Complex mining area full-coverage operation-oriented systematized path planning method

Abstract

The invention provides a systematized path planning method for full-coverage operation of a complex mining area, which comprises the following steps of: performing regional decomposition processing on a grid map of a large-scale mining area containing irregular obstacles so as to obtain a plurality of sub-regions without I-type obstacle interference and center coordinate positions of II-type obstacles in each sub-region; constructing a topological matrix among all the sub-regions, and determining an optimal access sequence for traversing all the sub-regions; determining a full-coverage path of the deep-sea mining vehicle in each sub-region according to the optimal access sequence for traversing all the sub-regions and the central coordinate position of the class-II obstacle in each sub-region; the method has the beneficial effects that a complete path planning framework is provided for the deep-sea mining vehicle to carry out efficient and continuous autonomous operation in a large-scale and multi-obstacle mining area environment; according to the method, the space continuity and the motion performability of the path are improved, and the practicability and the robustness of a path planning system in a real complex scene are remarkably enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of path planning, and in particular to a systematic path planning method for full-coverage operations in complex mining areas. Background Art

[0002] With the increasing depletion of global terrestrial mineral resources, deep-sea mining areas, as a key frontier for future strategic resource extraction, are attracting widespread attention from both the scientific and industrial communities. Deep-sea mining vehicles, as core operational equipment, must accomplish tasks such as resource coverage, obstacle avoidance, and path planning in a complex and volatile seabed environment. Full-coverage path planning, a key technology for ensuring mining efficiency and comprehensive resource extraction, faces significant challenges. For example, tracked deep-sea mining vehicles (CMMs) must navigate large-scale mining areas, navigating challenging terrain, soft subsurfaces, dense obstacles, and frequent environmental disturbances. These challenges not only impact vehicle stability and maneuverability but also place higher demands on path continuity, integrity, and optimization efficiency. Furthermore, the unique characteristics of deep-sea operations necessitate that path planning not only meet feasibility, safety, and obstacle avoidance requirements, but also balance energy consumption, path continuity, and minimal disturbance to the seabed ecosystem. Appropriate full-coverage path planning not only maximizes the effective operating area, improves mining efficiency, and controls slippage and energy consumption, but also ensures controllable and executable trajectory. Regarding the technical difficulties faced by deep-sea mining vehicles in planning paths for full coverage of large-scale mining areas, current research still faces two core challenges:

[0003] 1. Lack of a unified and efficient full-process planning framework: Although existing studies have proposed a variety of path planning strategies, such as those based on grid division, neural network modeling, hexagonal coverage, and multi-robot collaborative mechanisms, these methods mostly focus on a certain stage of path generation and have not yet formed a complete closed-loop system from mining area division, regional connection path optimization to detailed traversal within the sub-area. In particular, in the process of connecting multi-stage tasks, there are often problems of insufficient information transmission or inconsistent goals between modules, which affect the overall planning efficiency and execution effect. In addition, the existing planning methods are still insufficient in the comprehensive trade-off between energy consumption control, path continuity, and obstacle avoidance capabilities, and it is difficult to meet the high requirements of deep-sea mining operations for systematicity and coordination.

[0004] 2. Insufficient adaptability to complex and irregular environments: Most current mainstream path planning methods assume that the terrain of the operating area is regular and obstacles are sparsely distributed. They are unable to cope with the complex working conditions of deep-sea mining areas with fragmented terrain, dense obstacles, or dynamic changes. In such environments, traditional algorithms are prone to partial coverage omissions, lengthy detours, and even inability to effectively complete the task. In addition, although the randomization strategy has a certain degree of robustness, it cannot guarantee coverage efficiency and operational reliability in high-cost, long-term deep-sea operations. Therefore, improving the algorithm's ability to adapt to environmental uncertainty and complexity is a key problem that needs to be overcome in the current full-coverage path planning. Summary of the Invention

[0005] In view of the shortcomings of the existing technology mentioned above, the purpose of the present invention is to provide a systematic path planning method for full coverage operations in complex mining areas, which is used to solve the problems of the existing technology lacking a unified and efficient full-process planning framework and insufficient adaptability to complex and irregular environments.

[0006] To achieve the above-mentioned and other related purposes, the present invention provides the following technical solutions:

[0007] A systematic path planning method for full coverage operations in complex mining areas comprises the following steps: obtaining a raster map of a large-scale mining area containing irregular obstacles; performing regional decomposition processing on the raster map, and obtaining a plurality of sub-areas free of interference from Class I obstacles and the center coordinate positions of Class II obstacles in each sub-area based on the regional decomposition processing results; storing the adjacent relationships of all sub-areas in the form of an adjacency matrix, constructing a topological matrix between all sub-areas based on the adjacency matrix, and determining an optimal access sequence for traversing all sub-areas based on the topological matrix between all sub-areas; and determining a full coverage path for a deep-sea mining vehicle in each sub-area based on the optimal access sequence for traversing all sub-areas and the center coordinate positions of Class II obstacles in each sub-area.

[0008] In one embodiment of the present invention, the grid image is subjected to regional decomposition processing, and a plurality of sub-areas free of interference from Class I obstacles and the center coordinate position of a Class II obstacle in each sub-area are obtained based on the regional decomposition processing results, including: preprocessing the grid image, obtaining the geometric outline of the obstacle based on the preprocessing results, and storing it in a matrix form; and obtaining a plurality of sub-areas free of interference from Class I obstacles and the center coordinate position of a Class II obstacle in each sub-area based on the geometric outline of the obstacle and using a Boustrophedon algorithm based on obstacle features.

[0009] In one embodiment of the present invention, after preprocessing the grid map, before obtaining a plurality of sub-areas free of Class I obstacle interference and the center coordinate positions of Class II obstacles in each sub-area according to the geometric contours of the obstacles and using the Boustrophedon algorithm based on obstacle features, it also includes: classifying the obstacles in the large-scale mining area according to the area of ​​the obstacles and in combination with the unit distance operating area threshold of the mining vehicle, and obtaining Class I obstacles and Class II obstacles according to the classification results.

[0010] In one embodiment of the present invention, the grid image is preprocessed, and the geometric outline of the obstacle is obtained according to the preprocessing result and stored in a matrix form, including: filling the irregular obstacle area that only partially occupies the grid cell, and performing a cell expansion operation on the filled obstacle area, and then extracting the geometric outline of the obstacle from the expanded edge and storing it in a matrix form.

[0011] In one embodiment of the present invention, the method comprises: determining the inflection point coordinates from the geometric outline of the obstacle stored in a matrix form, and dividing the large-scale mining area into several initial sub-areas free from Class I obstacle interference according to the inflection point coordinates of the Class I obstacle; merging adjacent partitions of several initial sub-areas free from Class I obstacle interference, and obtaining several sub-areas free from Class I obstacle interference according to the merging results.

[0012] In one embodiment of the present invention, determining the optimal access order of traversing all sub-regions based on the topological matrix between all sub-regions includes: determining the optimal access order of traversing all sub-regions based on the topological matrix between all sub-regions and using an ant colony algorithm, wherein the problem of the optimal access order of traversing all sub-regions can be regarded as an open traveling salesman problem, and the ant colony algorithm is used to solve the open traveling salesman problem.

[0013] In one embodiment of the present invention, determining a full-coverage path for the deep-sea mining vehicle within each sub-area based on an optimal access sequence for traversing all sub-areas and the central coordinate positions of Class II obstacles within each sub-area includes dynamically switching between coverage search and inter-area transfer using a finite state machine based on the optimal access sequence for traversing all sub-areas and the central coordinate positions of Class II obstacles within each sub-area, optimizing the traversal path in combination with a multimodal heuristic function and a steering motion model, thereby obtaining a full-coverage path for the deep-sea mining vehicle within each sub-area.

[0014] In one embodiment of the present invention, the method of dynamically switching between coverage search and inter-area transfer using a finite state machine, and optimizing the traversal path using a multimodal heuristic function and a steering motion model, thereby obtaining a fully covered path for the deep-sea mining vehicle within each sub-area, includes: constructing multiple operating states using the finite state machine; planning a traversal path for the deep-sea mining vehicle within each sub-area based on the constructed multiple operating states; and during the path planning process, when an operating state associated with an obstacle avoidance strategy is triggered, calling a corresponding function within the obstacle avoidance strategy to plan a traversal path within each sub-area, until a fully covered path for the deep-sea mining vehicle within each sub-area is obtained.

[0015] In one embodiment of the present invention, the obstacle avoidance strategy is a three-level obstacle avoidance response system, which includes a local adjustment function, a regional replanning function and a global fallback function in sequence; the local adjustment function is successful when there is an unvisited and obstacle-free grid node in the neighborhood of the current grid node, otherwise the regional replanning function is triggered; the regional replanning function uses the A* algorithm to search for a new path in an area with a radius of R with the current node as the center, and the path cost threshold within the search range is set to u times the current path length; if a valid path cannot be found, the global fallback function is triggered; the global fallback function reversely searches the historical path, and the backtracking depth does not exceed v% of the historical path length, and falls back and replans on the condition that there is an unvisited and obstacle-free grid node in the neighborhood.

[0016] As described above, the systematic path planning method for full coverage operations in complex mining areas of the present invention has the following beneficial effects:

[0017] 1. Constructing a full-process planning system for full-coverage operation paths in large-scale mining areas: The present invention proposes a systematic solution that integrates regional decomposition of large-scale mining areas, optimization of connection paths between sub-areas, and detailed planning of traversal paths within sub-areas. For the first time, a closed-loop design from environmental modeling to path execution is realized in the full-coverage path planning task. By introducing a regional decomposition method based on the geometric characteristics of obstacles, a fine division of irregular and complex mining areas is achieved, effectively controlling the computational complexity. By converting the optimal visit sequence problem for traversing all sub-areas into an open traveling salesman problem, the overall path efficiency is improved with the help of an improved ant colony algorithm. Within the sub-areas, a finite state-driven path planning algorithm is used to ensure coverage integrity while improving dynamic response capabilities, thereby providing a complete path planning framework for mining vehicles to carry out efficient and continuous autonomous operations in large-scale, multi-obstacle mining environments.

[0018] 2. Improving the autonomous adaptability and environmental robustness of path planning: The present invention introduces a multi-level obstacle avoidance mechanism and a probabilistic grid modeling method in the path planning process, combined with a finite state machine and a dynamic heuristic function, which effectively enhances the system's adaptability to dynamic obstacles, sensor errors and environmental uncertainties. In particular, in the sub-area traversal stage, a three-level obstacle avoidance strategy of local adjustment, regional replanning and global fallback is designed to ensure that the mining vehicle can still quickly adjust its path when encountering sudden obstacles or local dead zones to avoid operation interruption. At the same time, the path smoothing function and steering optimization model are used to improve the spatial continuity and motion executableness of the path, significantly enhancing the practicality and robustness of the path planning system in real complex scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 Shown is an overall flow chart of a systematic path planning method for full coverage operations in complex mining areas disclosed in an embodiment of the present invention;

[0020] Figure 2 Shown is a detailed flow chart of the overall systemic path planning method for full coverage operations in complex mining areas disclosed in an embodiment of the present invention;

[0021] Figure 3 An environmental diagram showing an initial grid map in a systematic path planning method for full coverage operations in a complex mining area disclosed in an embodiment of the present invention;

[0022] Figure 4 Displayed is an obstacle outline map based on geometric features in a systematic path planning method for full coverage operations in complex mining areas disclosed in an embodiment of the present invention;

[0023] Figure 5 Shown is a Boustrophedon decomposition diagram based on obstacle features in a systematic path planning method for full coverage operations in complex mining areas disclosed in an embodiment of the present invention;

[0024] Figure 6 Shown is a topological relationship diagram between sub-areas in the systematic path planning method for full coverage operations in complex mining areas disclosed in an embodiment of the present invention;

[0025] Figure 7 Displayed is a planned path diagram traversing all sub-mining areas in a systematic path planning method for full coverage operations in a complex mining area disclosed in an embodiment of the present invention;

[0026] Figure 8 Shown is a traversal path graph generated by the FSM-DSCPP algorithm in the systematic path planning method for full coverage operations in complex mining areas disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0027] The following describes the embodiments of the present invention through specific examples. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. It should be noted that the following embodiments and features in the embodiments can be combined with each other unless there is a conflict.

[0028] The present invention relates to a systematic path planning method for full coverage operation in complex mining areas. The process is as follows: Figure 1 and Figure 2 As shown, the details are as follows:

[0029] Step 101: Obtain a raster map of a large-scale mining area containing irregular obstacles.

[0030] Step 102: performing a region decomposition process on the grid image, and obtaining a plurality of sub-regions free from interference of type I obstacles and the center coordinate position of the type II obstacle in each sub-region according to the region decomposition process result.

[0031] Specifically, the specific process of large-scale mining area decomposition is as follows: In order to improve the accuracy of environmental modeling for path planning, the grid map of irregular obstacles needs to be preprocessed: First, the irregular obstacle area that only partially occupies the grid cells is filled to ensure the complete representation of the obstacle in the image; then, the cell expansion operation is performed on the filled obstacle area to expand the obstacle boundary outward by one cell to enhance the safety margin; finally, the obstacle geometric outline is extracted from the expanded edge and stored in matrix form to provide structured input for subsequent path planning (such as Figure 3 Before the algorithm is implemented, the obstacles in the mining area need to be classified according to the area of ​​the obstacles, and the unit distance operation area threshold A of the mining vehicle should be combined. mine , classify the obstacles into Class I obstacles (large scale, S obs ≥A mine ) and Class II obstacles (small scale, S obs mine );

[0032] To construct a mining area environment representation model, an outer contour is used to describe the large-scale mining area boundary, and several inner contours are used to describe the geometric boundaries of obstacles, thereby determining the traversable free space between the inner and outer contours; each contour is stored in the form of a matrix M∈R^(n×2) (n is the number of vertices on the contour), and each row element corresponds to an inflection point coordinate, forming a complete closed polygon boundary; Figure 4 As shown, the type I obstacle (orange area in the figure) is represented by a two-dimensional matrix M = [x1, y1; x2, y1; x1, y2; x2, y2], where each M i Store the inflection point sequence of the corresponding obstacle; the type II obstacle (black dots in the figure) is simplified to the center coordinate set R = [x i ,y​i ], leaving it to the obstacle avoidance process in the path planning stage;

[0033] The Boustrophedon decomposition algorithm based on obstacle features combines the ideas of traditional Boustrophedon decomposition and obstacle feature extraction. Its implementation process is as follows: (1) Contour inflection point extraction and sorting: Inflection point detection is performed on the outer contour and all type I obstacle contours to obtain the inflection point coordinate set {x i} and arranged in ascending order along the x-axis to prepare for the subsequent sub-region division along the x-axis projection; (2) Vertical line segmentation and sub-region generation: at adjacent inflection points x i ,x i+1 A vertical dividing line is constructed along the x-axis direction, and the boundaries of the outer contour and its inner contour are intersected to generate the geometric boundaries of multiple independent partitions; when traversing the partitions, if the line segment part falls on [x i ,x i+1 ] interval, it is necessary to further calculate the precise intersection coordinates of the line segment and the longitudinal boundary, determine the topological relationship of the partition, and form a trapezoidal or polygonal sub-area;

[0034] (3) Partition merging and redundant boundary removal: During the segmentation process, adjacent partitions often share overlapping boundaries (e.g. Figure 5 -Partitions 2 and 4 in (a), in order to reduce the number of partitions and optimize the geometry, adjacent partitions with common boundaries can be merged. Specifically, if the boundary lines between adjacent partitions are geometrically the same straight line or partially overlap, only the two end points are retained and redundant vertices are deleted by updating the partition boundary coordinate matrix (see Figure 5 -(b)), simplifying the partition structure; (4) Handling of small obstacles: small obstacles classified as Class II are not explicitly removed during the decomposition process, but are bypassed by the obstacle avoidance algorithm in the subsequent path planning stage to avoid increasing the computational cost;

[0035] The present invention innovatively proposes a Boustrophedon algorithm based on obstacle features to decompose large-scale mining areas. The differentiated processing strategy has three advantages: first, only subdividing the large-area contour into sub-regions effectively controls the computational complexity and reduces the amount of projection line calculation and boundary intersection operations; second, the strategy based on inflection point segmentation and subsequent merging reduces the problem of over-subdivision of regions and makes the partition topology relationship clearer; finally, the real-time obstacle avoidance processing mechanism adopted for small-area obstacles significantly enhances the adaptability of the algorithm in dynamic environments or scenarios with scattered obstacle distribution; Figure 4 and Figure 5 As shown in the figure, after the vertical projection initial partition and the adjacent partition merge optimization, the final generated sub-region structure can provide an efficient initial division basis for the mining vehicle path coverage calculation under the condition of no obstacle interference.

[0036] Step 103: store the adjacent relationships of all sub-regions in the form of an adjacency matrix, construct a topology matrix between all sub-regions based on the adjacency matrix, and determine the optimal access order of traversing all sub-regions based on the topology matrix between all sub-regions.

[0037] Specifically, the specific steps of the connection path between sub-areas are as follows: (1) Constructing the topological matrix: After the decomposition algorithm processes the explicit partitioning of large-area obstacles and the dynamic neglect of small-area obstacles, it generates a series of sub-areas without interference from Class I obstacles. Although these sub-areas are geometrically independent, they have a clear topological relationship in the overall mining area coordinate system. For details, please refer to Figure 6 To systematically describe the topological relationships between all subregions, the adjacency relationships of all subregions are stored in the form of an adjacency matrix, denoted as A∈R^(n×n), where n is the number of subregions. If the left and right boundaries of subregions i and j partially overlap, then A(i,j)=1; otherwise, A(i,j)=0. In addition, the "adjacency" between factor regions i and j is symmetric, so A(i,j)=A(j,(i), thus obtaining an n×n symmetric matrix as follows:

[0038]

[0039] (2) Sub-region coverage order optimization: The optimal access order problem for traversing all sub-regions can be regarded as a variant of the traditional traveling salesman problem (TSP). The difference is that it no longer requires returning to the starting node after traversing all nodes, so it is transformed into an open traveling salesman problem (Open-TSP). The present invention uses the ant colony algorithm to solve the Open-TSP. Its pheromone positive feedback mechanism and flexible heuristic design can maintain high efficiency and adaptability under a large number of sub-regions.

[0040] In the parameter initialization stage, set the initial pheromone value τ ij (0) = τ0>0, and define the heuristic factor η ij =1 / d ij , where d ij Represents the conversion cost between sub-regions, and evenly distributes ants at different starting points to increase spatial exploration; at the same time, introduces adaptive parameters: α(t), β(t), ρ(t), which are dynamically adjusted during the iteration process, and sets the maximum number of iterations T max When constructing candidate solutions, set the ant probability transfer rule, that is, the probability of the kth ant transferring from the current node i to the node j

[0041]

[0042] in, represents the feasible neighborhood set of the current ant k; through this rule, the ant can comprehensively consider the pheromone concentration and heuristic information, thereby dynamically adjusting the path selection; after each ant completes the path construction, the total cost of its path is calculated to evaluate the objective function. Assume that the path of the kth ant is The total cost of the path L(S k ) is expressed as:

[0043] In the pheromone update phase, a global update strategy is used to update the pheromone concentration τ of each edge (i, j). ij (t) is adjusted and the update formula is as follows: in, Indicates that the kth ant is on path S k The pheromone increment of the edge (i, j) is calculated by the following formula: The parameter Q is a constant used to adjust the release of pheromone, and m is the total number of ants. In order to further improve the convergence performance of the algorithm, the elite strategy is introduced to the global optimal solution S best For pheromone enhancement: Among them, e is the number of elite ants, δ ij (S best ) is the indicator function, when edge (i,j)∈S best 1 when , otherwise 0, set the pheromone boundary [τ min , τ max ], to prevent premature convergence;

[0044] When the number of sub-mining areas is 10, the algorithm finds the optimal access order of the sub-mining areas (for example, 4→6→2→1→3→5→7→9→10→8) under the premise of meeting the global optimal cost. Figure 7The red line in the middle and the overall search quality is guaranteed by the repeated iterative update of pheromones and the positive feedback mechanism; this coverage order allows the mining vehicle to traverse all sub-mining areas without returning to the starting area (Open-TSP feature), effectively reducing unnecessary round-trip paths, thereby reducing operation energy consumption and time costs; there are also connection path planning and dynamic obstacle avoidance problems between subsequent sub-areas (that is, how to connect the path from sub-area 4 to sub-area 6), which can be combined with the adaptive two-level path planning strategy proposed by the team later. This adaptive two-level path planning strategy has been made public. For details, please refer to [2] Changyu Lu, Jianmin Yang, et al. Adaptive bi-level path optimization for deep-sea mining vehicles in non-uniform grids considering ocean currents and dynamic obstacles [J] Ocean Engineering 2025, 315: 119835. (SCI), and combine the sub-area access order obtained by the present invention with the global path planning and local obstacle avoidance of the mining vehicle to meet the comprehensive requirements of deep-sea mining operations in terms of coverage and traffic safety.

[0045] Step 104 : determining a full coverage path of the deep-sea mining vehicle in each sub-area based on the optimal access sequence of all sub-areas and the center coordinate position of the Class II obstacle in each sub-area.

[0046] Specifically, 3. The specific steps of traversing the path in the sub-area are as follows: Sub-area traversal path planning aims to solve two core problems: (1) achieving complete coverage in each sub-area; (2) achieving rapid local replanning based on obstacles in the area; The present invention proposes a finite state driven dual-state collaborative path planning algorithm (FSM-DSCPP) algorithm to achieve dynamic switching between coverage search and inter-area transfer through a finite state machine, and optimize the traversal path by combining a multimodal heuristic function and a steering motion model;

[0047] Mathematical model of FSM-DSCPP algorithm: FSM-DSCPP algorithm uses finite state machine (FSM) to complete the management and collaborative planning of mining vehicle operation mode; its core idea is to integrate coverage search and inter-region transfer by constructing multiple working states and combining heuristic search and path optimization strategy to achieve flexibility of path planning and completeness of coverage in complex mining environment; define state set S = {S standby , S coverage , S unvisited , S avooidance , S error}, state transition function T: S×C→S, where C is the trigger condition set; the state machine structure is as follows: S standby : Standby state, waiting for external instructions or environment updates; S coverage :Using heuristic coverage search strategy, through grid partitioning method and dynamic function optimization, to maximize coverage and minimize path redundancy; S unvisited :When a local dead zone (ón′∈Neighbor(n current ), O n′ =0∧H n′ =0), the system is based on the condition Transfer to S unvisited State, call the A* algorithm to plan the optimal transfer path based on the evaluation function f(n)=g(n)+h(n); S avoidance : Obstacle avoidance state, when obstacle detected = Triggered when true; S error : Error handling state, activated when the system is abnormal; S coverage and S unvisited The seamless switching between the two states is achieved through the path smoothing function Γ(P coverage ,P transfer ) is implemented, which uses cubic spline interpolation and posture optimization technology to ensure the spatial and kinematic continuity of the path and eliminate redundant backtracking; the state transition is based on the environment map M and the strategy function π(s t ,M t ) to ensure that the system can adaptively adjust the working mode according to real-time environmental changes; when planning the traversal path of the deep-sea mining vehicle in each sub-area, first run S coverage Working state, that is, using heuristic coverage search strategy, through grid partitioning method and dynamic function optimization, maximize coverage and minimize path redundancy; when running S coverage In working state, if S is triggered unvisited Working status and S avoidance When in working state, the corresponding function in the obstacle avoidance strategy is called to plan the traversal path in each sub-area until a full coverage path of the deep-sea mining vehicle in each sub-area is obtained;

[0048] Obstacle avoidance strategy: To ensure the safe and efficient movement of mining vehicles in complex environments, the present invention proposes a dual grid verification mechanism, which uses the static obstacle matrix O i,j and the dynamic historical path matrix H i,j Strict constraints on the expansion of grid nodes; the initial environment map is represented by the grid matrix O∈{0,1} m×n , where O i,j =1 indicates that there is a static obstacle in grid (i, j), and the historical path matrix H∈{0,1}m×n , where H i,j =1 indicates that the grid (i, j) has been visited; in the path extension process, only when the grid (i, j) satisfies O i,j =0∧H i,j = 0 condition, thus avoiding static obstacles and preventing path duplication; in addition, the probabilistic grid map P is introduced. i,j ∈[0,1] quantifies the impact of sensor detection error and environmental uncertainty on the grid occupancy state, and only when the occupancy probability is below the threshold (P i,j <P th ) nodes can be regarded as safe nodes;

[0049] In order to realize the real-time avoidance of mining vehicles during the process, a three-level obstacle avoidance response system is designed. By defining a hierarchical function set F local 、F region and F global To achieve adaptive obstacle avoidance, the mathematical definitions of the functions at each level are as follows:

[0050] Local Adjustment F local (n, O): defined as when there is an unvisited and barrier-free node in the neighborhood of the current node (i.e. O n′ =0), the local adjustment is successful; otherwise, regional replanning is triggered.

[0051] Regional Replanning regopn (n start ,n goal ,O,H): In the area with radius R and the current node as the center, the A* algorithm is used to search for a new path. The path cost threshold within the search range is set to 1.5 times the current path length. If no valid path can be found, the global fallback function is triggered. regopn (n start ,n goal ,O,H)=A * (n start ,n goal ,O,H)|n∈Region(n start ,R),Cost(n)≤1.5·Cost current (8);

[0052] Global FallbackF global (path,n current ): Reverse search in the historical path, with the backtracking depth not exceeding 30% of the historical path length, to perform fallback replanning under the condition that there are unvisited and barrier-free nodes in the neighborhood.

[0053] The system execution process adopts a hierarchical progressive strategy: local adjustments are prioritized to quickly avoid obstacles. If the planning goals cannot be met, regional replanning is triggered. When regional replanning fails, a global fallback mechanism is activated to trace back to a feasible node in the historical path and replan. This three-layer structure realizes the organic integration of local avoidance, regional detour and global recovery, effectively improving the fault tolerance and robustness of the system in a dynamic environment. The full coverage path generated by the FSM-DSCPP algorithm in the sub-area is as follows: Figure 8 (b) As shown in the visualization results, the mining vehicle successfully avoided all obstacles and achieved complete coverage of all accessible grids in the sub-area. During the path planning process, when encountering a local dead zone, the algorithm quickly switched to the A* algorithm for regional transfer and planned the optimal path to the next uncovered area, as shown in Figure 2. Figure 8 The orange path line in (b) reflects its dynamic response and adaptability; the entire traversal process takes only 1.2 seconds, the path length is 625, the turn cost is 135, and the path repetition rate is 2.24%; the above indicators demonstrate the superiority of the algorithm in terms of efficiency, path optimization, and coverage capability.

[0054] In summary, the present invention proposes a systematic path planning method for full coverage operations in large-scale mining areas. Through a three-stage collaborative strategy of regional decomposition, connection path planning, and regional traversal planning, it effectively solves the problems of rough partitioning, path redundancy, and inflexible obstacle avoidance in current path planning. The technical route is as follows: Figure 2 As shown; the three core technical innovations and advantages of this method are as follows:

[0055] 1. Adaptive regional decomposition method based on obstacle characteristics: First, to address the problem of numerous and complex obstacles in large-scale mining areas, this paper proposes a Boustrophedon regional decomposition algorithm based on the geometric characteristics of obstacles. By differentially processing Class I (large-scale) and Class II (small-scale) obstacles, it achieves structured modeling and efficient sub-region division of the operating area in complex environments.

[0056] 2. Sub-region connection path optimization strategy based on the Open-TSP model: After completing regional decomposition, the present invention innovatively transforms the connection path planning problem between multiple sub-regions into an Open Traveling Salesman Problem (Open-TSP). Sub-region visit sequence optimization and path calculation are performed based on an improved ant colony optimization algorithm. This method uses an adjacency matrix to construct the topological relationship of sub-regions and designs a pheromone feedback regulation mechanism, a heuristic factor control strategy, and an elite reinforcement strategy, effectively improving the algorithm's global search capability and local optimization accuracy.

[0057] 3. Finite-state driven sub-area traversal path collaborative planning algorithm (FSM-DSCPP): For the operation path planning within each sub-area, the present invention further proposes a finite-state driven dual-state collaborative path planning algorithm (FSM-DSCPP) that integrates behavior management and path search. The algorithm constructs multiple working states such as coverage search, area transfer, and obstacle avoidance, and uses a finite state machine to flexibly schedule the behavior switching of the mining vehicle to achieve efficient path construction in a dynamic environment. The algorithm embeds multimodal heuristic functions and path smoothing interpolation functions to improve the spatial continuity and steering stability of the path. For the obstacle avoidance problem, a three-layer obstacle avoidance response mechanism including local adjustment, area replanning, and global fallback is designed to ensure path safety while reducing the risk of operation interruption caused by local dead zones or sudden obstacles. In addition, by introducing a probabilistic grid map model, the perception error and environmental uncertainty problems are effectively handled, further enhancing the usability of the system in actual scenarios.

[0058] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the present invention. Any equivalent modifications or variations made by persons skilled in the art without departing from the spirit and technical concepts disclosed herein shall be encompassed by the claims of the present invention.

Claims

1. A systematic path planning method for full coverage operations in complex mining areas, characterized by: The following steps are involved: Obtain a raster map of a large-scale mining area containing irregular obstacles; Performing a regional decomposition process on the grid map, and obtaining a plurality of sub-regions free of interference from Class I obstacles and the center coordinate position of the Class II obstacle in each sub-region according to the regional decomposition process result; The adjacent relationships of all sub-regions are stored in the form of an adjacency matrix, a topological matrix between all sub-regions is constructed according to the adjacency matrix, and an optimal access order for traversing all sub-regions is determined according to the topological matrix between all sub-regions; The full coverage path of the deep-sea mining vehicle in each sub-area is determined based on the optimal access sequence of all sub-areas and the center coordinate position of the Class II obstacles in each sub-area.

2. A systematic path planning method for full coverage operations in complex mining areas according to claim 1, characterized in that: The grid image is subjected to a region decomposition process, and a plurality of sub-regions free of interference from type I obstacles and the center coordinate position of the type II obstacle in each sub-region are obtained according to the region decomposition process result, including: Preprocessing the grid image, obtaining the geometric outline of the obstacle according to the preprocessing result, and storing it in a matrix form; According to the geometric outline of the obstacle and using the Boustrophedon algorithm based on obstacle characteristics, several sub-areas without interference from type I obstacles and the center coordinate position of the type II obstacle in each sub-area are obtained.

3. The method for systematic path planning for full coverage operations in complex mining areas according to claim 2, characterized in that: After preprocessing the grid image, and before obtaining a plurality of sub-areas free of Type I obstacle interference and the center coordinate position of a Type II obstacle in each sub-area using a Boustrophedon algorithm based on obstacle features according to the geometric outline of the obstacle, the method further includes: The obstacles in large-scale mining areas are classified according to their area and the unit distance operating area threshold of mining vehicles. Class I obstacles and Class II obstacles are obtained based on the classification results.

4. The method for systematic path planning for full coverage operations in complex mining areas according to claim 2, characterized in that: The preprocessing of the grid image, obtaining the geometric outline of the obstacle according to the preprocessing result, and storing it in a matrix form includes: The irregular obstacle area that only partially occupies the grid cells is filled, and the cell dilation operation is performed on the filled obstacle area. Then, the geometric outline of the obstacle is extracted from the dilated edge and stored in the form of a matrix.

5. The method for systematic path planning for full coverage operations in complex mining areas according to claim 3 is characterized by: The method of obtaining a plurality of sub-areas free of Class I obstacle interference and the center coordinate position of the Class II obstacle in each sub-area using the Boustrophedon algorithm based on obstacle characteristics according to the geometric outline of the obstacle includes: Determine the inflection point coordinates from the geometric outline of the obstacle stored in the form of a matrix, and divide the large-scale mining area into several initial sub-areas without interference from type I obstacles according to the inflection point coordinates of type I obstacles; The adjacent partitions of several initial sub-areas without type I obstacle interference are merged, and several sub-areas without type I obstacle interference are obtained according to the merging results.

6. The method for systematic path planning for full coverage operations in complex mining areas according to claim 1, characterized in that: The determining of the optimal access order of traversing all sub-areas according to the topological matrix between all sub-areas includes: An optimal access sequence for traversing all sub-regions is determined based on a topological matrix between all sub-regions and using an ant colony algorithm. The problem of the optimal access sequence for traversing all sub-regions can be regarded as an open traveling salesman problem, and the ant colony algorithm is used to solve the open traveling salesman problem.

7. The method for systematic path planning for full coverage operations in complex mining areas according to claim 1, characterized in that: The method of determining a full coverage path of the deep-sea mining vehicle in each sub-area based on the optimal access sequence of all sub-areas and the center coordinate position of the Class II obstacles in each sub-area includes: The method uses a finite state machine to dynamically switch between coverage search and inter-area transfer based on the optimal access sequence of all sub-areas and the center coordinate position of the Class II obstacles in each sub-area. The traversal path is optimized by combining a multimodal heuristic function and a steering motion model, thereby obtaining a full coverage path for the deep-sea mining vehicle in each sub-area.

8. The method for systematic path planning for full coverage operations in complex mining areas according to claim 7, characterized in that: The finite state machine is used to dynamically switch between coverage search and inter-area transfer, and the multimodal heuristic function and steering motion model are combined to optimize the traversal path, thereby obtaining a full coverage path for the deep-sea mining vehicle in each sub-area, including: A finite state machine is used to construct multiple working states. Based on the constructed multiple working states, the traversal path of the deep-sea mining vehicle in each sub-area is planned. During the path planning process, when a working state related to the obstacle avoidance strategy is triggered, the corresponding function in the obstacle avoidance strategy is called to plan the traversal path in each sub-area until a full coverage path is obtained for the deep-sea mining vehicle in each sub-area.

9. The method for systematic path planning for full coverage operations in complex mining areas according to claim 8, characterized in that: The obstacle avoidance strategy is a three-level obstacle avoidance response system, which includes a local adjustment function, a regional replanning function, and a global fallback function in sequence; The local adjustment function is that when there is an unvisited and unobstructed grid node in the neighborhood of the current grid node, the local adjustment is successful, otherwise the regional replanning function is triggered; The regional replanning function uses the A* algorithm to search for a new path within an area with a radius R centered on the current node, with the path cost threshold within the search range set to u times the current path length; if no valid path can be found, the global fallback function is triggered; The global fallback function searches backward in the historical path, and the backtracking depth does not exceed v% of the historical path length, so as to perform fallback replanning under the condition that there are unvisited and barrier-free grid nodes in the neighborhood.

Images