Rapid regulation and control method for earth-rock excavation and filling balance of new energy mine clamping in steep, narrow and rare areas

Abstract

The invention discloses a rapid regulation and control method for earth-rock excavation and filling balance of new energy mine clamping in steep, narrow and rare areas, and belongs to the technical field of automatic control of mining and earthwork engineering machinery. The method aims at solving the technical problems that in the prior art, due to the fact that manual judgment and regular measurement are relied on, digging and filling balance regulation and control efficiency is low, and real-time performance is poor. According to the technical scheme, the method is characterized in that real-time point cloud data are acquired through a three-dimensional laser scanner; the vehicle-mounted processing unit compares the real-time point cloud data with a pre-stored design terrain model within 0.3 s to 1.5 s to generate a real-time digging and filling distribution diagram; and calculating a digging and filling balance ratio on the basis, when the ratio exceeds a range of 0.95-1.05, planning an adjustment path pointing to a region with the maximum digging and filling difference, and controlling the vehicle to execute at a speed of 5-15 km / h. The method is mainly used for new energy mine truck construction operation in steep and narrow mining areas, and rapid and automatic regulation and control of earth-rock excavation and filling balance are achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automatic control of mining and earthwork machinery. More specifically, the present application relates to a method for rapid regulation and control of earthwork excavation and filling balance in steep and narrow areas. BACKGROUND

[0002] In open-pit mining, large earthwork engineering and other operations, the balance regulation of earthwork excavation and filling is one of the core operation links. Achieving dynamic balance of excavation and filling volume plays a key role in ensuring operation continuity, improving transportation efficiency and reducing operating costs. The traditional excavation and filling balance operation mode mainly relies on manual experience and periodic measurement. Specifically, the dispatch personnel command the mine truck vehicles to load, transport and unload the earthwork according to the stage survey results and experience judgment. This mode can still maintain basic operation in open and flat conventional operation areas, but it exposes many technical bottlenecks when facing steep and narrow special terrain.

[0003] Firstly, in steep and narrow areas, the terrain changes rapidly and the field of view is limited, making it difficult to achieve real-time dynamic balance of earthwork by relying on manual judgment and periodic measurement. The fundamental reason is that manual judgment has a large delay and cannot quickly perceive the subtle changes in the terrain, while the traditional surveying and mapping cycle is as long as several hours or even several days, resulting in a serious lag in the obtained excavation and filling state information. When the excavation and filling imbalance occurs, it cannot be immediately detected, and adjustments are made only after the problem accumulates to a certain extent, often requiring more time and resources to correct, which seriously affects the operation efficiency. Attempts have been made to improve this problem by increasing the frequency of manual inspection or increasing the density of measurement points, but face the difficulties of high labor costs, data update speed still cannot meet the requirements of real-time regulation and control, and personnel safety is difficult to guarantee in dangerous steep areas.

[0004] Secondly, to achieve automatic regulation and control, a high-precision design terrain model of the operation area needs to be established as a reference. In existing technologies, a single platform (such as only a drone or only ground equipment) is often used for terrain mapping. However, in steep and narrow areas, single-platform mapping has inherent limitations: drone aerial surveying is easily blocked by cliff walls, resulting in blind areas and incomplete models; and relying solely on ground scanning is inefficient and difficult to obtain a macro overall terrain. If multi-source point cloud data obtained from different platforms is fused, new technical problems will be introduced. Different sources of point cloud data differ in resolution, coordinate system and perspective, and direct fusion will result in registration errors. This error is particularly pronounced in complex terrain and is passed on and accumulated in the subsequent data processing chain, ultimately causing the design terrain model as a reference to have reduced accuracy, causing all automatic judgments and decisions based on this model to be built on an unreliable foundation. Solving the seamless fusion of multi-source data and precision control is a key difficulty in improving system reliability.

[0005] Moreover, even if an initial high-precision model is established, in a dynamic earthwork environment, the terrain is always changing rapidly. A static model will quickly "fail" as the work progresses, and cannot truly reflect the current excavation and filling state. Therefore, dynamic updating of the designed terrain model is a necessary requirement to maintain system effectiveness. However, dynamic updating of the model itself also brings stability challenges. Environmental noise at the work site, sensor measurement errors, and temporary disturbances caused by vehicle work, can all be misjudged by the model as valid terrain changes and updated. This false update will cause the model to distort, and in turn trigger oscillations in system decision-making - for example, vehicles are repeatedly assigned to "false" work areas generated by noise. How to design an updating mechanism that can both track the real terrain evolution and effectively filter out interference to maintain the long-term stability of the model is a serious challenge faced by automated control systems in practical applications. SUMMARY

[0006] An object of the present application is to provide a steep and narrow area new energy mine card earthwork excavation and filling balance rapid regulation method, which realizes second-level perception and response to the terrain state of the steep and narrow area, and significantly improves the timeliness and automation level of the excavation and filling balance operation.

[0007] In order to achieve these objects and other advantages of the present application, according to one aspect of the present application, the present application provides a preparation method of a steep and narrow area new energy mine card earthwork excavation and filling balance rapid regulation method, comprising the following steps: Step one, through the three-dimensional laser scanner installed on the new energy mine card, continuously scanning the terrain of the annular work area with the new energy mine card as the center and a radius of 50 m to 100 m at a scanning frequency of 20 Hz to 40 Hz, and acquiring real-time point cloud data; Step two, transmitting the real-time point cloud data to the vehicle-mounted processing unit of the new energy mine card through the vehicle-mounted communication system, and the vehicle-mounted processing unit completes the processing of the real-time point cloud data within 0.3 s to 1.5 s, compares the real-time point cloud data with the pre-stored designed terrain model, and generates a real-time excavation and filling distribution map; Step three, the vehicle-mounted processing unit calculates the current excavation volume and filling volume based on the real-time excavation and filling distribution map, and calculates the ratio of the excavation volume to the filling volume as the excavation and filling balance ratio; when the excavation and filling balance ratio is less than 0.95 or greater than 1.05, the vehicle-mounted processing unit determines the adjustment path and adjustment speed of the new energy mine card according to the real-time point cloud data, the adjustment path points to the area with the largest difference between the excavation volume and the filling volume, and the adjustment speed is controlled between 5 km / h and 15 km / h; Step 4: The on-board processing unit sends the adjustment path and speed to the control system of the new energy mining truck. The control system drives the electric drive mechanism and steering mechanism of the new energy mining truck according to the adjustment path and speed to execute the adjustment path and speed.

[0008] Preferably, the pre-stored design terrain model in step two is established through the following steps; The entire work area was scanned using UAV-borne lidar, while a ground-based 3D laser scanner was used to supplement the scan of steep cliffs and narrow ravines to obtain high-precision initial point cloud data. The point cloud data acquired by the UAV-borne lidar and the ground-based 3D laser scanner are fused and registered to generate an initial digital elevation model of the work area with a resolution of 2 cm to 5 cm. The digital design terrain model provided by the engineering design party is overlaid and compared with the initial digital elevation model. Areas with elevation deviations exceeding 10 cm to 15 cm are manually verified and corrected to generate a pre-stored design terrain model as a benchmark.

[0009] Preferably, the point cloud data fusion and registration includes a precision control step; During the fusion registration process, the registration error between the UAV-borne LiDAR and the ground-based 3D laser scanner point cloud data is calculated. When the registration error is greater than 3 cm to 5 cm, the iterative nearest point algorithm is used for re-registration. The re-registered point cloud data is subjected to an overlap area consistency check, and point cloud data with an elevation difference of more than 2 cm to 4 cm in the overlap area are removed. In the final fused point cloud data, the registration error of 90% to 95% of the points is no greater than 1.5 times the resolution of the initial digital elevation model.

[0010] Preferably, the vehicle-mounted processing unit is equipped with a model update module, which dynamically updates the pre-stored design terrain model based on real-time point cloud data; wherein the dynamic update follows the following steps; After completing one path adjustment, the on-board processing unit performs a second differential calculation on the real-time point cloud data collected and registered during this round of operation and the design terrain model before dynamic update to generate an incremental terrain dataset. Statistical analysis was performed on the terrain increment dataset to calculate its average height change and the continuous area of ​​the changed region; only when the absolute value of the average height change is greater than 5 cm to 10 cm and the continuous area of ​​the changed region is greater than 2 m², the change is considered complete. 2 up to 5 m 2 Only when this condition is met is the incremental data deemed valid. The incremental terrain dataset that has passed the validity verification is superimposed onto the pre-stored design terrain model in a weighted fusion manner to complete the iterative update of the model; The model updating module further comprises an update stability control mechanism; the mechanism records the average change height of the current terrain incremental data set at each iterative update, and compares it with the change height recorded in the previous two updates; when the change height sequence of the last three updates alternates in sign and the absolute values of the adjacent two changes are both greater than 8 cm to 12 cm, it is determined to be oscillation and the update amplitude limit is started; the update amplitude limit is realized by introducing a weighting factor starting from 0.1 and increasing by 0.1 step by step until 0.5, which is used to attenuate the fusion weight of the current terrain incremental data set.

[0011] Preferably, the generation of the real-time excavation and filling distribution map in step two comprises the following steps: The vehicle-mounted processing unit converts the real-time point cloud data into a grid digital elevation model with a resolution of 10 cm to 20 cm; The grid digital elevation model and the pre-stored design terrain model are subjected to elevation difference calculation on a grid cell basis under the same coordinate system to generate an elevation difference matrix; The elevation difference matrix is subjected to region growing algorithm processing based on neighborhood connectivity to cluster adjacent grid cells with an elevation change value greater than 15 cm to 30 cm into independent change regions; The average elevation change value of each independent change region is calculated, and regions with a negative average elevation change value are marked as excavation regions, and regions with a positive average elevation change value are marked as filling regions; The spatial position information of all excavation regions and filling regions is combined to generate a real-time excavation and filling distribution map.

[0012] Preferably, the calculation of the current excavation volume and filling volume in step three is realized by the following steps: The vehicle-mounted processing unit extracts the boundary contours of all excavation regions and filling regions based on the real-time excavation and filling distribution map; Triangular meshing is performed on each region to generate an irregular triangular mesh model; Based on the spatial relationship between the triangular mesh model and the pre-stored design terrain model, the prismatic volume integration method is used to calculate the excavation volume of each excavation region and the filling volume of each filling region; In the calculation process, a real-time calibration mechanism based on the vehicle-mounted weighing system is introduced, which realizes dynamic updating of the earthwork density value by the following method: during the loading process of the new energy mine truck, the vehicle-mounted processing unit synchronously acquires the actual load mass data measured by the vehicle-mounted weighing system, and dynamically updates the earthwork density value in combination with the calculated volume of the corresponding excavation region; The updated earthwork density value is applied to the subsequent volume calculation process to realize dynamic calibration of the excavation and filling volume calculation.

[0013] Preferably, the earthwork density real-time calibration mechanism further comprises an error compensation step; the vehicle-mounted processing unit establishes an earthwork density error propagation model, which takes into account the measurement error of 1% to 2% of the vehicle-mounted weighing system and the volume calculation error of 2% to 3% of the point cloud data; when dynamically updating the earthwork density value, the Kalman filtering algorithm is used to filter the density measurement value; the compensated and corrected volume calculation results of the excavated volume and the filled volume are obtained by combining the error propagation model with the filtered density value; after compensation and correction, the estimated range of the volume calculation error is between 3% and 5%.

[0014] Preferably, the determination of the adjusted path in step three is achieved by the following steps: The vehicle-mounted processing unit identifies the target area with the largest difference between the excavated volume and the filled volume based on the real-time excavation and filling distribution map, and takes the center coordinates of the target area as the end point of the path; A path planning algorithm is used to search for a path in the obstacle map generated by real-time point cloud data, and the cost function of the path planning algorithm includes two calculation items: path length value and terrain slope value; The initial path output by the path planning algorithm is subjected to safety verification based on vehicle dynamics constraints, which excludes path segments that exceed the maximum climbing angle of 30% to 35% of the new energy mine truck, and excludes path segments that are less than the minimum turning radius of 8 m to 10 m of the new energy mine truck; A curve fitting algorithm is used to process the path that passes the safety verification to generate an executable path with continuous curvature; During the execution of the executable path by the new energy mine truck, the three-dimensional laser scanner performs real-time obstacle detection within a range of 20 m to 30 m in the direction of travel at a frequency of 5 Hz to 10 Hz; When an obstacle is detected, the vehicle-mounted processing unit initiates local path re-planning, generates an obstacle-avoiding path, and continues to execute.

[0015] Preferably, the path planning algorithm uses a hierarchical planning strategy; the hierarchical planning strategy includes a global planning layer and a local planning layer; the global planning layer uses a path planning algorithm to perform coarse-grained path planning with a resolution of 1 m to 2 m, and the planning period is 10 s to 15 s; the local planning layer uses the same path planning algorithm to perform fine-grained path planning with a resolution of 0.2 m to 0.5 m, and the planning period is 1 s to 2 s; the global planning layer and the local planning layer are coupled through a sequence of path points, and the local planning layer performs real-time fine planning between global path points; the vehicle-mounted processing unit allocates dynamically adjustable computing resources to the global planning layer and the local planning layer; when the frequency of local planning layer triggering re-planning exceeds 3 times within 1 second, the system automatically tilts more computing resources to the local planning layer.

[0016] Preferably, the system stability monitoring and control mode switching step is further included; The vehicle-mounted processing unit monitors the following system key indicators in real time: fluctuation frequency of the excavation-filling balance ratio, data integrity rate of real-time point cloud data, and triggering frequency of local path re-planning; When the following conditions are met simultaneously, it is determined that the system is in a state of instability risk: within 30 s, the fluctuation frequency of the excavation-filling balance ratio exceeds 10 times per minute, the data integrity rate of real-time point cloud data is less than 85%, and the triggering frequency of local path re-planning exceeds 5 times per minute; When the system is determined to be in a state of instability risk, the vehicle-mounted processing unit automatically switches from the full-automatic control mode to the assisted control mode; In the assisted control mode, the vehicle-mounted processing unit simplifies the path to a straight line path from the current position to the nearest excavation-filling area, and limits the adjustment speed to 5-8 km / h; The system continuously monitors the key indicators, and when the indicators return to normal and remain stable for 10 s, it automatically switches back to the full-automatic control mode.

[0017] The present application at least includes the following beneficial effects: The steep and narrow area new energy mine card earthwork excavation-filling balance rapid regulation method disclosed by the present application realizes second-level perception and response to the terrain state of the steep and narrow area by constructing a closed-loop control loop of scanning-processing-decision-execution, significantly improving the timeliness and automation level of the excavation-filling balance operation. By fusing unmanned aerial vehicle and ground scanning data and implementing strict registration accuracy control, a high-precision and complete initial terrain reference is established, which guarantees the reliability of subsequent judgment from the source. The model dynamic updating mechanism based on the saliency criterion and stability control is introduced, which enables the reference model to continuously evolve with the operation process while effectively suppressing model distortion and decision oscillation caused by data noise. The grid processing and region growing algorithm are used to realize accurate identification and automatic classification of the excavation-filling area. Combined with the volume calculation method of real-time calibration and error compensation of the vehicle-mounted scale, the calculation accuracy of the earthwork volume is greatly improved. Through the integration of hierarchical path planning with vehicle dynamics constraints and real-time obstacle detection, the executability and safety of the planned path are ensured. Finally, the multi-indicator monitoring and degradation strategy at the system level provides the final guarantee for stable operation under complex conditions, forming a complete technical closed loop from bottom-layer data to top-layer decision, and realizing rapid, accurate and robust regulation of earthwork excavation-filling balance under harsh terrain conditions.

[0018] Other advantages, objects, and features of the application will be apparent from the following specification, and upon examination of the appropriate documents. DETAILED DESCRIPTION

[0019] The application will be further described in detail below with reference to specific embodiments, so that those skilled in the art can implement the application according to the description.

[0020] It should be understood that the terms such as "have", "contain" and "include" used herein do not exclude the presence or addition of one or more other elements or combinations thereof.

[0021] It should be noted that the experimental methods in the following embodiments are conventional methods, and the reagents and materials are commercially available unless otherwise specified.

[0022] The application provides a steep and narrow region new energy mine card earthwork excavation and filling balance rapid regulation method, comprising the following steps: Step one, through the three-dimensional laser scanner installed on the new energy mine card, the scanning frequency of 20 Hz to 40 Hz is used to continuously scan the terrain of the annular operation area with the new energy mine card as the center and the radius of 50 m to 100 m, and the real-time point cloud data is obtained; Step two, the real-time point cloud data is transmitted to the vehicle-mounted processing unit of the new energy mine card through the vehicle-mounted communication system, the vehicle-mounted processing unit completes the processing of the real-time point cloud data within 0.3 s to 1.5 s, compares the real-time point cloud data with the pre-stored design terrain model, and generates a real-time excavation and filling distribution map; Step three, the vehicle-mounted processing unit calculates the current excavation volume and filling volume based on the real-time excavation and filling distribution map, and calculates the ratio of the excavation volume to the filling volume as the excavation and filling balance ratio; when the excavation and filling balance ratio is lower than 0.95 or higher than 1.05, the vehicle-mounted processing unit determines the adjustment path and adjustment speed of the new energy mine card according to the real-time point cloud data, the adjustment path points to the area with the largest difference between the excavation volume and the filling volume, and the adjustment speed is controlled between 5 km / h and 15 km / h; Step four, the vehicle-mounted processing unit sends the adjustment path and adjustment speed to the control system of the new energy mine card, and the control system drives the electric drive mechanism and steering mechanism of the new energy mine card according to the adjustment path and adjustment speed to execute the adjustment path and adjustment speed.

[0023] In the above technical solution, the three-dimensional laser scanner can be a 32-line or 64-line mechanical rotating laser radar. The scanner is fixedly installed on the top protection frame of the mine card cab through a rigid support, and the scanning center axis is inclined downward by an angle of 5 to 10 degrees with the horizontal plane to optimize the scanning coverage of the nearby ground. The scanner works at a fixed frequency of 20 Hz, emits laser pulses and receives return signals, and generates point cloud data containing three-dimensional coordinates and reflection intensity. The point cloud data is transmitted to the vehicle-mounted processing unit through shielded twisted pair, and about 300,000 valid points are generated per second, which is sufficient to describe the terrain features within a radius of 80 meters.

[0024] The vehicle-mounted processing unit can be an industrial-grade vehicle-mounted computer with a multi-core CPU and a dedicated GPU computing module. The processing unit is installed in a shockproof cabinet in the cab of the mining truck and is connected with the scanner and the control system through vehicle-mounted Ethernet. The real-time point cloud data processing process includes: first, filtering outliers and voxel grid downsampling of the input point cloud to reduce the data volume to 60% of the original data; then, using the iterative closest point algorithm, the processed real-time point cloud is registered with the pre-stored design terrain model, and the registration error is controlled within 3 cm; finally, the real-time excavation and filling distribution map is generated by calculating the height difference, and the entire processing process takes about 0.8 seconds.

[0025] The excavation and filling balance calculation and path planning module is integrated in the vehicle-mounted processing unit and realized through special algorithm software. The module first extracts all the elevation change areas from the real-time excavation and filling distribution map and uses the region growing algorithm to cluster adjacent change areas; then, based on the triangular mesh model, the volume of each area is calculated, and the total excavation volume and filling volume are accumulated; when the excavation and filling balance ratio reaches 0.97, the path planning program is started. The path planning uses the A-star algorithm, considering the path length and terrain slope factors, to generate a safe path from the current position to the target area, with the driving speed set to 8 km / h. The planned path and speed parameters are sent to the vehicle control system through the CAN bus.

[0026] Through the above implementation, a complete automatic control chain from terrain perception to vehicle execution is established, and continuous monitoring and rapid response to the terrain state of the working area are realized. The use of industrial-grade hardware devices ensures the reliability of the system in harsh working conditions, and the optimized algorithm process guarantees the timeliness of data processing and decision-making. This method can effectively replace the traditional manual judgment mode and realize the automatic regulation and control of earthwork excavation and filling balance in steep and narrow special terrain conditions, improving the safety and efficiency of the operation.

[0027] In other technical solutions, the pre-stored design terrain model in step two is established by the following steps: The working area is scanned by an unmanned aerial laser radar, and the steep cliff and narrow ravine areas are supplemented by a ground three-dimensional laser scanner to obtain high-precision initial point cloud data. The point cloud data obtained by the unmanned aerial laser radar and the ground three-dimensional laser scanner are fused and registered to generate an initial digital elevation model of the working area with a resolution of 2 cm to 5 cm. The digital design terrain model provided by the engineering design party is superimposed and compared with the initial digital elevation model, and the areas with an elevation deviation of more than 10 cm to 15 cm are manually checked and corrected to generate the pre-stored design terrain model as a reference.

[0028] In the data acquisition stage, the unmanned aerial laser radar can be a lightweight laser scanning system carried by a multi-rotor unmanned aerial vehicle platform. The ground three-dimensional laser scanner can be a pulse scanning measurement system. The unmanned aerial vehicle flies at a predetermined route above the operation area, the flight height is maintained at 80-100 meters, the scanning field angle is 60 degrees, and the point cloud density is 200 points per square meter. The ground scanner is installed on a stable tripod and placed at the bottom of the ditch and the edge of the cliff, etc. The scanning distance is set to 50-150 meters, and the key topography is scanned in multiple stations. The point cloud data obtained by the two devices is stored in standard LAS format.

[0029] In the data processing stage, the point cloud fusion registration can use professional point cloud processing software running on a workstation with a high-performance graphics processor. The software first coordinates the point cloud data obtained by the unmanned aerial vehicle and the ground scanner, and performs initial registration through feature point matching. Then, the iterative closest point algorithm is used for fine registration, the iteration number is set to 100 times, and the registration error threshold is set to 5 cm. The registered point cloud is grid processed to generate a digital elevation model with a resolution of 3 cm. During processing, the point cloud in the overlapping area is checked for consistency, and abnormal points with a height difference greater than 4 cm are removed.

[0030] In the verification and correction stage, the digital design topographic model provided by the engineering design party can be a three-dimensional curved surface model in CAD format. The design model and the measured digital elevation model are imported into the same coordinate system for superposition and comparison, and the height difference between the two is detected by spatial analysis function. When the height deviation of the detected area exceeds 12 cm, the measurement engineer will manually analyze and judge according to the actual situation. During the correction process, the measured terrain data is given priority, and the areas in the design model that do not obviously conform to the actual terrain characteristics are adjusted appropriately, and finally a design topographic model that conforms to the actual terrain is generated as a reference.

[0031] Through the cooperative operation of unmanned aerial vehicles and ground scanning equipment, the problem of visual blind area in complex terrain existing in single scanning mode is effectively overcome, ensuring the integrity and accuracy of the initial topographic data. Using professional point cloud processing procedures, accurate fusion of multi-source data is achieved, and a high-resolution digital terrain model is established. Through the manual verification and correction link, the significant deviation between design and field is further eliminated, providing a reliable reference for subsequent real-time excavation and filling balance control, and ensuring the operation accuracy of the entire system from the data source.

[0032] In some other technical solutions, the fusion and registration of the point cloud data includes an accuracy control step. In the fusion registration process, the registration error between the point cloud data of the unmanned aerial laser radar and the ground three-dimensional laser scanner is calculated; when the registration error is greater than 3 cm to 5 cm, the iterative closest point algorithm is used for re-registration; the point cloud data after re-registration is subjected to overlap area consistency check, and the point cloud data with an elevation difference of more than 2 cm to 4 cm in the overlap area is eliminated; in the finally fused point cloud data, the registration error of 90% to 95% of the point positions is not greater than 1.5 times the resolution of the initial digital elevation model.

[0033] In the above technical solution, in the registration error calculation and re-registration stage, a professional three-dimensional data processing software with point cloud registration function can be selected. The software runs on a workstation configured with a sixteen-core processor and sixty-four GB of memory. The software first calculates the initial registration error between the unmanned aerial vehicle and the ground scanning point cloud, and when it is detected that the registration error reaches 4 cm, the iterative closest point algorithm is automatically triggered for re-registration. In the re-registration process, the maximum number of iterations is set to two hundred, and the convergence threshold is 0.001 meters, and by continuously optimizing the rigid body transformation matrix, the matching error between the two point clouds is minimized.

[0034] In the overlap area consistency check link, a point cloud difference analysis algorithm can be selected to process the data after re-registration. The algorithm first establishes an index of the overlap area of the two point clouds, and then calculates the three-dimensional Euclidean distance between the corresponding point pairs. For point pairs with an elevation difference of more than 3 cm, the system automatically marks them as inconsistent points. The inconsistent point elimination process adopts a bidirectional checking mechanism, i.e. difference comparison is performed simultaneously from both the unmanned aerial vehicle point cloud and the ground point cloud to ensure the uniformity and fairness of the elimination standard.

[0035] In the final precision control link, a point cloud quality evaluation module can be selected to statistically analyze the fused overall data. The module randomly selects one thousand uniformly distributed sample points from the fused point cloud, and calculates the positional deviation between them and the reference point cloud. When it is detected that the registration error of 92% of the sample points is controlled within 2 cm, it is determined that the fusion result meets the requirements. For the areas that do not meet the requirements, the system generates an accuracy distribution map to guide the operator to carry out local fine processing.

[0036] By establishing a strict registration precision control process, the error accumulation problem in the multi-source point cloud data fusion process is effectively solved. The automatic re-registration mechanism ensures the accuracy of point cloud matching, and the overlap area consistency check eliminates the systematic deviation between different data sources. Finally, the statistical quality control method ensures the overall reliability of the fused data, providing an accurate data basis for subsequent digital elevation model generation, thereby improving the quality of the benchmark data of the entire earthwork excavation and filling balance control system.

[0037] In some other technical solutions, a model updating module is arranged in the vehicle-mounted processing unit, which dynamically updates the pre-stored design terrain model based on real-time point cloud data; wherein the dynamic updating follows the following steps: After completing one path adjustment, the vehicle-mounted processing unit performs secondary difference calculation on the real-time point cloud data collected and registered in this round of operation and the design terrain model before dynamic updating, to generate a terrain incremental data set; Statistical analysis is performed on the terrain incremental data set to calculate the average change height and the continuous area of the change region; only when the absolute value of the average change height is greater than 5 cm to 10 cm, and the continuous area of the change region is greater than 2 m 2 to 5 m 2 , is the incremental data determined to be valid; The terrain incremental data set that passes the validity verification is superimposed into the pre-stored design terrain model in a weighted fusion manner to complete the iterative updating of the model; The model updating module further includes an updating stability control mechanism; the mechanism records the average change height of the current terrain incremental data set at each iterative updating, and compares it with the change heights recorded in the previous two updates; when the sign of the change height sequence of the three consecutive updates alternates between positive and negative, and the absolute values of the changes of the adjacent two times are both greater than 8 cm to 12 cm, it is determined to be oscillation and the updating amplitude limit is started; the updating amplitude limit is realized by introducing a weighting factor starting from 0.1 and increasing by 0.1 step by step until 0.5, which is used to attenuate the fusion weight of the terrain incremental data set of this time.

[0038] In the above technical solutions, in the data acquisition and difference calculation stage, an industrial-grade vehicle-mounted computer with point cloud processing function can be selected as the hardware platform. The computer is installed in a shockproof cabinet in the cab of the mine truck, and is connected with the three-dimensional laser scanner through a gigabit Ethernet interface. When the mine truck completes one path adjustment, the system automatically starts the data acquisition process, and registers the latest acquired real-time point cloud data with the design terrain model stored in the solid state disk. After registration is completed, the elevation difference between the two is calculated by a spatial difference algorithm to generate a terrain incremental data set, which is stored in the system memory in the form of a two-dimensional matrix, and each grid cell corresponds to a 10 cm by 10 cm area on the actual ground.

[0039] In the validity determination link, an embedded analysis module can be selected to perform statistical processing on the terrain incremental data set. The module first calculates the average height change of all grid cells in the data set. When the average change height reaches 8 cm, further connected region analysis is performed. The system uses an eight-neighborhood region growing algorithm to identify continuous change regions and calculates the projected area of each connected region. The system sets the validity determination threshold as the average change height not less than 8 cm and the continuous change area not less than 4 square meters. Only the data that meets both conditions will be marked as valid incremental data.

[0040] In the model update execution phase, a weighted fusion algorithm can be selected to integrate the valid incremental data into the original model. The algorithm assigns a weight of 0.3 to the new incremental data and a weight of 0.7 to the original model data, and calculates the new elevation value of each grid cell by weighted average. During the update process, the system establishes a data version management mechanism, and a backup file is generated each time the update is performed, ensuring that the system can roll back to the previous stable version in case of an exception. The entire update process is executed in the background thread of the vehicle-mounted computer, without affecting the normal operation of other real-time control functions.

[0041] In the oscillation detection phase, an embedded processing module with real-time analysis function can be selected. The module is installed in the expansion slot of the vehicle-mounted processing unit and communicates with the main processor through the PCIe interface. The module runs a specially written oscillation detection algorithm, which continuously records the average change height of each terrain incremental data set and stores the last three update data in a circular buffer. When the system detects that the change height values of the last three updates are +12 cm, -11 cm, and +13 cm in sequence, the system automatically determines that an oscillation pattern has occurred, and the alternation of change direction and the change amplitude both exceed the set threshold of 10 cm.

[0042] In the update amplitude limitation execution link, a programmable logic controller or a microprocessor with similar functions can be selected. The processor is installed on the circuit board of the vehicle-mounted control unit and connected with the main system through a digital signal interface. When receiving the oscillation determination signal, the processor starts the weighted factor control program, which starts from the initial value of 0.1 and gradually increases the weighted factor by 0.1 each time. In each model update, the system multiplies the current weighted factor with the terrain incremental data set to achieve gradual control of the update amplitude until the weighted factor reaches the maximum value of 0.5.

[0043] In the update history management link, a non-volatile memory can be selected as the data storage medium. The memory is installed in the shockproof case of the vehicle-mounted system and connected with the main system through a SATA interface. The system records the time stamp, change data and weighted factor value of each update in a fixed format to form a complete history sequence. These data are stored in a ring buffer mode, retaining the last one hundred update records, while important data are backed up to the vehicle-mounted solid state disk at regular intervals for subsequent analysis and diagnosis.

[0044] By establishing a model update mechanism based on strict effectiveness criteria, the dynamic synchronization of the designed terrain model and the actual terrain is realized, ensuring the present situation of the reference data. The statistical analysis method is used to screen the change area, effectively distinguishing the real earthwork change from the temporary interference, and avoiding the model distortion caused by data noise. The weighted fusion method maintains the stability of the model while absorbing new data, and the version management mechanism provides data security protection, which together constitute a reliable dynamic maintenance system of terrain model.

[0045] By establishing an oscillation detection mechanism based on continuous change trend analysis, abnormal fluctuation state in the model update process can be identified in time. The gradual adjustment of the weighted factor control strategy effectively suppresses the update amplitude while maintaining the model update ability, avoiding the model distortion problem caused by excessive update. Complete update history records provide data support for system state monitoring and fault diagnosis, which together constitute a model update system that can maintain long-term stable operation.

[0046] In some other technical solutions, the step two of generating the real-time excavation and filling distribution map comprises the following steps: The vehicle-mounted processing unit converts the real-time point cloud data into a grid digital elevation model with a resolution of 10 cm to 20 cm; The grid digital elevation model and the pre-stored designed terrain model are subjected to elevation difference calculation on a grid unit basis in the same coordinate system to generate an elevation difference matrix; The elevation difference matrix is subjected to region growing algorithm processing based on neighborhood connectivity, and adjacent grid units with an elevation change value greater than 15 cm to 30 cm are clustered into independent change areas; The average elevation change value of each independent change area is calculated, and the area with a negative average elevation change value is marked as an excavation area, and the area with a positive average elevation change value is marked as a filling area; The spatial position information of all excavation and filling areas is combined to generate a real-time excavation and filling distribution map.

[0047] In the data preprocessing stage, an industrial control computer with point cloud processing function can be selected. The computer is installed on the standardized guide rail of the vehicle-mounted control cabinet and connected with the scanner through a copper core shielded cable. The system performs down-sampling on the real-time collected point cloud data through voxel grid filtering to generate a grid digital elevation model with a resolution of 15 cm. The elevation value of each grid unit is obtained by averaging all the point cloud data falling into the grid, and the point cloud density information of each grid unit is recorded, and the grid with a point cloud density lower than 5 points per square meter is specially marked.

[0048] In the elevation difference calculation stage, a professional geographic information system core processing library can be selected. The software module runs on a special computing card of the vehicle-mounted computer and accesses the data files of the grid digital elevation model and the pre-stored design terrain model through memory mapping. The system first unifies the two models to the same plane coordinate system and elevation reference, and then performs elevation subtraction operation on each grid unit to generate an elevation difference matrix. For the grid unit with an elevation change value exceeding 20 cm, the system marks it as a significant change area and establishes an index for subsequent processing.

[0049] In the region identification and classification stage, a computer vision algorithm library with image processing function can be selected. The algorithm library is integrated into the main program through an application program interface and uses eight-neighborhood connectivity criterion for region growing processing. The system clusters adjacent significant change grids into independent regions, and the minimum area threshold of each region is set to 3 square meters. The average elevation change value of all grids inside each independent region is calculated, and when the average change value is negative and less than-5 cm, it is marked as a excavation area, and when the average change value is positive and greater than 5 cm, it is marked as a fill area. The final real-time excavation and filling distribution map is output in the form of a vector layer, containing the geometric boundary and attribute information of each region.

[0050] By establishing a systematic processing flow from point cloud to grid, the rapid identification and accurate quantification of the terrain change area are realized. The grid-based data processing method improves the calculation efficiency and ensures the real-time response of the system. The region growing algorithm based on neighborhood connectivity can accurately identify continuous change regions and avoid scattered false judgments. Combined with the automatic classification mechanism of the elevation change direction, the intelligent discrimination of excavation and filling areas is realized, providing a reliable data basis for subsequent earthwork balance calculation.

[0051] In some other technical solutions, the calculation of the current excavation volume and the fill volume in step three is realized by the following steps: The vehicle-mounted processing unit extracts the boundary contour of all excavation regions and fill regions based on the real-time excavation and filling distribution map; Triangular meshing is performed on each region to generate an irregular triangular mesh model; Based on the spatial relationship between the triangular mesh model and the pre-stored design terrain model, the prismatic volume integration method is used to calculate the volume of each excavation area and the volume of each filling area. During the calculation process, a real-time calibration mechanism based on the on-board weighing system is introduced. This mechanism achieves the following: during the loading process of the new energy mine truck, the on-board processing unit synchronously acquires the actual load mass data measured by the on-board weighing system, and dynamically updates the earthwork density value in combination with the calculated volume of the corresponding excavation area. The updated earthwork density value is applied to the subsequent volume calculation process, realizing dynamic calibration of the excavation and filling volume calculation.

[0052] In the boundary extraction and gridding processing stage, an industrial control computer with computer vision processing function can be selected. The computer is installed in the 19-inch standard rack of the on-board control cabinet and connected with the data acquisition system through shielded twisted pair. The system first reads the vector boundary data of each area from the real-time excavation and filling distribution map, simplifies the boundary using the Douglas-Peucker algorithm, and sets the simplification distance tolerance to 0.5 meters. Then, each area is subjected to constrained Delaunay triangulation to generate an irregular triangular mesh model composed of triangular patches, with the maximum triangular side length limited to 1.5 meters to ensure that the model can accurately describe the terrain features without being overly complex.

[0053] In the volume calculation stage, a professional three-dimensional spatial analysis software library can be selected. The software library runs in the 64-bit operating system environment of the on-board computer and is integrated into the main control program through the application program interface. Based on the generated triangular mesh model and the pre-stored design terrain model, the prismatic volume integration method is used for calculation. The specific process is to project each triangular mesh onto the design terrain surface to form a triangular prism, calculate the volume of each triangular prism, and accumulate the total volume of the region. During the calculation process, the direction of each triangular prism is judged, the volume of the excavation area is taken as negative, and the volume of the filling area is taken as positive. Finally, the total excavation volume and the total filling volume are respectively obtained.

[0054] In the density calibration stage, resistance strain type weighing sensors installed on the suspension system of the mine truck can be selected. These sensors are connected to the on-board weighing terminal through a waterproof junction box, and the weighing terminal communicates with the main control computer through the CAN bus. During the loading process of the mine truck, the system real-time acquires the weighing data, and when the detection of the load mass stability time exceeds 3 seconds, the mass data at this time is recorded. The system compares the mass data with the calculated volume of the corresponding excavation area, obtains the current earthwork density value by dividing the mass by the volume, and uses the density value to calibrate the subsequent volume calculation results.

[0055] The complete processing chain from boundary extraction to volume calculation is established to realize the accurate quantification of excavation and filling volume. The triangular mesh model can better adapt to complex terrain features and improve the accuracy of volume calculation. Combined with the real-time density calibration mechanism of the vehicle-mounted weighing system, the calculation error caused by the change of soil and stone density is effectively eliminated, ensuring the reliability of the excavation and filling balance judgment. The entire calculation process realizes automatic processing, providing accurate data support for soil and stone balance regulation.

[0056] In other technical solutions, the soil density real-time calibration mechanism further includes an error compensation step; the vehicle-mounted processing unit establishes a soil and stone density error propagation model, which considers the measurement error of the vehicle-mounted weighing system of 1% to 2% and the volume calculation error of point cloud data of 2% to 3%; when dynamically updating the soil and stone density value, the Kalman filtering algorithm is used to filter the density measurement value; the filtered density value is combined with the error propagation model to compensate and correct the calculation results of the excavation volume and the filling volume; after compensation and correction, the estimated range of volume calculation error is between 3% and 5%.

[0057] In the error propagation model establishment stage, an industrial control computer with matrix operation capability can be selected. The computer is installed in the shockproof cabinet of the vehicle-mounted system and connected with each sensor interface module through the backplane bus. The system constructs a mathematical model based on error propagation theory, which sets the measurement error of the vehicle-mounted weighing system to 1.5% and the volume calculation error of point cloud data to 2.5%. The influence relationship of these error sources on the final density calculation is described by establishing a covariance matrix, and the matrix parameters are initialized and set according to the sensor calibration data and historical measurement statistical results.

[0058] In the Kalman filtering processing stage, an embedded processor with digital signal processing function can be selected. The processor is installed on the special slot of the main control board and communicates with the main processor through the high-speed serial interface. The filtering algorithm sets the process noise variance to 0.01 and the measurement noise variance to 0.04, and recursively processes the density measurement value through state prediction and state update steps. Each time new density measurement data is input, the filter will calculate the optimal density estimation value according to the system model and observation model, and update the error covariance matrix at the same time.

[0059] In the compensation correction execution stage, an arithmetic logic unit with floating point operation function can be selected. The unit is integrated in the vehicle-mounted processor and connected with the cache memory through an internal bus. The system substitutes the density estimation value output by the Kalman filter into the error propagation model to calculate the correction coefficient of the current volume calculation result. The correction coefficient is applied to the real-time volume calculation result through a multiplier, and the corrected result is range-checked to ensure that the output value is within a reasonable range. The volume data compensated and corrected is stored in a dual-port RAM for subsequent processing.

[0060] By establishing a complete error propagation model, the system can accurately describe the cumulative effect of errors in each measurement link on the final result. The Kalman filter algorithm is used to optimize the density measurement value, effectively reducing the interference of random errors. The compensation correction mechanism based on the error model significantly improves the accuracy of the volume calculation result, providing a more reliable data basis for excavation and filling balance control. The entire error compensation process is automated, ensuring that the system can maintain stable measurement accuracy under complex working conditions.

[0061] In other technical solutions, the determination of the adjusted path in step three is achieved by the following steps: The vehicle-mounted processing unit identifies the target area with the largest difference between excavation volume and filling volume based on the real-time excavation and filling distribution map, and takes the center coordinates of the target area as the end point of the path; A path planning algorithm is used to search for a path in the obstacle map generated from real-time point cloud data. The cost function of the path planning algorithm includes two calculation items: path length cost value and terrain slope cost value; The initial path output by the path planning algorithm is subjected to safety verification based on vehicle dynamics constraints. This verification excludes path segments that exceed the maximum climbing angle of the new energy mine truck by 30% to 35%, and excludes path segments that are less than the minimum turning radius of the new energy mine truck by 8 m to 10 m; The path that passes the safety verification is processed by a curve fitting algorithm to generate an executable path with continuous curvature; During the execution of the executable path by the new energy mine truck, the three-dimensional laser scanner performs real-time obstacle detection within a range of 20 m to 30 m in the direction of travel at a frequency of 5 Hz to 10 Hz; When an obstacle is detected, the vehicle-mounted processing unit initiates local path re-planning, generates an obstacle-avoiding path, and continues execution.

[0062] In the target area identification and path planning stage, a vehicle-mounted industrial computer with parallel computing capability can be selected. The computer is installed in a shockproof cabinet at the rear of the cab and connected with the perception system through high-speed Ethernet. The system first calculates the volume difference of each region based on the real-time excavation and filling distribution map, selects the region with the largest difference value as the target, and takes the center of the circumscribed rectangle as the path endpoint. The A* algorithm is used for path planning, and the search is performed in the obstacle map generated by real-time point cloud. The path length weight in the algorithm cost function is set to 0.6, the terrain slope weight is set to 0.4, and the search step is set to 0.5 meters.

[0063] In the safety verification and path optimization stage, a path processing module with geometric analysis function can be selected. The module runs in the real-time operating system of the vehicle-mounted computer and communicates with the planning module through shared memory. The system checks the initial path planned and excludes road segments with a slope exceeding 32% and road segments with a turning radius less than 9 meters. The path that passes the verification is smoothed by B-spline curve fitting algorithm, the curve order is set to 3, and the control point interval is set to 2 meters, ensuring that the generated path curvature is continuous and meets the vehicle kinematics characteristics.

[0064] In the path execution and obstacle response stage, a combination of three-dimensional laser scanner and real-time control system can be selected. The scanner is installed on the front protection frame of the vehicle and connected with the control unit through a special cable. During the vehicle driving along the planned path, the scanner performs fan scanning on the front 25 meters range at a frequency of 8Hz, and when the detected obstacle distance from the path center line is less than 1.5 meters, the system immediately starts local re-planning. The local planning uses the same path search algorithm to generate an obstacle-avoiding path between global path points, ensuring continuous vehicle operation.

[0065] Through the integrated path planning and safety verification mechanism, the feasibility of the planned path in complex terrain is ensured. The multi-factor weighted cost function balances the length and safety of the path, and the strict dynamic constraint verification avoids the vehicle from getting stuck in impassable working conditions. Real-time obstacle detection and rapid re-planning capability ensure the continuity of the driving process. The whole path determination method realizes efficient operation under the premise of safety.

[0066] In some other technical solutions, the path planning algorithm adopts a hierarchical planning strategy; the hierarchical planning strategy includes a global planning layer and a local planning layer; the global planning layer adopts a path planning algorithm to perform coarse-grained path planning with a resolution of 1 m to 2 m, and the planning period is 10 s to 15 s; the local planning layer adopts the same path planning algorithm to perform fine-grained path planning with a resolution of 0.2 m to 0.5 m, and the planning period is 1 s to 2 s; the global planning layer and the local planning layer are coupled through a path point sequence, and the local planning layer performs real-time fine planning between global path points; the vehicle-mounted processing unit allocates dynamically adjustable computing resources to the global planning layer and the local planning layer; when the frequency of triggering re-planning by the local planning layer exceeds 3 times within 1 second, the system automatically inclines more computing resources to the local planning layer.

[0067] In the hierarchical planning system construction stage in the above technical solution, a vehicle-mounted industrial computer with multi-core processing capability can be selected as a hardware platform. The computer is installed in a shockproof control cabinet and connected with each functional module through a backplane bus. The system divides the planning task into two independent layers: the global planning layer runs on two special computing cores to process the terrain data of the entire work area with a resolution of 1.5 meters, and the planning period is set to 12 seconds; the local planning layer runs on another group of computing cores to process the area around the vehicle with a resolution of 0.3 meters, and the planning period is set to 1.5 seconds. The two layers exchange path point sequence data through shared memory.

[0068] In the planning task execution stage, the same path planning algorithm can be selected but run at different scales. The global planning layer adopts the A* algorithm to process a low-resolution obstacle map to generate a rough path sequence containing 20 to 30 path points, and each path point is spaced 8 to 10 meters apart. The local planning layer receives the global path point sequence and performs fine planning between adjacent global path points to generate a detailed path containing 5 to 8 path points, and the path point spacing is controlled to be 1 to 2 meters. Both layers of planners adopt the same cost function configuration to ensure consistency of the planning strategy.

[0069] In the computing resource dynamic allocation stage, a system resource management module can be selected for real-time monitoring and allocation. The module runs on an independent management core and continuously monitors the frequency of re-planning of the local planning layer. When it is detected that the number of re-planning reaches 3 times within 1 second, the resource management module automatically adjusts the allocation ratio of the computing cores to temporarily allocate the two computing cores of the global planning layer to the local planning layer. The system also sets a duration threshold for resource allocation, and when the high-frequency re-planning state lasts for 5 seconds, the system automatically restores the original resource allocation scheme.

[0070] By establishing a hierarchical planning architecture, the rational allocation of different granularity path planning tasks is realized, which not only ensures the rationality of the global path, but also ensures the real-time of local obstacle avoidance. A dynamic resource allocation mechanism is adopted, so that the system can adaptively adjust the computing resources according to the actual working conditions, and the response speed of local planning is prioritized in complex environments. This design effectively balances the contradiction between planning quality and computing efficiency, and provides reliable technical support for the stable operation of vehicles in steep and narrow areas.

[0071] In some other technical solutions, a system stability monitoring and control mode switching step is further included; The vehicle-mounted processing unit monitors the following system key indicators in real time: the fluctuation frequency of the excavation-filling balance ratio, the data integrity rate of real-time point cloud data, and the triggering frequency of local path re-planning; When the following conditions are met simultaneously, it is determined that the system is in an unstable risk state: the fluctuation frequency of the excavation-filling balance ratio exceeds 10 times per minute, and the data integrity rate of real-time point cloud data is less than 85% within 30 consecutive seconds, and the triggering frequency of local path re-planning exceeds 5 times per minute; When the system is determined to be in an unstable risk state, the vehicle-mounted processing unit automatically switches from the full-automatic control mode to the auxiliary control mode; In the auxiliary control mode, the vehicle-mounted processing unit simplifies the adjustment path to a straight line path from the current position to the nearest excavation-filling area, and limits the adjustment speed to 5-8 km / h; The system continuously monitors the key indicators, and when the indicators return to normal and remain stable for 10 seconds, it automatically switches back to the full-automatic control mode.

[0072] In the system state monitoring stage of the above technical solution, a multi-channel data acquisition card can be selected in combination with an industrial-grade microprocessor. The monitoring system is installed in a special slot of the vehicle-mounted control cabinet, and is connected with various sensors through analog input modules and digital input modules. The system collects excavation-filling balance ratio data at a frequency of 1 Hz, calculates the fluctuation frequency per minute, simultaneously calculates the point cloud data integrity rate, i.e. the ratio of effective points to theoretical total points, and records the triggering frequency of local path re-planning. These data are transmitted to the main processing unit through the RS485 bus for real-time analysis and storage.

[0073] In the unstable state determination stage, a programmable controller with logic judgment function can be selected. The controller is installed on the guide rail of the shockproof case and connected with the main system through an optical coupling isolation module. The system sets three threshold values for unstable state determination: the fluctuation frequency of the excavation-filling balance ratio is 12 times per minute, the point cloud data integrity rate is 85%, and the triggering frequency of local path re-planning is 6 times per minute. When these three conditions are met simultaneously within 40 consecutive seconds, the controller outputs an unstable state signal, which is transmitted to the system main controller through a relay module.

[0074] In the control mode switching and recovery stage, a dual-mode programmable controller can be selected. The controller is installed adjacent to the main control board and controls mode switching through a digital output module. When switching to the auxiliary control mode, the system uses a straight path planning algorithm to select the excavation and filling area closest to the current position of the vehicle. The straight distance calculation uses the Euclidean algorithm. The vehicle speed is limited to 7 kilometers per hour. The system continuously monitors key indicators, and when all indicators return to normal and remain stable for 12 seconds, the controller automatically switches back to the fully automatic control mode.

[0075] By establishing a multi-parameter joint determination system stability monitoring mechanism, abnormal operating conditions of the system can be identified in a timely manner. An automatic mode switching strategy is used to quickly degrade to a more reliable control mode when the system is at risk of instability, ensuring job safety. Setting reasonable recovery conditions ensures that the system can automatically recover to normal operation after the working conditions improve, reducing the need for manual intervention. This complete monitoring and protection mechanism improves the adaptability and reliability of the system in complex working conditions.

[0076] Example 1 In the steep and narrow mine area, before implementing the method, first deploy the system. Install a 32-line mechanical rotating laser radar on the top of the new energy mine truck cab protection frame. The device is fixed through a rigid support, and the scanning center axis is inclined downward by 8 degrees. Install an industrial vehicle-mounted computer in the shockproof cabinet inside the cab, equipped with a multi-core CPU and a special GPU computing module, connected with the laser radar through a gigabit Ethernet, and communicates with the vehicle control system through a CAN bus. The control system directly drives the electric drive mechanism and steering mechanism of the mine truck to complete the hardware platform construction.

[0077] Before starting the operation, use a multi-rotor unmanned aerial vehicle to carry a lightweight laser scanning system to scan the entire operation area. The flight height is maintained at 90 meters, and the point cloud density is set to 200 points per square meter. At the same time, use a pulse scanning measurement system on the ground to supplement the scanning of steep cliffs and narrow gullies, with a scanning distance of 100 meters. Import the obtained multi-source point cloud data into professional point cloud processing software, and fuse and register through the iterative closest point algorithm. The registration error threshold is set to 4 cm, and the point cloud data with a height difference exceeding 3 cm is removed. Finally, an initial digital elevation model with a resolution of 3 cm is generated. Superimpose this model with the CAD format three-dimensional surface model provided by the engineering design party, and manually check and correct the areas with a height deviation exceeding 12 cm. The pre-stored design terrain model is generated and stored in the vehicle-mounted computer solid state disk.

[0078] During real-time operation, the 3D laser scanner continuously scans the terrain at a frequency of 30 Hz for a 80-meter radius annular area, obtaining real-time point cloud data. The on-board processing unit completes data processing within 0.8 seconds: first, it converts the real-time point cloud into a grid digital elevation model with a resolution of 15 cm, and performs elevation difference calculation with the pre-stored design terrain model to generate an elevation difference matrix; then it uses the region growing algorithm to cluster adjacent grid cells with an elevation change greater than 20 cm into independent change areas; according to the positive and negative of the average elevation change value, it labels the excavation area and the filling area respectively, and generates a real-time excavation and filling distribution map. Based on this distribution map, the regional boundary contour is extracted, and an irregular triangular mesh model is generated by triangular meshing processing. The prism volume integration method is used to calculate the excavation volume and filling volume, and a real-time calibration mechanism based on the on-board weighing system is introduced to measure the soil density. Combined with the Kalman filter algorithm, the density measurement value is filtered to realize dynamic calibration and error compensation of volume calculation.

[0079] When the calculated excavation and filling balance ratio reaches 0.97, the system starts the path planning program. The on-board processing unit identifies the target area with the largest difference in excavation and filling volume based on the real-time excavation and filling distribution map, and uses a hierarchical planning strategy for path planning. The global planning layer performs coarse-grained path planning at a resolution of 1.5 meters, with a planning period of 12 seconds; the local planning layer performs fine-grained path planning at a resolution of 0.3 meters, with a planning period of 1.5 seconds. The path planning uses the A* algorithm, and the cost function includes two calculation items: path length and terrain slope, with weights of 0.6 and 0.4 respectively. The planned path is subjected to safety verification to exclude path segments with a slope greater than 32% and path segments with a turning radius less than 9 meters, and then a B-spline curve fitting is performed to generate an executable path with continuous curvature. During the execution of the path by the vehicle at a speed of 8 kilometers per hour, the 3D laser scanner performs real-time obstacle detection at a frequency of 8 Hz for a range of 25 meters in front of the vehicle, and immediately starts local path re-planning when an obstacle is detected.

[0080] In the model updating link, the system performs second difference calculation on the latest collected real-time point cloud data and the current design terrain model after completing each path adjustment to generate a terrain incremental data set. When the average change height reaches 8 cm and the continuous change area reaches 4 square meters, it is determined as valid incremental data, which is fused and updated to the design terrain model with a weight of 0.3. At the same time, the system monitors the updating trend through a stability control mechanism, and when the direction alternates for three consecutive updates and the change amplitude exceeds 10 cm, it starts to limit the update amplitude, using a weighted factor that starts from 0.1 and gradually increases to control the update amplitude.

[0081] The system stability monitoring module monitors the fluctuation frequency of the cut-fill balance ratio, the completeness rate of the point cloud data, and the local path re-planning trigger frequency in real time. When the fluctuation frequency is detected to be 12 times per minute, the data completeness rate is lower than 85%, and the re-planning trigger frequency is 6 times per minute within 40 seconds, it is determined that the system enters an unstable risk state, and automatically switches to an auxiliary control mode. In this mode, straight path planning is adopted, and the speed is limited to 7 kilometers per hour, until the system indicators return to normal and remain stable for 12 seconds, and then automatically restore the full-automatic mode.

[0082] Through the above complete implementation process, the rapid and accurate regulation and control of the cut-fill balance of new energy mine cards in steep, narrow and rare complex terrain conditions is realized.

[0083] Comparative Example 1: Traditional manual judgment and periodic measurement method The comparative example adopts a traditional operation mode that completely relies on manual experience and periodic measurement. Two surveyors are provided on site, and a total station is used to measure the terrain every 8 hours, with a measurement point spacing of 20 meters. After the measurement data is recorded manually, it needs to be returned to the office for internal processing, and a topographic map is generated through CAD software and compared with the design drawing. The entire process takes about 10 hours. The mine card driver completely relies on the gestures and whistle instructions of the on-site commander for operation, and the commander judges the cut-fill area by visual observation. The vehicle driving path is selected by the driver according to experience, and the driving speed is randomly adjusted according to the road conditions at 5-25 kilometers per hour. The design model update period is one month, and the terrain changes during this period cannot be reflected in time.

[0084] Comparative Example 2: Single platform surveying and static model method The comparative example adopts a single unmanned aerial vehicle platform for terrain surveying and maintains a static model. A fixed-wing unmanned aerial vehicle carrying a laser radar is used for aerial surveying, with a flight height of 200 meters and a point cloud density of 50 points per square meter. Due to the high flight height and model limitations, the steep cliff cannot be effectively scanned, resulting in data missing in the model. The obtained point cloud directly generates a digital elevation model with a resolution of 15 centimeters, which is not verified and corrected with ground measurement data. During the subsequent three-month operation period, the design model remains in the initial state and is not updated with the progress of the operation, resulting in systematic deviation when compared in real time.

[0085] Comparative Example 3: Simplified calculation and fixed path planning method The present comparative example adopts a simplified calculation method and a fixed path planning strategy. The earthwork volume calculation adopts an average density method, and is uniformly valued at 2.0 tons / m3 without real-time calibration. The volume calculation is based on a regular grid model, and adopts a prism formula for simplified calculation, ignoring the terrain surface characteristics. The path planning is completed once before operation, and only considers the straight line distance during planning, ignoring the terrain slope factor. The vehicle travels at a fixed speed of 12 km / h, and needs to stop and wait for manual re-planning when encountering obstacles. Due to insufficient calculation accuracy and the inability of the path to adapt to real-time working conditions, frequent manual intervention and adjustment are required.

[0086] Although embodiments of the present application have been disclosed as above, they are not limited only to the uses listed in the specification and embodiments, and can be fully applied to various fields suitable for the present application, and additional modifications can be easily made by those skilled in the art, and thus the present application is not limited to specific details and the embodiments shown and described herein, without departing from the general concept defined by the claims and the equivalent scope.

Claims

1. A steep and narrow area new energy mine card earthwork excavation and filling balance rapid regulation method, characterized in that, The application relates to a method for dynamically adjusting the operation path of a new energy mining vehicle. Step 1: A three-dimensional laser scanner installed on the new energy mining vehicle continuously scans the terrain of a ring-shaped operation area with the new energy mining vehicle as the center and a radius of 50-100 m at a scanning frequency of 20-40 Hz to obtain real-time point cloud data; Step 2: The real-time point cloud data is transmitted to a vehicle-mounted processing unit of the new energy mining vehicle through a vehicle-mounted communication system, and the vehicle-mounted processing unit completes the processing of the real-time point cloud data within 0.3-1.5 s, compares the real-time point cloud data with a pre-stored design terrain model, and generates a real-time excavation and filling distribution map; Step 3: The vehicle-mounted processing unit calculates the current excavation volume and filling volume based on the real-time excavation and filling distribution map, and calculates the ratio of the excavation volume to the filling volume as an excavation and filling balance ratio; when the excavation and filling balance ratio is lower than 0.95 or higher than 1.05, the vehicle-mounted processing unit determines an adjustment path and an adjustment speed of the new energy mining vehicle according to the real-time point cloud data, the adjustment path points to the region with the largest difference between the excavation volume and the filling volume, and the adjustment speed is controlled to be between 5-15 km / h; Step 4: The vehicle-mounted processing unit sends the adjustment path and the adjustment speed to a control system of the new energy mining vehicle, and the control system drives the electric drive mechanism and the steering mechanism of the new energy mining vehicle according to the adjustment path and the adjustment speed to execute the adjustment path and the adjustment speed.

2. The steep and narrow area new energy mine card earthwork excavation and filling balance rapid regulation method of claim 1, wherein, The pre-stored design terrain model in step 2 is established by the following steps: An unmanned aerial vehicle-borne laser radar is used to scan the whole operation area, and a ground three-dimensional laser scanner is used to supplement the scanning of steep cliff walls and narrow gully regions to obtain high-precision initial point cloud data; The point cloud data obtained by the unmanned aerial vehicle-borne laser radar and the ground three-dimensional laser scanner are fused and registered to generate an initial digital elevation model of the operation area with a resolution of 2-5 cm; A digital design terrain model provided by an engineering design party is superimposed and compared with the initial digital elevation model, regions with an elevation deviation of more than 10-15 cm are manually checked and corrected, and a pre-stored design terrain model serving as a reference is generated.

3. The steep and narrow area new energy mine card earthwork excavation and filling balance rapid regulation method of claim 2, wherein, The fusion and registration of the point cloud data include an accuracy control step; In the fusion and registration process, the registration error between the point cloud data of the unmanned aerial vehicle-borne laser radar and the ground three-dimensional laser scanner is calculated; when the registration error is greater than 3-5 cm, the iterative closest point algorithm is used for re-registration; the point cloud data after re-registration is subjected to overlap area consistency checking, and the point cloud data with an elevation difference of more than 2-4 cm in the overlap area is removed; in the finally fused point cloud data, the registration error of 90-95% of the points is not greater than 1.5 times the resolution of the initial digital elevation model.

4. The steep and narrow area new energy mine card earthwork excavation and filling balance rapid regulation method of claim 2, wherein, A model updating module is arranged in the vehicle-mounted processing unit, which dynamically updates the pre-stored design terrain model based on the real-time point cloud data; wherein the dynamic updating follows the following steps: After completing an adjustment path, the vehicle-mounted processing unit performs secondary differential calculation on the real-time point cloud data collected and registered in the current operation and the design terrain model before dynamic updating to generate a terrain incremental data set; The terrain incremental data set is statistically analyzed to calculate the average change height and the continuous area of the change region; only when the absolute value of the average change height is greater than 5 cm to 10 cm, and the continuous area of the change region is greater than 2 m 2 to 5 m 2 , the incremental data is determined to be valid; The terrain incremental data set that passes the validity verification is superimposed on the pre-stored design terrain model in a weighted fusion manner to complete the iterative update of the model; The model update module further comprises an update stability control mechanism; the mechanism records the average change height of the current terrain incremental data set at each iterative update, and compares it with the change heights recorded in the previous two updates; when the change height sequence of the last three updates alternates in sign and the absolute values of the adjacent two changes are both greater than 8 cm to 12 cm, it is determined that there is oscillation and the update amplitude limit is started; the update amplitude limit is realized by introducing a weighting factor that starts from 0.1 and increases by 0.1 step by step until 0.5, which is used to attenuate the fusion weight of the current terrain incremental data set.

5. The steep and narrow area new energy mine card earthwork excavation and filling balance rapid regulation method of claim 1, wherein, The generation of the real-time excavation and filling distribution map in step two comprises the following steps: The vehicle-mounted processing unit converts the real-time point cloud data into a grid digital elevation model with a resolution of 10 cm to 20 cm; The grid digital elevation model and the pre-stored design terrain model are subjected to elevation difference calculation on a grid cell basis in the same coordinate system to generate an elevation difference matrix; The elevation difference matrix is subjected to region growing algorithm processing based on neighborhood connectivity to cluster adjacent grid cells with an elevation change value greater than 15 cm to 30 cm into independent change regions; The average elevation change value of each independent change region is calculated, and regions with a negative average elevation change value are marked as excavation regions, and regions with a positive average elevation change value are marked as filling regions; The spatial position information of all excavation regions and filling regions is combined to generate a real-time excavation and filling distribution map.

6. The steep and narrow area new energy mine card earthwork excavation and filling balance rapid regulation method of claim 1, wherein, The calculation of the current excavation volume and filling volume in step three is realized by the following steps: The vehicle-mounted processing unit extracts the boundary contours of all excavation regions and filling regions based on the real-time excavation and filling distribution map; Triangular meshing processing is performed on each region to generate an irregular triangular mesh model; Based on the spatial relationship between the triangular mesh model and the pre-stored design terrain model, the prismatic volume integration method is used to calculate the excavation volume of each excavation region and the filling volume of each filling region; In the calculation process, a real-time calibration mechanism based on the vehicle-mounted weighing system is introduced, which realizes the dynamic updating of the earthwork density value by the following method: during the loading process of the new energy mine truck, the vehicle-mounted processing unit synchronously acquires the actual load mass data measured by the vehicle-mounted weighing system, and dynamically updates the earthwork density value in combination with the calculated volume of the corresponding excavation region; The updated earthwork density value is applied to the subsequent volume calculation process to realize dynamic calibration of the excavation and filling volume calculation.

7. The steep and narrow area new energy mine card earthwork excavation and filling balance rapid regulation method of claim 6, wherein, The earthwork density real-time calibration mechanism further comprises an error compensation step; the vehicle-mounted processing unit establishes an earthwork density error propagation model, which considers a measurement error of 1% to 2% of the vehicle-mounted weighing system and a volume calculation error of 2% to 3% of the point cloud data; when dynamically updating the earthwork density value, a Kalman filtering algorithm is used to filter the density measurement value; the filtered density value is combined with the error propagation model to compensate and correct the calculation results of the excavated volume and the filled volume; after compensation and correction, the estimated range of the volume calculation error is between 3% and 5%.

8. The steep and narrow area new energy mine card earthwork excavation and filling balance rapid regulation method of claim 1, wherein, The determination of the adjusted path in step three is achieved by the following steps: The vehicle-mounted processing unit identifies a target area with the largest difference between the excavated volume and the filled volume based on the real-time excavation and filling distribution map, and takes the center coordinates of the target area as the end point of the path; A path planning algorithm is used to search for a path in the obstacle map generated from the real-time point cloud data, and the cost function of the path planning algorithm includes two calculation items, namely the path length value and the terrain slope value; The initial path output by the path planning algorithm is subjected to safety verification based on vehicle dynamics constraints, which excludes path segments exceeding 30% to 35% of the maximum climbing angle of the new energy mining truck and path segments less than 8 m to 10 m of the minimum turning radius of the new energy mining truck; A curve fitting algorithm is used to process the path that passes the safety verification to generate an executable path with continuous curvature; During the execution of the executable path by the new energy mining truck, the three-dimensional laser scanner performs real-time obstacle detection within a range of 20 m to 30 m in the direction of travel at a frequency of 5 Hz to 10 Hz; When an obstacle is detected, the vehicle-mounted processing unit starts local path re-planning, generates an obstacle-avoiding path, and continues to execute.

9. The steep and narrow area new energy mine card earthwork excavation and filling balance rapid regulation method of claim 8, wherein, The path planning algorithm adopts a hierarchical planning strategy; the hierarchical planning strategy includes a global planning layer and a local planning layer; the global planning layer uses a path planning algorithm to perform coarse-grained path planning with a resolution of 1 m to 2 m, and the planning period is 10 s to 15 s; The local planning layer uses the same path planning algorithm to perform fine-grained path planning with a resolution of 0.2 m to 0.5 m, and the planning period is 1 s to 2 s; The global planning layer and the local planning layer are coupled through a sequence of path points, and the local planning layer performs real-time fine planning between global path points; the vehicle-mounted processing unit allocates dynamically adjustable computing resources to the global planning layer and the local planning layer; when the local planning layer triggers re-planning more than 3 times within 1 second, the system automatically tilts more computing resources to the local planning layer.

10. The steep and narrow area new energy mine card earthwork excavation and filling balance rapid regulation method of claim 1, wherein, It also includes a system stability monitoring and control mode switching step; The vehicle-mounted processing unit monitors the following system key indicators in real time: fluctuation frequency of the excavation-filling balance ratio, data integrity rate of real-time point cloud data, and trigger frequency of local path re-planning; When the following conditions are met simultaneously, it is determined that the system is in an unstable risk state: within 30 s, the fluctuation frequency of the excavation-filling balance ratio exceeds 10 times per minute, the data integrity rate of the real-time point cloud data is less than 85%, and the trigger frequency of the local path re-planning exceeds 5 times per minute; When the system is determined to be in a state of instability risk, the vehicle-mounted processing unit automatically switches from the full-automatic control mode to the auxiliary control mode; In the auxiliary control mode, the vehicle-mounted processing unit simplifies the path to a straight line path from the current position to the nearest excavation and filling area, and limits the adjustment speed to 5-8 km / h; The system continues to monitor key indicators, and when the indicators return to normal and remain stable for 10 s, it automatically switches back to the full-automatic control mode.