Multi-objective optimization mine card path planning method and system, terminal and medium

Abstract

The invention relates to the field of path planning, and particularly provides a multi-objective optimization mine truck path planning method and system, a terminal and a medium, and the method comprises the steps: constructing a comprehensive cost function according to the weighted fusion of path length cost, energy consumption cost and safety cost; periodically calling a pre-trained road condition prediction model, inputting the prediction period time mode features, the prediction period weather features and the current road condition features into the road condition prediction model, and outputting a mining area road condition prediction result; adjusting the weight coefficient, the energy consumption cost or the safety cost of each cost item in the comprehensive cost function according to the mining area road condition prediction result to form a new comprehensive cost function; and carrying out mine truck path planning by using the latest comprehensive cost function through an A * algorithm. According to the method, multi-target balance of mine truck path planning in distance, energy consumption and safety is achieved, meanwhile, environment changes of a mining area are dynamically adapted, the operation cost and the safety risk are effectively reduced, and the actual requirements of complex transportation of the mining area are met.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of path planning, in particular to a multi-objective optimization mine truck path planning method, system, terminal and medium. BACKGROUND

[0002] The main technical solutions for mine truck path planning in mining areas are based on traditional shortest path algorithm, improved algorithm based on single optimization objective and ordinary multi-objective optimization algorithm. The traditional shortest path algorithm only takes path length as the only optimization objective and does not consider energy consumption and safety during mine truck transportation. In a complex mining environment, the shortest path may pass through a high-energy consumption slope section or a dangerous area with low safety, which is difficult to meet the actual transportation requirements. The improved algorithm based on single optimization objective optimizes energy consumption or safety, such as establishing an energy consumption model to screen the path or avoiding known dangerous areas to plan the path. However, these solutions cannot comprehensively plan multiple objectives, which may lead to the selection of a longer path to reduce energy consumption or a significant increase in transportation distance and energy consumption to pursue safety. The ordinary multi-objective optimization algorithm, such as genetic algorithm and particle swarm algorithm, can handle multiple objectives, but it is a static optimization and cannot dynamically adapt to real-time changes in the mining environment. Moreover, it has high computational overhead and cannot meet the real-time planning requirements. SUMMARY

[0003] To solve the above problems, the present application provides a multi-objective optimization mine truck path planning method, system, terminal and medium, which realizes the multi-objective balance of distance, energy consumption and safety in mine truck path planning, dynamically adapts to changes in the mining environment, effectively reduces operating costs and safety risks, and meets the actual requirements of complex transportation in mining areas.

[0004] In a first aspect, the technical solution of the present application provides a multi-objective optimization mine truck path planning method, which includes the following steps: constructing a comprehensive cost function based on the weighted fusion of path length cost, energy consumption cost and safety cost; periodically calling a pre-trained road condition prediction model, inputting the predicted time period time feature, predicted time period weather feature and current road condition feature into the road condition prediction model, and outputting a mine area road condition prediction result; adjusting the weight coefficients of each cost term, energy consumption cost or safety cost in the comprehensive cost function according to the mine area road condition prediction result, and forming a new comprehensive cost function; using the latest comprehensive cost function to plan the mine truck path through the A* algorithm.

[0005] In an optional embodiment, In a second aspect, the technical solution of the present application provides a multi-objective optimization mine truck path planning system, which includes: a cost function construction unit configured to construct a comprehensive cost function according to a weighted fusion of the path length cost, the energy consumption cost and the safety cost; a mine area road condition prediction unit configured to periodically call a pre-trained road condition prediction model, input time pattern features, weather features and current road condition features of a prediction period into the road condition prediction model, and output a mine area road condition prediction result; a cost function updating unit configured to adjust the weight coefficients of the cost items, the energy consumption cost or the safety cost in the comprehensive cost function according to the mine area road condition prediction result, and form a new comprehensive cost function; a path planning unit configured to use the latest comprehensive cost function to perform mine card path planning by using the A* algorithm.

[0006] In a third aspect, the technical solution of the present application provides a terminal, comprising: a memory configured to store a multi-objective optimization mine card path planning program; a processor configured to implement the steps of the multi-objective optimization mine card path planning method when executing the multi-objective optimization mine card path planning program.

[0007] In a fourth aspect, the technical solution of the present application provides a computer readable storage medium, wherein the readable storage medium stores a multi-objective optimization mine card path planning program, and the multi-objective optimization mine card path planning program is executed by a processor to implement the steps of the multi-objective optimization mine card path planning method.

[0008] As can be seen from the above technical solutions, the present application has the following advantages: By constructing a comprehensive cost function of weighted sum of path length cost, energy consumption cost and safety cost, the three core targets of distance, energy consumption and safety in mine card transportation are integrated into a unified planning framework, which not only avoids the problems of high energy consumption and high risk caused by the shortest path, but also solves the phenomenon of losing one thing to gain another caused by single target optimization, realizes the comprehensive balance of multiple targets, and is more suitable for the actual needs of complex transportation in mine areas; By periodically calling a pre-trained road condition prediction model, inputting time pattern features, weather features and current road condition features of a prediction period, obtaining a mine area road condition prediction result, and dynamically adjusting the weight coefficients of the heuristic function or the cost items based on the result, the comprehensive cost function is always matched with the real-time road conditions, and then the A* algorithm is used for path planning, which not only ensures the planning efficiency, but also makes the planning result dynamically adapt to the dynamic changes of the mine area environment, solving the limitations of traditional static optimization; Through quantitative consideration and dynamic adjustment of energy consumption cost, unnecessary energy consumption of the mine truck in a slope, complex terrain and other road sections can be effectively reduced, and operating costs such as fuel are reduced; meanwhile, the safety cost is included and adaptively adjusted, which can guide the mine truck to avoid dangerous areas or high-risk road sections, reduce safety hazards in the transportation process, and reduce the probability of accidents. BRIEF DESCRIPTION OF DRAWINGS

[0009] In order to more clearly illustrate the technical solutions of the present application, the drawings needed to be used in the description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0010] Figure 1 A multi-objective optimization mine truck path planning method flowchart is provided for the embodiment of the present application.

[0011] Figure 2 A path planning architecture diagram is provided.

[0012] Figure 3 An LSTM network model structure diagram is provided.

[0013] Figure 4 A multi-objective optimization mine truck path planning system structure schematic diagram is provided for the embodiment of the present application.

[0014] Figure 5 A terminal structure diagram is provided for the embodiment of the present application. DETAILED DESCRIPTION

[0015] In order to make the application purpose, features and advantages of the present application more obvious and easy to understand, the technical solutions protected by the present application will be described in detail below with specific embodiments and drawings. Obviously, the following described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0016] Unless otherwise defined, all technical and scientific terms used in the present application have the same meaning as understood by those skilled in the art to which the present application belongs. The terms used in the specification of the present application are only for the purpose of describing the specific embodiments, and are not intended to limit the present application.

[0017] Figure 1 A multi-objective optimization mine truck path planning method flowchart is provided for the embodiment of the present application. Wherein, Figure 1The execution subject can be a multi-objective optimization mine truck path planning system. The multi-objective optimization mine truck path planning method provided by the embodiment of the present application is executed by a computer device, and accordingly, the multi-objective optimization mine truck path planning system runs in the computer device. According to different requirements, the order of steps in the flowchart can be changed, and some can be omitted.

[0018] As shown in the figure, the method comprises the following steps. Figure 1

[0019] S1, a comprehensive cost function is constructed according to the weighted fusion of path length cost, energy consumption cost and safety cost.

[0020] S2, a pre-trained road condition prediction model is periodically called, time pattern features, weather features and current road condition features in a prediction period are input into the road condition prediction model, and a mine area road condition prediction result is output.

[0021] S3, the weight coefficients of each cost item in the comprehensive cost function, the energy consumption cost or the safety cost are adjusted according to the mine area road condition prediction result, and a new comprehensive cost function is formed.

[0022] S4, the latest comprehensive cost function is used for mine truck path planning by A* algorithm.

[0023] Further, as a refinement and expansion of the specific implementation of the above embodiment, in order to completely describe the specific implementation process in the embodiment, another embodiment of the multi-objective optimization mine truck path planning method is provided, Figure 2 a path planning architecture schematic diagram, the real-time based Figure 2 The architecture shown in the figure is implemented, and the architecture comprises a data acquisition and modeling layer, a core algorithm layer, and an output and application layer. The data acquisition and modeling layer obtains various types of original data required for path planning, the core algorithm layer processes data, makes predictions and planning decisions, and the output and application layer executes the planning results and monitors the execution process.

[0024] The data acquisition and modeling layer comprises: A three-dimensional laser scanning / satellite communication module is used to collect mine area geographic information and construct a high-precision digital ground model. A real-time environment sensor module is deployed in the mine area to collect weather and road condition data in real time. The weather data includes temperature, humidity and rainfall; the road condition data includes road surface adhesion coefficient and flatness. A high-precision positioning device module is used to obtain real-time position information of the mine truck. A time / weather information module provides standard time information and accesses weather forecast data.

[0025] ​The data acquisition and modeling layer outputs a completed raster map of the mining area, in which each raster cell contains attributes such as terrain type, elevation, slope, and risk factor.

[0026] The core algorithm layer includes an LSTM traffic prediction model and an improved hybrid A* algorithm module. The LSTM traffic prediction model receives time, weather, and real-time traffic features from the data acquisition layer, and outputs a prediction of mine traffic conditions for a future period through a trained deep learning model. The improved hybrid A* algorithm module receives a mine grid map, traffic prediction results, and information on the starting and ending points of the mining trucks. By using an improved heuristic function that integrates path length, energy consumption, and safety, path search is performed on the grid map while considering the kinematic constraints of the mining trucks.

[0027] The output and application layer includes a mining truck navigation control unit. This unit receives the optimal path from the core algorithm layer and converts it into control commands to guide the mining truck along the planned path. Furthermore, it can continuously monitor environmental changes during the mining truck's journey and send a replanning trigger signal to the core algorithm layer when sudden changes in weather or road conditions are detected.

[0028] In this embodiment, a road condition prediction model is pre-built and trained. This model uses an LSTM network architecture. Figure 3 This is a schematic diagram of an LSTM network model. Its input is environmental information, and its output is a prediction of road conditions in the mining area, including a comprehensive environmental score. Path length influencing factor Energy consumption influencing factors and safety influencing factors Specifically, the architecture of the LSTM network is as follows: the number of neurons in the input layer is the same as the dimension of the input features; two hidden layers are set, with the first layer containing 128 LSTM units and the second layer containing 64 LSTM units; a dropout mechanism is used between the hidden layers, with a dropout rate of 0.25; the output layer is set to 4 neurons, and the output value of the first neuron is the comprehensive environmental score of the mining area during the current time period. (0~100 points), the output value of the second neuron is the path length influencing factor during that time period. (0~1), the output value of the 3rd neuron is the energy consumption influencing factor during this time period. (0~1), the output value of the 4th neuron is the safety impact factor during this time period. (0~1).

[0029] The environmental information input into the traffic prediction model includes the time pattern characteristics of the prediction period, the weather characteristics of the prediction period, and the current traffic characteristics.

[0030] To ensure the consistency of input features for the LSTM model, time features, weather features, and road condition features were standardized. Specifically: for time pattern features, using "hours" as the quantization unit (value range: 0~23, corresponding to 00:00~23:00), one-hot encoding was used to convert them into 24-dimensional binary vectors, achieving discretized representation of the time dimension; for weather features, three core quantification indicators—temperature, humidity, and rainfall—were selected, and their original value ranges were determined based on actual meteorological statistics from the mining area, then min-max normalization was used to transform them to... The normalization formula for the interval is: ,in, For the original value of the feature, , These are the minimum and maximum values ​​of this feature in the entire dataset, respectively. These are the normalized values. Specific characteristics are defined as follows: temperature is quantified as daily average temperature (unit: °C); humidity is quantified as daily relative humidity (unit: %); rainfall is quantified as daily cumulative rainfall (unit: mm). For road condition characteristics, three core quantitative indicators are selected: congestion index, road surface adhesion coefficient, and road surface smoothness. Their original value ranges are determined based on actual operational statistics of the mining area, and are also converted to the normalized values ​​described above using the min-max normalization method. The specific characteristics are defined as follows: the congestion index reflects the traffic efficiency of mining area roads, which is quantified by the actual number of mining trucks operating on the mining area roads; the road surface adhesion coefficient reflects the frictional resistance of the road surface to vehicle tires, which can be directly measured and quantified by a road surface friction coefficient tester; and the road surface smoothness reflects the vehicle driving stability, energy consumption, and path travel time, which can be directly measured and quantified by a laser smoothness tester.

[0031] When training the traffic prediction model, data such as time, weather, and traffic conditions are divided into training, validation, and test sets in a 7:2:1 ratio. The Adam optimizer is used with an initial learning rate of 0.001; the loss function is mean squared error; the batch size is set to 32; and the number of iterations is 50 epochs. During training, if the validation set loss does not decrease for 5 consecutive epochs, the learning rate is reduced to 0.5 times its original value; if it does not decrease for 10 consecutive epochs, the training is terminated early, and the optimal model parameters are saved.

[0032] In this embodiment, a high-precision raster map model of the mining area is pre-constructed, including three stages: data acquisition, data processing and rasterization, and raster cell attribute labeling.

[0033] The first phase involves multi-source data collection.

[0034] 3D laser scanning: The mining area is scanned in all directions and at high density using a ground-based or airborne 3D laser scanner to obtain high-precision point cloud data of the surface. The point cloud density is recommended to be no less than 10 points / square meter to ensure the restoration of terrain details.

[0035] Satellite remote sensing or aerial photogrammetry: Acquire satellite remote sensing images or high-resolution aerial photographs covering the entire mining area to extract large-scale ground feature information, road outlines, and boundaries of hazardous areas.

[0036] Field surveys and sensor data: For key areas, precise field measurements are conducted using equipment such as total stations and GPS-RTK. Simultaneously, sensors are deployed along mining roads to collect supplementary information such as road surface material and compaction degree.

[0037] The second stage involves data processing and raster map generation.

[0038] This stage processes the collected raw data to generate a regular raster map. First, data preprocessing is performed, including denoising, filtering, and stitching of the point cloud data to form a complete point cloud model of the mining area. Orthorectification and fusion of the remote sensing imagery are then performed to eliminate distortion. Next, based on the processed point cloud data, a digital elevation model (DEM) of the mining area is generated using an interpolation algorithm. This model stores the elevation value of each point in a regular grid. Then, based on the DEM, the entire geographic space of the mining area is divided into a regular two-dimensional raster cell array. The size of the raster cell can be set according to the accuracy requirements, preferably 1 meter × 1 meter. Each raster cell is stored as an independent data structure in the database, used to store a series of attributes.

[0039] The third stage is the annotation of raster cell attributes.

[0040] Calculate and assign the following key properties to each grid cell for use in subsequent path planning algorithms.

[0041] 1) Terrain type: Based on elevation, slope and remote sensing image interpretation, an enumerated type value is assigned to each raster.

[0042] For example, 0-flat area, 1-sloping area, 2-pitted area, 3-backfilled area, 4-waterlogged area, 5-mineral pile area.

[0043] 2) Elevation and Slope: The elevation value of the center point of the grid is extracted directly from the digital elevation model. The slope is calculated by determining the elevation difference between the grid and its neighboring grids.

[0044] 3) Risk coefficient: It is negatively correlated with the distance from the grid cell to the danger zone and positively correlated with the terrain complexity of the grid cell. Danger zones include blasting areas, steep slopes and mountain edge areas.

[0045] Specifically, the formula for calculating the risk coefficient is as follows:

[0046] In the formula, For the first The initial risk coefficient for each grid cell ranges from [value range missing]. .

[0047] The distance influences the weighting coefficient. The weighting coefficients are determined by the degree of terrain complexity, and the following conditions are met: .

[0048] For the first The normalized distance from each grid cell to the danger zone is calculated as follows: , For the first The shortest straight-line distance from the center of each grid cell to the boundary of each danger zone. This represents the maximum distance from all grid cells within the mining area to the danger zone. The closer to the danger zone, the more The larger.

[0049] For the first The normalized value of terrain complexity for each raster cell is calculated as follows: .

[0050] in, For the first The terrain complexity index of each grid cell is calculated by weighting its slope, surface roughness, and undulation. The calculation method is as follows: , This is the slope value. This is the surface roughness coefficient. For undulation, These are the weighting coefficients for each item. Among them, the slope value is an important factor affecting vehicle stability and energy consumption. The surface roughness coefficient can be quantified by the elevation standard deviation of the point cloud within the calculation unit, reflecting the degree of road bumpiness. The undulation can be quantified by the elevation range within a certain range within the calculation unit, reflecting the local undulation changes of the terrain.

[0051] 、 These represent the minimum and maximum values ​​of the terrain complexity index for all raster cells within the mining area. The more complex the terrain, the more... The larger.

[0052] 4) Accessibility indicator: Based on the terrain type and risk factor, assign a Boolean value (accessible / inaccessible) to each grid cell. For example, mark flooded areas and cliff edges as inaccessible.

[0053] The final generated raster map model of the mining area is a digital map containing several rasters and attribute information. This model serves as the static environment basis for path search using the hybrid A* algorithm.

[0054] In this embodiment, given that the hybrid A* algorithm already considers the kinematic constraints of mining trucks, energy consumption and safety objectives are further integrated into the path length objective of the original heuristic function. The energy consumption objective is based on a vehicle dynamics principle to construct an energy consumption model, while the safety objective is characterized by the sum of the risk coefficients of the grid cells traversed by the path. Specifically, to comprehensively quantify the factors that significantly impact the actual operation of mining trucks, such as transportation efficiency, vehicle energy consumption, and safety, the multi-objective function is decomposed into three parts. The first part is the path length objective, which intuitively reflects transportation efficiency; the second part is the energy consumption objective based on vehicle operating power; and the third part is the safety objective based on the risk coefficients of the grid cells.

[0055] The overall cost function is expressed as:

[0056] In the formula, To start from the initial position and pass through the current node Total cost estimate to reach the target location; From the initial position to the current node The actual total cost; To start from the current node Heuristically estimate the total cost to reach the target location.

[0057] Regarding the estimated total cost Its expression is:

[0058] in, For the node The estimated path length to the target location. For the node The estimated energy cost to reach the target location. For the node The estimated safety cost to reach the target location; These are the weighting coefficients for each cost item.

[0059] For the goal of path length, it is measured by path length.

[0060] Regarding energy consumption targets, a side estimate is made by summing the work done over a period of time using environmental factors such as frictional resistance and air resistance, expressed as: ,in For rolling resistance power, For air resistance power, The climbing resistance power, To accelerate the resistance power, The integral symbol is used. To calculate the start time, To calculate the termination time, this expression represents the sum of the vehicle's operating power over a continuous period of time.

[0061] The above energy consumption target integral is the theoretical total energy consumption model. In actual path planning, the heuristic function used... The calculations need to be simplified, including: 1) Ignoring acceleration resistance. Since the precise speed curve cannot be predicted during the path planning stage, acceleration resistance power is usually ignored. Or incorporate its impact into an average efficiency factor; 2) Ignore air resistance, given the low operating speed of mining trucks in the mining area, air resistance power This is usually negligible.

[0062] Therefore, the energy consumption cost of this embodiment The main estimation focuses on the work done by rolling resistance and climbing resistance. This estimation is based on elevation information provided by the raster map, and the calculation formula is as follows:

[0063] In the formula, Indicates starting from the current path node The estimated path to the target node; For the first Path length of a segment of the journey For the first The elevation difference between the starting and ending points of a journey. The rolling resistance coefficient, The total mass of the mining truck. This represents gravitational acceleration. It should be noted that elevation difference, rolling resistance coefficient, etc., can be obtained from the raster map.

[0064] For safety objectives, the risk coefficient attribute of the grid cells traversed by the planned path is used for measurement, and its expression is: In the formula, For the first Risk coefficient of each grid cell This represents the total number of raster cells traversed by the path. The risk factor has been explained previously and will not be repeated here.

[0065] During the operation of the mining truck, a pre-trained road condition prediction model is periodically invoked to predict the future environmental conditions of the mining area, thereby determining whether adjustments to the comprehensive cost function are necessary. Specifically, the road condition prediction model is automatically invoked at fixed time intervals (preferably 1 hour) to adapt to scenarios with periodic patterns such as mining shifts and weather changes. Simultaneously, event triggering is supported; that is, when the real-time environmental monitoring system detects sudden changes in key indicators such as rainfall and wind speed exceeding preset thresholds, the current cycle will be immediately interrupted, triggering a temporary model invocation and path replanning.

[0066] At the beginning of each prediction period, the temporal pattern features, weather features, and current road condition features for the prediction period are prepared and standardized preprocessed. The preprocessing process is the same as that during the prediction model training phase and will not be repeated here. The preprocessed feature vector is then input into the pre-trained LSTM road condition prediction model, which outputs a prediction of the mining area environment for the next hour. This prediction is a vector containing four elements: Comprehensive Environmental Score (0~100 points): Reflects the overall quality of the mining area's operating environment; Path length influence factor (0~1): Characterizes the degree of influence of the current environment on the path length target; Energy consumption influencing factors (0~1): Characterizes the degree of impact of the current environment on vehicle energy consumption targets; Safety Influencing Factors (0~1): Characterizes the degree of impact of the current environment on transportation safety objectives.

[0067] In each cycle, after the prediction model outputs the prediction results, the weight coefficients of each cost term in the comprehensive cost function, energy consumption targets, or safety targets are adjusted based on these results. Specifically, this includes: a) Analyze the combined environmental score for this period and the previous period. Whether the degree of deviation is within the preset deviation range; b) If so, no adjustment is needed; c) If not, determine whether the degree of deviation exceeds the deviation threshold; d) If so, adjust the weighting coefficients of each cost item using the following formula:

[0068]

[0069]

[0070] e) If the degree of deviation exceeds the deviation threshold and decreases relatively, the energy consumption cost and safety cost are increased by a first preset ratio; if the degree of deviation exceeds the deviation threshold and increases relatively, the energy consumption cost and safety cost are decreased by a second preset ratio.

[0071] For example, one specific way to dynamically adjust the heuristic function is to use the comprehensive environmental score of the mining area output by the LSTM model. The values ​​of environmental attribute factors are adjusted. If the predicted comprehensive environmental score of the mining area for a certain period deviates from the comprehensive environmental score of the mining area for the previous period within ±10%, no adjustment is made; if the score decreases and the relative decrease is greater than 10%, the energy consumption cost and safety cost are increased by 10%; if the score increases and the relative increase is greater than 10%, the energy consumption cost and safety cost are reduced by 10%.

[0072] The second specific method for dynamically adjusting the heuristic function is: based on the comprehensive environmental score of the mining area output by the LSTM model. The weight coefficients of the heuristic function are updated and adjusted. If the predicted comprehensive environmental score of the mining area for a certain period deviates from the comprehensive environmental score of the mining area for the previous period within ±10%, no adjustment is made; if the score deviation is greater than 10%, the path length influence factor is applied. Energy consumption influencing factors Safety influencing factors For weighting coefficients , , Updates and adjustments will be made.

[0073] This embodiment periodically detects the deviation of the environmental score and uses this as a trigger signal, enabling the system to proactively perceive changes in the environmental situation and initiate adjustments, replanning the path. This overcomes the poor adaptability of static planning in complex and variable mining environments. It also features two adjustment methods: cost term scaling and weight coefficient reallocation. Cost term scaling is a fast and efficient response; when the environment deteriorates, it directly increases energy consumption and safety costs, quickly guiding the algorithm to avoid high-risk, high-energy-consumption areas and prioritizing safety. Weight coefficient reallocation enables adaptive switching of the focus of multi-objective optimization; for example, when the model predicts that safety will become the primary concern... The factor will increase, thereby automatically increasing the weight of security. This ensures that the planned path has a high level of safety.

[0074] In this embodiment, a hybrid A* algorithm is used to search on a discrete grid map, but each node is associated with a continuous vehicle state (such as position and orientation), and motion primitives such as Reeds-Shepp (RS) curves are used to generate path segments that conform to vehicle kinematics.

[0075] Step 1, Algorithm initialization.

[0076] The algorithm receives input, including the starting point and target point, a grid map, and the latest heuristic function parameters. The starting point and target point are the current pose and target pose of the mining truck, respectively, and the latest heuristic function parameters are the comprehensive cost function dynamically adjusted based on road condition prediction results.

[0077] Add the starting node to the open list, which is used to store nodes to be explored.

[0078] Step 2, Forward exploration and node expansion.

[0079] Select comprehensive cost from the open list The smallest node is used as the current node for expansion.

[0080] Based on the kinematic model of the mining truck (such as the turning radius), starting from the state of the current node, a series of possible motion primitives (such as short arcs at different steering angles) are generated. These primitives represent the trajectory segments that the mining truck may travel in the next time period.

[0081] For each child node generated by a motion primitive, calculate its cost.

[0082] a) Calculate the actual cumulative cost For each child node generated by the motion primitive Calculate the distance from the starting position to the child node. The actual cumulative cost that has already occurred. This cost is the exact cost known during the pathfinding process, and is calculated using the following formula:

[0083] In the formula, Move from parent node to current child node The additional cost. The calculation also applies the heuristic function (that is, the estimation of the total cost). The consistent multi-objective fusion principle is calculated as follows:

[0084] In the formula, This is the length of this path segment. This is the estimated energy consumption for this route segment. This is the sum of the risk coefficients of the grid cells traversed in this path segment. Weight coefficient It uses the same dynamically adjusted value as the heuristic function.

[0085] It should be noted that the actual cumulative cost The calculation method and Maintain consistency, from the starting point to the node. The weighted sum of the path length, energy consumption, and safety costs of the route already traveled.

[0086] c) Calculate the heuristic cost Estimate from the current child node The cost of reaching the final target location That is, it is calculated using the following formula:

[0087] in, For the node The estimated path length to the target location. For the node Estimated energy cost to reach the target location.

[0088] d) Calculate the total cost of the nodes node Total cost It is the sum of the actual cost and the heuristic cost mentioned above, representing the cost of passing through the nodes. Total estimated cost to reach the destination:

[0089] according to The value of is used to prioritize the node with the lowest total cost from the open list for subsequent expansion, thereby ensuring that the search always moves in the direction of overall optimization.

[0090] The above text describes in detail an embodiment of a multi-objective optimization mining card path planning method. Based on the multi-objective optimization mining card path planning method described in the above embodiment, this invention also provides a multi-objective optimization mining card path planning system corresponding to the method.

[0091] Figure 4 This is a schematic block diagram of a multi-objective optimized mining truck route planning system provided in an embodiment of the present invention. In this embodiment, the multi-objective optimized mining truck route planning system 400 can be divided into multiple functional units according to the functions it performs, such as... Figure 4 As shown. The unit referred to in this invention refers to a series of computer program segments that can be executed by at least one processor and perform a fixed function, and which are stored in memory.

[0092] Cost function construction unit 410 is used to construct a comprehensive cost function based on the weighted fusion of path length cost, energy consumption cost, and security cost.

[0093] The mining area road condition prediction unit 420 is used to periodically call the pre-trained road condition prediction model, inputting the time pattern features of the prediction period, the weather features of the prediction period, and the current road condition features into the road condition prediction model, and outputting the mining area road condition prediction results.

[0094] The cost function update unit 430 is used to adjust the weight coefficients, energy consumption costs, or safety costs of each cost item in the comprehensive cost function based on the prediction results of the road conditions in the mining area, so as to form a new comprehensive cost function.

[0095] The path planning unit 440 is used to plan mining routes using the latest comprehensive cost function through the A* algorithm.

[0096] The multi-objective optimization mining truck path planning system of this embodiment is used to implement the aforementioned multi-objective optimization mining truck path planning method. Therefore, the specific implementation of this system can be found in the embodiment section of the multi-objective optimization mining truck path planning method above. Thus, the specific implementation can be referred to the description of the corresponding embodiments, and will not be elaborated here.

[0097] Furthermore, since the multi-objective optimized mining truck path planning system of this embodiment is used to implement the aforementioned multi-objective optimized mining truck path planning method, its function corresponds to the function of the above method, and will not be repeated here.

[0098] Figure 5 This is a schematic diagram of the structure of a terminal 500 provided in an embodiment of the present invention, including: a processor 510, a memory 520, and a communication unit 530. The processor 510 is used to implement the process steps of the above-described embodiment of the multi-objective optimized mining truck path planning method when implementing the multi-objective optimized mining truck path planning program stored in the memory 520.

[0099] This invention also provides a computer storage medium, which may be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc. The computer storage medium stores a multi-objective optimized mining truck path planning program. When executed by a processor, the multi-objective optimized mining truck path planning program implements the process steps of the above-described multi-objective optimized mining truck path planning method embodiment.

[0100] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A multi-objective optimization method for mining truck path planning, characterized in that, Includes the following steps: A comprehensive cost function is constructed by weighted fusion of path length cost, energy consumption cost, and security cost; The pre-trained road condition prediction model is periodically invoked, and the time pattern features, weather features, and current road condition features of the prediction period are input into the road condition prediction model to output the road condition prediction results for the mining area. Based on the road condition prediction results in the mining area, the weight coefficients, energy consumption costs, or safety costs of each cost item in the comprehensive cost function are adjusted to form a new comprehensive cost function. Mining card path planning is performed using the latest comprehensive cost function via the A* algorithm.

2. The multi-objective optimization mining truck path planning method according to claim 1, characterized in that, The overall cost function is expressed as: In the formula, To start from the initial position and pass through the current node Total cost estimate to reach the target location; From the initial position to the current node The actual total cost; To start from the current node The heuristic estimate of the total cost to the target location is expressed as follows: in, For the node The estimated path length to the target location. For the node The estimated energy cost to reach the target location. For the node The estimated safety cost to reach the target location; These are the weighting coefficients for each cost item.

3. The multi-objective optimization mining truck path planning method according to claim 2, characterized in that, The energy cost is expressed as: In the formula, Indicates starting from the current path node The estimated path to the target node; For the first Path length of a segment of the journey For the first The elevation difference between the starting and ending points of a journey. The rolling resistance coefficient, The total mass of the mining truck. This is the acceleration due to gravity.

4. The multi-objective optimization mining truck path planning method according to claim 2, characterized in that, The security cost is expressed as: In the formula, For the first Risk coefficient of each grid cell This represents the total number of grid cells traversed by the path.

5. The multi-objective optimization mining truck path planning method according to claim 4, characterized in that, No. The risk coefficient of each grid cell is calculated using the following formula: In the formula, The distance influences the weighting coefficient. The weighting coefficient is determined by the degree of terrain complexity. For the first Normalized distance from each grid cell to the danger zone; For the first The normalized value of terrain complexity for each raster cell is represented as: ; in, For the first The terrain complexity index of each grid cell is calculated by weighting its slope, surface roughness, and undulation. The calculation method is as follows: , This is the slope value. This is the surface roughness coefficient. For undulation, These are the weighting coefficients for each item; , These represent the minimum and maximum values ​​of the terrain complexity index for all raster cells within the mining area.

6. The multi-objective optimization mining truck path planning method according to any one of claims 2 to 5, characterized in that, The road condition prediction model outputs a comprehensive environmental score for the mining area. Path length influencing factor Energy consumption influencing factors and safety influencing factors ; Based on the mine area road condition forecast results, the weight coefficients, energy consumption targets, or safety targets of each cost item in the comprehensive cost function are adjusted, specifically including: The combined environmental score for this period and the previous period was measured. Whether the degree of deviation is within the preset deviation range; If so, no adjustment will be made; If not, determine whether the degree of deviation exceeds the deviation threshold; If so, adjust the weighting coefficients of each cost item using the following formula: Furthermore, if the degree of deviation exceeds the deviation threshold and decreases relatively, the energy consumption cost and safety cost will be increased by a first preset ratio. If the deviation exceeds the deviation threshold and increases relatively, the energy consumption cost and safety cost will be reduced according to the second preset ratio.

7. The multi-objective optimization mining truck path planning method according to claim 6, characterized in that, The road condition prediction model uses an LSTM network architecture.

8. A multi-objective optimization mining truck path planning system, characterized in that, include: The cost function construction unit is used to construct a comprehensive cost function based on the weighted fusion of path length cost, energy consumption cost, and security cost. The mine road condition prediction unit is used to periodically call the pre-trained road condition prediction model. It inputs the time pattern features of the prediction period, the weather features of the prediction period, and the current road condition features into the road condition prediction model and outputs the mine road condition prediction results. The cost function update unit is used to adjust the weight coefficients, energy consumption costs, or safety costs of each cost item in the comprehensive cost function based on the prediction results of road conditions in the mining area, so as to form a new comprehensive cost function. The path planning unit is used to plan mining routes using the latest comprehensive cost function via the A* algorithm.

9. A terminal, characterized in that, include: Memory, used to store multi-objective optimized mining card path planning programs; A processor, configured to implement the steps of the multi-objective optimized mining truck path planning method as described in any one of claims 1 to 7 when executing the multi-objective optimized mining truck path planning program.

10. A computer-readable storage medium, characterized in that, The readable storage medium stores a multi-objective optimized mining truck path planning program, which, when executed by a processor, implements the steps of the multi-objective optimized mining truck path planning method as described in any one of claims 1 to 7.

Images