back to the previous stable parameter set to ensure the safety and reliability of the learning process. Offline updates include: Offline learning is conducted on a daily or weekly basis, utilizing large-scale accumulated historical data to make batch and in-depth adjustments to more fundamental parameters that require long-term data support for accurate optimization.

[0057] Optimize parameters: Weighting coefficients involved in calculating the overall risk value and ; The basic weight of dust concentration in the first weighting parameter calculation formula and visibility base weight ; Weighting coefficients in the reward function , , , ; The execution process includes: during the daily off-peak period (in this embodiment, the criteria for determining the off-peak period are that the average speed of vehicles is ≤3 km / h for 10 consecutive minutes, the frequency of operations is ≤20%, and the vehicle load is ≤50%, then the current period is confirmed as an effective off-peak period), the system aggregates the fleet's performance data and corresponding environmental data for the whole day to the cloud or edge server; using batch reinforcement learning or supervised learning algorithms, the above parameters are refitted based on massive historical data, with the goal of optimizing the long-term average performance (or cumulative reward); the optimized new parameters are deployed during the off-peak period of the next day, thereby completing the iterative update of system parameters and performance evolution.

[0058] Through the aforementioned dual-layer learning mechanism combining online and offline learning, this invention transforms the autonomous driving system from a static program with fixed parameters into an intelligent agent capable of learning from experience. It can automatically adapt to the characteristics of different mining areas and track environmental changes caused by mine tunnel wear and equipment aging. Ultimately, it achieves excellent performance with high precision, high smoothness, and low energy consumption throughout its entire life cycle, while significantly reducing manual maintenance and calibration costs.

[0059] Example 2: This embodiment provides a system for implementing the dynamic positioning switching control method for autonomous driving mining vehicles in Embodiment 1. The system includes: The data acquisition module is configured to acquire dust concentration change rate data and visibility change rate data for a set historical time period, as well as the first dust concentration change rate data and the first visibility change rate data for the current area, and the first dust concentration data and the first visibility data at the current moment. This module is directly connected to the vehicle-mounted dust concentration sensor and vision sensor and is responsible for the synchronous acquisition and preliminary processing of data.

[0060] The prediction module, which is communicatively connected to the data acquisition module, is configured to receive historical and current data provided by the data acquisition module and run a short-time series prediction model (such as Kalman filtering or lightweight long short-term memory network) as in step S1 of Embodiment 1 to predict the second dust concentration change rate data and the second visibility change rate data for a future set time period.

[0061] The processing module is communicatively connected to both the data acquisition module and the prediction module. This module is configured to execute the core control logic detailed in steps S2 to S4, S403, and S404 of Embodiment 1, and its specific functions include: Access and query the preset area threshold table to determine the risk level and corresponding triggering conditions of the current area based on the vehicle's current location; The second rate of change data output by the prediction module is compared with the triggering condition, and a trigger signal is generated at the first time point to start the pre-switching window. Within the pre-switching window, the primary positioning mode and the backup positioning mode are controlled to run in parallel; Based on the first dust concentration data and the first visibility data, the first weight parameter and the second weight parameter are dynamically calculated. The output data of the primary and backup positioning modes are weighted, fused, smoothed, and subject to weight constraints. Monitor the output difference and external parameter residuals, and when the preset switching conditions are met, perform the switch from the primary positioning mode to the backup positioning mode at the second time point; If the switching conditions are not met, the main positioning mode will be maintained and the vehicle will be controlled to travel at the preset final speed limit. The warning level determination process is executed, and the current warning level is determined based on the final risk assessment value, and the corresponding risk strategy is triggered. The system performs a self-learning optimization step, autonomously iteratively optimizing key parameters based on accumulated performance data. The primary positioning mode and the backup positioning mode serve as positioning execution units of the system and are scheduled by the processing module. The primary positioning mode mainly consists of a lidar and a high-precision map, while the backup positioning mode mainly consists of a millimeter-wave radar and / or a visual odometry and a high-precision map. Together, they provide the processing module with the first output data and the second output data. Each module is connected to the external parameter calibration via time synchronization to form a closed-loop intelligent control system, which together realizes the full process function from steps S1 to S4 as shown in Example 1.

[0062] In the description of this application, it should be noted that the terms "upper," "lower," etc., indicating the orientation or positional relationship are based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this application. Unless otherwise expressly specified and limited, the terms "installed," "connected," and "linked" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication between two elements. For those skilled in the art, the specific meaning of the above terms in this application can be understood according to the specific circumstances.

[0063] It should be noted that in this application, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0064] The above description is merely a specific embodiment of this application, enabling those skilled in the art to understand or implement this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A control method for dynamic positioning switching of an automated driving mining vehicle, characterized in that, Includes the following steps: Based on historical dust concentration change rate data and visibility change rate data for a set time period, as well as the first dust concentration change rate data and the first visibility change rate data for the current area, predict the second dust concentration change rate data and the second visibility change rate data for a set time period in the future. The risk level of the current area and the corresponding triggering conditions are determined based on the preset regional threshold table. The triggering conditions are compared with the second dust concentration change rate data and the second visibility change rate data. If any change rate data reaches or exceeds the triggering conditions corresponding to the risk level of the current area, a trigger signal is generated at the first time point and the pre-switching window is started. Based on the pre-switching window control, the main positioning mode and the backup positioning mode are operated in parallel. The first dust concentration data and the first visibility data at the current moment are obtained. Based on the first dust concentration data and the first visibility data, the first weight parameter of the main positioning mode and the second weight parameter of the backup positioning mode are calculated. Based on the first weight parameter and the second weight parameter, the first output data of the main positioning mode and the second output data of the backup positioning mode are weighted, fused, smoothed and weighted constrained. Monitor the output difference between the first output data and the second output data, as well as the residual external parameters of the backup positioning mode. When the output difference and the residual external parameters meet the preset switching conditions, perform the switching from the primary positioning mode to the backup positioning mode at the second time point. If the preset switching conditions are not met, the main positioning mode will continue to operate, and the vehicle will be controlled to travel at the preset final speed limit.

2. The control method for dynamic positioning switching of autonomous mining vehicles according to claim 1, characterized in that, The preset regional threshold table includes the risk level, triggering conditions, and basic speed limit of the current region. The risk level of the current region includes low-risk, medium-risk, and high-risk regions, wherein: The triggering conditions for the high-risk area include: the second dust concentration change rate data is greater than or equal to the first set benchmark threshold, and the second visibility change rate data is less than or equal to the second set benchmark threshold; the basic speed limit for the high-risk area is the first set speed value. The triggering conditions for the medium-risk area include: the second dust concentration change rate data is greater than or equal to the third set benchmark threshold, and the second visibility change rate data is less than or equal to the fourth set benchmark threshold; the basic speed limit for the medium-risk area is the second set speed value. The triggering conditions for the low-risk area include: the second dust concentration change rate data is greater than or equal to the fifth set benchmark threshold, and the second visibility change rate data is less than or equal to the sixth set benchmark threshold; the basic speed limit for the low-risk area is the third set speed value.

3. The control method for dynamic positioning switching of autonomous mining vehicles according to claim 1, characterized in that, The first weight parameter is calculated using the following formula: ; in, This is the first dust concentration data. As the basic weight for dust concentration, This is the dust concentration sensitivity coefficient. As the basic weight for visibility, Visibility sensitivity coefficient This is the primary visibility data. Ideal visibility for mine tunnel design The lower limit of the weight preset for the primary positioning mode. The preset weight limit for the primary positioning mode; The second weighting parameter is calculated using the following formula: ； in, This is the first weight parameter.

4. The control method for dynamic positioning switching of autonomous mining vehicles according to claim 3, characterized in that, The smoothing process is specifically implemented using an exponential smoothing algorithm, where the smoothing coefficient α ranges from 0.1 to 0.

3. The weight constraint processing includes constraining the gradient of the change of the first weight parameter and the second weight parameter within the pre-switching window, such that the absolute value of the change in weight between adjacent calculation cycles does not exceed 0.

15.

5. The control method for dynamic positioning switching of autonomous mining vehicles according to claim 1, characterized in that, The preset switching condition is: within a set number of sampling frames, the output difference between the first output data and the second output data is less than a first set difference value, and the residual of the external parameters of the backup positioning mode is less than a second set difference value.

6. The control method for dynamic positioning switching of autonomous mining vehicles according to claim 1, characterized in that, The process of generating a trigger signal and starting the pre-switching window at the first time point also includes: The comprehensive risk value is calculated based on the first dust concentration data, the first visibility data, the second dust concentration change rate data, and the second visibility change rate data. The comprehensive risk value is weighted and adjusted to obtain the final risk assessment value; The warning level for the current area is determined based on the final risk assessment value. The warning levels include high-risk warning level, medium-risk warning level and low-risk warning level. Based on the aforementioned warning level, the corresponding risk strategy will be implemented.

7. The control method for dynamic positioning switching of autonomous mining vehicles according to claim 6, characterized in that, The process of determining the warning level of the current area based on the final risk assessment value, and executing the corresponding risk strategy based on the warning level, specifically includes: If the final risk assessment value is greater than or equal to the first assessment threshold and less than the second assessment threshold, it is determined to be a low-risk warning level, and a low-risk strategy is executed. The low-risk strategy includes: limiting the vehicle speed to the final speed limit determined based on the risk level and low-risk warning level of the current area, maintaining the parallel operation of the main positioning mode and the backup positioning mode, and performing a first-level safety anchoring on the vehicle. The first-level safety anchoring includes limiting the output power of the drive system according to a set ratio and triggering the vehicle parking brake system to enter a preload state. If the final risk assessment value is greater than or equal to the second assessment threshold and less than the third assessment threshold, it is determined to be a medium-risk warning level, and a medium-risk strategy is implemented. The medium-risk strategy includes: limiting the vehicle speed to the final speed limit determined based on the risk level of the current area and the medium-risk warning level, forcibly maintaining the parallel operation of the primary positioning mode and the backup positioning mode and prohibiting mode switching, and implementing a second-level safety anchor for the vehicle, which includes cutting off the driving power and implementing the parking brake. If the final risk assessment value is greater than or equal to the third assessment threshold, it is determined to be a high-risk warning level and a high-risk strategy is executed. The high-risk strategy includes: immediately implementing a third-level safety anchor for the vehicle and / or controlling the vehicle to the final speed limit determined based on the risk level and high-risk warning level of the current area, and evacuating along a preset emergency path. The third-level safety anchor includes cutting off all unsafe power and locking all wheel brakes. If the risk level is already high, and the final risk assessment value is greater than or equal to the second assessment threshold, and less than or equal to the difference between the third assessment threshold and the set delay amount, then the risk level will be downgraded to medium risk and a medium risk strategy will be implemented. If the risk level is already at the medium risk warning level, and the final risk assessment value is greater than or equal to the first assessment threshold, and less than or equal to the difference between the second assessment threshold and the set delay amount, then the risk level will be downgraded to the low risk warning level, and a low risk strategy will be implemented.

8. The control method for dynamic positioning switching of autonomous mining vehicles according to claim 6, characterized in that, The final speed limit is determined through the following steps: Determine the visibility scaling factor based on the first visibility data; The level of suppression coefficient is determined based on the current warning level of the region; The corresponding basic speed limit is determined based on the current risk level of the area. Based on the basic speed limit, visibility scaling factor, and level suppression factor, calculate the preliminary speed limit value; The initial speed limit is compared with the preset minimum safe speed, and the maximum value is taken as the final speed limit.

9. The control method for dynamic positioning switching of autonomous mining vehicles according to claim 7, characterized in that, It also includes a self-learning optimization step, which includes autonomously iteratively optimizing the parameters in the first weight parameter calculation formula and / or the parameters and / or smoothing coefficients involved in calculating the comprehensive risk value based on the performance data accumulated during operation to characterize positioning error, control smoothness, and system energy consumption.

10. A system for a control method of dynamic positioning switching based on autonomous driving mining vehicles, characterized in that, include: The data acquisition module is configured to acquire dust concentration change rate data and visibility change rate data for a set historical time period, first dust concentration change rate data and first visibility change rate data for the current area, and first dust concentration data and first visibility data for the current moment. The prediction module is configured to predict the second dust concentration change rate data and the second visibility change rate data for a future set time period; The processing module is configured to determine the risk level of the current area and the triggering conditions corresponding to the risk level based on a preset area threshold table, and compare the triggering conditions with the second dust concentration change rate data and the second visibility change rate data. If either change rate data reaches or exceeds the triggering conditions corresponding to the risk level of the current area, a trigger signal is generated at the first time point and a pre-switching window is started. The main positioning mode and the backup positioning mode are controlled to run in parallel based on the pre-switching window. At the same time, the first weight parameter of the main positioning mode and the second weight parameter of the backup positioning mode are calculated based on the first dust concentration data and the first visibility data. The first output data of the main positioning mode and the second output data of the backup positioning mode are weighted, fused, smoothed and weighted constrained based on the first weight parameter and the second weight parameter. Monitor the output difference between the first output data and the second output data, as well as the residual external parameters of the backup positioning mode. When the output difference and the residual external parameters meet the preset switching conditions, perform the switching from the primary positioning mode to the backup positioning mode at the second time point. If the preset switching conditions are not met, the main positioning mode will continue to operate, and the vehicle will be controlled to travel at the preset final speed limit.

Images