Walking path planning method suitable for cabin unmanned excavator

Abstract

The invention discloses a walking path planning method suitable for a cabin unmanned excavator, and the method comprises the steps: employing an inertial navigation module to combine with a laser radar to scan cabin environment information, dividing the cabin environment information into a plurality of grid units, constructing a cabin environment space coordinate system, and forming a three-dimensional environment map; the grid units in the space coordinate system serve as basic units, the obstacle position points are assigned, and planning area path data are obtained; taking an original point as an initial node of the planning area path data, and obtaining path data from the initial node to a target node; smoothing path information of the path data by adopting a path planning algorithm, and planning a path with the minimum elevation change as a walking path; self-adaptive dynamic adjustment is conducted on the walking path of the unmanned excavator, an evaluation function is established, the optimal path track is obtained, the walking dynamic state of the unmanned excavator can be mastered in real time, it is ensured that the unmanned excavator is always kept stable in the walking process, and overturning and jamming are avoided.

Description

Technical Field

[0001] The invention relates to the technical field of path planning, and in particular to a walking path planning method suitable for an unmanned excavator in a cabin. Background Art

[0002] With the rapid development of modern ship transportation, cargo loading and unloading operations in the cabin have become more frequent and complicated. The traditional manual operation method is not only inefficient, but also has high safety risks. The introduction of unmanned excavators provides an efficient and safe solution for cargo loading and unloading in ships. However, due to the complex environment and diverse terrain in the cabin, unmanned excavators face many challenges in walking path planning. Therefore, it is of great practical significance to develop a walking path planning method suitable for unmanned excavators in cabins.

[0003] At present, there are some path planning technologies for unmanned excavators on the market, but most of them are used in open sites or mining environments. The existing path planning methods for cabin unmanned excavators mainly have the following problems:

[0004] (1) In open areas, GPS positioning technology can provide relatively accurate location information. However, in the cabin, due to severe signal obstruction, the GPS positioning accuracy is greatly reduced and cannot meet the positioning requirements of the unmanned excavator.

[0005] (2) Existing path planning algorithms are mostly based on plane path planning, which does not fully consider the impact of elevation changes on the movement of the excavator. As a result, the planned path is not stable enough, and the excavator is prone to overturning or jamming during movement. Summary of the invention

[0006] The purpose of the present invention is to provide a walking path planning method suitable for an unmanned excavator in a cabin, so as to solve the technical problem in the prior art that the unmanned excavator is prone to overturning or jamming during walking, thereby affecting the working efficiency and safety.

[0007] In order to solve the above technical problems, the present invention specifically provides the following technical solutions:

[0008] The present invention provides a walking path planning method applicable to a cabin unmanned excavator, comprising the following steps:

[0009] An inertial navigation module is used in combination with a laser radar to scan the cabin environment information, and the cabin environment information is divided into a plurality of grid units. The position point of the unmanned excavator is used as the origin, and the length of the grid unit is used as the unit length to construct a cabin environment space coordinate system to form a three-dimensional environment map;

[0010] Taking the grid unit in the spatial coordinate system as the basic unit, assigning values ​​to the obstacle position points, and obtaining the path data of the planned area;

[0011] The planned area path data takes the origin as the initial node, calculates the elevation data from the initial node to the target node, and obtains the path data from the initial node to the target node;

[0012] The path planning algorithm is used to smooth the path information of the path data, and a path with the smallest elevation change is planned as the walking path of the unmanned excavator;

[0013] The walking path of the unmanned excavator is adaptively and dynamically adjusted, an evaluation function is established, and an optimal path trajectory is obtained.

[0014] As a preferred solution of the present invention, an inertial navigation module is used in combination with a laser radar to scan the cabin environment information, and the cabin environment information is divided into a plurality of grid units, including:

[0015] An inertial navigation system is installed inside the cabin to provide navigation information, and a laser radar fusion sensor is used to scan the cabin environment information in real time;

[0016] The cabin environment information is rasterized one by one from left to right in a flat-top regular hexagonal pattern, and six vertices of the regular hexagon are connected with circles to form a horizontal grid;

[0017] Each of the regular hexagons is numbered by row, and double numbered according to the odd and even row numbers, so as to obtain a plurality of grid units in sequence.

[0018] As a preferred solution of the present invention, taking the position point of the unmanned excavator as the origin and the grid unit length as the unit length, a cabin environment space coordinate system is constructed to form a three-dimensional environment map, including:

[0019] Taking the center of the regular hexagonal grid occupied by the position of the unmanned excavator as the origin and the length of the diagonal of the regular hexagon as the length of the grid unit, a spatial rectangular coordinate system is established;

[0020] The regular hexagonal grid is extended in the positive direction of the spatial rectangular coordinate system by adopting a combination form of right shifting of even-numbered rows;

[0021] The regular hexagonal grid at the obstacle point position is marked in the spatial coordinate system. If the obstacle area in the regular hexagonal grid is less than one tenth of the grid area, the regular hexagonal grid is not marked.

[0022] As a preferred solution of the present invention, the grid unit in the spatial coordinate system is used as a basic unit, and the obstacle position point is assigned a value to obtain the planning area path data, including:

[0023] The matrix composed of regular hexagonal grids is converted into an adjacency matrix in the spatial coordinate system, the location of the unmanned excavator is taken as the starting origin, and the path from the starting origin to the target point is planned in the adjacency matrix using an ant colony algorithm;

[0024] The obstacle positions are assigned within the path from the starting point to the target point, and a safety distance is set around the starting point, the arrived path is recorded, and the path data of the planned area is obtained.

[0025] As a preferred solution of the present invention, the planning area path data takes the origin as the initial node, calculates the elevation data from the initial node to the target node, and obtains the path data from the initial node to the target node, including:

[0026] Divide the path set S according to the planned area path data and calculate the path length matrix C i , for each unmanned excavator j, insert an empty path P into the path set S j , select the target node v 0 , and insert the target node into the corresponding empty path P j middle;

[0027] The adjacency matrix is ​​used to continuously search for the next nearest point that meets the time window limit and the capacity limit, and inserts it into the empty path P. j Until no point that meets the time window limit or the empty path P is found j The maximum capacity reaches W;

[0028] In the empty path P j Insert the target node v at the end of 0 , get a path S in the path set S j , for the path S j The path data in the path set S is obtained by calculating the elevation data.

[0029] As a preferred solution of the present invention, evaluating the path data to obtain evaluation point path data includes:

[0030] Create Openlist and Closedlist for the path data to store evaluated path nodes and nodes to be evaluated, set cost values ​​for the path nodes in the Openlist to find the shortest path, and complete the path planning until the end point is reached;

[0031] A cost function h(n,m) is established for the shortest path, and a dynamic weighting process is performed on the cost function h(n,m), and its expression is:

[0032] h(n,m)=h(n)*exp[k(1-m)*h(n)]

[0033] Among them, h(n) represents the cost function from the current evaluation node to the end point, which is obtained by calculating the Euclidean distance from the current evaluation node to the end point, m represents the ratio of the obstacle area under the current regular hexagonal grid to the entire map area, and k represents the dynamic weighting constant;

[0034] The minimum value of the cost function is updated to the Openlist, and the evaluated path nodes in the Openlist are horizontally expanded and diagonally expanded to find jump points;

[0035] The cost value of the jump point is calculated and added to the Openlist to obtain the evaluation point path data.

[0036] As a preferred solution of the present invention, a path planning algorithm is used to perform path smoothing processing on the evaluation point path data, and a path with the smallest elevation change is planned as the walking path of the unmanned excavator, including:

[0037] The evaluation point path data is extracted to form a path point set P, and smoothing is performed on the path turning points in the path point set P, and the expression is:

[0038] B(t)=(1-t) 3 P 0 +3t(1-t) 2 P 1 +3t 2 (1-t)P 2 +t 3 P 3

[0039] Among them, t represents the time parameter, which ranges from [0,1], P 0 , P 1 , P 2 , P 3 Represents any path turning point in the path point set P;

[0040] By controlling P 0 , P 1 , P 2 , P 3 The position of the control path is fitted at the inflection point;

[0041] A safety radius is set at the turning point of the path, and the elevation data change value of the turning point of the path is repeatedly calculated, and the path with the smallest elevation data change is used as the walking path of the drone excavator.

[0042] As a preferred solution of the present invention, the walking path of the unmanned excavator is adaptively and dynamically adjusted, including:

[0043] The shortest distance between the unmanned excavator and the obstacle in the walking path of the unmanned excavator is recorded as D min , through the shortest distance D min Set the speed evaluation function v r , whose expression is:

[0044]

[0045] Among them, v 0 represents the starting speed weight, θ represents the speed parameter, D s represents the safety threshold, n represents the weight coefficient of the safety area, v max represents the maximum speed of the unmanned excavator, and v represents the linear acceleration of the unmanned excavator;

[0046] Evaluate the function v according to the speed r The walking path of the unmanned excavator is adjusted in real time in combination with the safety distance to obtain dynamic walking path data.

[0047] As a preferred solution of the present invention, an evaluation function is established for the dynamic walking path data to obtain an optimal path trajectory, including:

[0048] The dynamic walking path data is input into the path evaluation function hed(v,w), the global path nodes are collected, and the local optimal path is planned. The path evaluation function hed(v,w) is expressed as:

[0049] hed(v,w)=v r *vel(v,w)+δ*dist_s(v,w)+w*dist_r(v,w)

[0050] Among them, δ and w represent the weight coefficients of the evaluation function, vel(v,w) represents the speed evaluation function, dist_s(v,w) represents the shortest distance evaluation function between the predicted trajectory endpoint and the environmental obstacle, and dist_r(v,w) represents the trajectory walking path azimuth evaluation function;

[0051] The local path is optimized according to the path evaluation function, and the global optimal path is obtained through an adaptive fusion algorithm.

[0052] As a preferred solution of the present invention, the adaptive fusion algorithm selects the optimal path, including:

[0053] The optimal local paths selected by the path evaluation function are merged into a global path set, key path nodes are extracted, trajectory prediction is performed on the key path nodes, and the optimal path is screened;

[0054] The optimal path is converted into an evaluation function based on linear velocity and angular velocity according to the walking speed of the unmanned excavator, the positions of obstacles around the trajectory are output, and the global optimal path is planned according to the safety distance.

[0055] Compared with the prior art, the present invention has the following beneficial effects:

[0056] The present invention adopts a high-precision inertial navigation module combined with a laser radar to obtain the position information of the unmanned excavator in the cabin in real time. Through multi-sensor data fusion, the limitations of a single sensor are overcome to ensure positioning accuracy and stability, and a high-precision three-dimensional environmental map of the cabin is generated. By calculating the elevation data in the map, the terrain changes in the cabin are identified. The adaptive fusion algorithm is used to compensate for the lack of local path planning capabilities in a dynamic environment, and the problem that the dynamic window does not have a good global guidance is solved, so that the unmanned excavator can plan a smooth path with global optimality while avoiding unknown obstacles and moving obstacles. At the same time, by adaptively and dynamically adjusting the walking path of the unmanned excavator, the walking dynamics of the unmanned excavator can be grasped in real time. By dynamically adjusting the path planning, it is ensured that the unmanned excavator always remains stable during the walking process to avoid overturning and jamming. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for the embodiments or the description of the prior art are briefly introduced below. Obviously, the drawings in the following description are only exemplary, and a person of ordinary skill in the art can also derive other implementation drawings based on the provided drawings without creative work.

[0058] Figure 1 A flow chart of a method for planning a travel path for an unmanned excavator provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] like Figure 1 As shown, the present invention provides a walking path planning method applicable to a cabin unmanned excavator, comprising the following steps:

[0061] An inertial navigation module is used in combination with a laser radar to scan the cabin environment information, and the cabin environment information is divided into a plurality of grid units. The position point of the unmanned excavator is used as the origin, and the length of the grid unit is used as the unit length to construct a cabin environment space coordinate system to form a three-dimensional environment map;

[0062] In this embodiment, a high-precision inertial navigation module is used in combination with a laser radar to obtain real-time position information of the unmanned excavator in the cabin. Through multi-sensor data fusion, the limitations of a single sensor are overcome, positioning accuracy and stability are ensured, and a high-precision three-dimensional environmental map of the cabin is generated.

[0063] In this embodiment, the actual three-dimensional environment map is abstracted into a digital map, the cabin environment information is rasterized, and the environment is divided into a number of cells of the same size. The obstacle information in the map is represented by 0 and 1 in the grid, 0 represents a passable area, and 1 represents an impassable area. The division of the grid size will directly determine the accuracy of the environment information and the efficiency of path finding.

[0064] Taking the grid unit in the spatial coordinate system as the basic unit, assigning values ​​to the obstacle position points, and obtaining the path data of the planned area;

[0065] In this embodiment, by assigning values ​​to obstacles in grid cells, setting a safe travel distance for the unmanned excavator, and planning regional path data within the safe distance, the unmanned excavator can be effectively prevented from being restricted in the walking path due to its own volume.

[0066] The planned area path data takes the origin as the initial node, calculates the elevation data from the initial node to the target node, and obtains the path data from the initial node to the target node;

[0067] In this embodiment, by calculating the elevation data from the initial node to the target node, the terrain changes in the cabin are identified, thereby improving the accuracy of path planning.

[0068] The path planning algorithm is used to smooth the path information of the path data, and a path with the smallest elevation change is planned as the walking path of the unmanned excavator;

[0069] The walking path of the unmanned excavator is adaptively and dynamically adjusted, an evaluation function is established, and an optimal path trajectory is obtained.

[0070] In this embodiment, the walking path of the unmanned excavator is adaptively adjusted using a dynamic window to obtain local path planning, thereby ensuring that the unmanned excavator can reach the target location quickly, accurately and safely.

[0071] The inertial navigation module is combined with a laser radar to scan the cabin environment information, and the cabin environment information is divided into multiple grid units, including:

[0072] An inertial navigation system is installed inside the cabin to provide navigation information, and a laser radar fusion sensor is used to scan the cabin environment information in real time;

[0073] The cabin environment information is rasterized one by one from left to right in a flat-top regular hexagonal pattern, and six vertices of the regular hexagon are connected with circles to form a horizontal grid;

[0074] Each of the regular hexagons is numbered by row, and double numbered according to the odd and even row numbers, so as to obtain a plurality of grid units in sequence.

[0075] In this embodiment, a high-precision inertial navigation module is combined with a laser radar to obtain the position information of the unmanned excavator in the cabin in real time. The positioning accuracy and stability are ensured through multi-sensor data fusion. The environmental information is rasterized in the form of regular hexagons, the row and column coordinates are doubled using the double coordinate method, and the doubled horizontal distance is used instead of the continuous coordinates, thereby reducing the computational complexity in the grid marking process.

[0076] Taking the position of the unmanned excavator as the origin and the grid unit length as the unit length, a cabin environment space coordinate system is constructed to form a three-dimensional environment map, including:

[0077] Taking the center of the regular hexagonal grid occupied by the position of the unmanned excavator as the origin and the length of the diagonal of the regular hexagon as the length of the grid unit, a spatial rectangular coordinate system is established;

[0078] The regular hexagonal grid is extended in the positive direction of the spatial rectangular coordinate system by adopting a combination form of right shifting of even-numbered rows;

[0079] The regular hexagonal grid at the obstacle point position is marked in the spatial coordinate system. If the obstacle area in the regular hexagonal grid is less than one tenth of the grid area, the regular hexagonal grid is not marked.

[0080] In this embodiment, the regular hexagonal grid is sorted according to the right-offset combination of even rows, and the pointed regular hexagon with a side length of 1 is taken as the basic square to establish a spatial rectangular coordinate system, which can not only realize environmental modeling but also improve the accuracy of node position information.

[0081] The grid unit in the spatial coordinate system is used as a basic unit, and the obstacle position point is assigned a value to obtain the planning area path data, including:

[0082] The matrix composed of regular hexagonal grids is converted into an adjacency matrix in the spatial coordinate system, the location of the unmanned excavator is taken as the starting origin, and the path from the starting origin to the target point is planned in the adjacency matrix using an ant colony algorithm;

[0083] The obstacle positions are assigned within the path from the starting point to the target point, and a safety distance is set around the starting point, the arrived path is recorded, and the path data of the planned area is obtained.

[0084] In this embodiment, the position of the unmanned excavator is used as the starting point in the spatial coordinate system, and the ant colony algorithm is used for path planning. The grid matrix is ​​first converted into an adjacency matrix, and the pheromone and heuristic information of the ant colony algorithm are initialized at the same time. Each ant starts from the starting point until it reaches the end point or stops at a point where there is no next step to advance. The successfully reached path and route are recorded and compared with the current shortest path. After all ants complete the iteration in each round, the pheromone is updated, and the shortest path is output after the maximum number of iterations is reached.

[0085] The planning area path data takes the origin as the initial node, calculates the elevation data from the initial node to the target node, and obtains the path data from the initial node to the target node, including:

[0086] Divide the path set S according to the planned area path data and calculate the path length matrix C i , for each unmanned excavator j, insert an empty path P into the path set S j , select the target node v 0 , and insert the target node into the corresponding empty path P j middle;

[0087] The adjacency matrix is ​​used to continuously search for the next nearest point that meets the time window limit and the capacity limit, and inserts it into the empty path P. j Until no point that meets the time window limit or the empty path P is found j The maximum capacity reaches W;

[0088] In the empty path P j Insert the target node v at the end of 0 , get a path S in the path set S j , for the path S j The path data in the path set S is obtained by calculating the elevation data.

[0089] In this embodiment, a proximity matrix data structure is used to store graph information. Each path node can be represented as a fixed point, and its demand, time window and other information are stored on the vertex. At the same time, the adjacency relationship between it and other vertices is recorded, which can save storage space and improve query efficiency.

[0090] Evaluating the path data to obtain evaluation point path data includes:

[0091] Create Openlist and Closedlist for the path data to store evaluated path nodes and nodes to be evaluated, set cost values ​​for the path nodes in the Openlist to find the shortest path, and complete the path planning until the end point is reached;

[0092] A cost function h(n,m) is established for the shortest path, and a dynamic weighting process is performed on the cost function h(n,m), and its expression is:

[0093] h(n,m)=h(n)*exp[k(1-m)*h(n)]

[0094] Wherein, h(n) represents the cost function from the current evaluation node to the end point, which is obtained by calculating the Euclidean distance from the current evaluation node to the end point, m represents the ratio of the obstacle area under the current regular hexagonal grid to the entire map area, and k represents the dynamic weighting constant;

[0095] The minimum value of the cost function is updated to the Openlist, and the evaluated path nodes in the Openlist are horizontally expanded and diagonally expanded to find jump points;

[0096] The cost value of the jump point is calculated and added to the Openlist to obtain the evaluation point path data.

[0097] In this embodiment, dynamic weighting is used for the shortest path, so that when the current node is far away from the target point, it moves toward the target point faster, and when it is close to the target point, the weight is dynamically modified to reduce the heuristics of the function and increase the search range.

[0098] In this embodiment, when the value of m becomes larger, the value of h(n,m) becomes smaller, the heuristic of the function decreases, the search time increases slightly, and the path length is shorter; when the value of m becomes smaller, the proportion of obstacles in the map decreases, the heuristic of the function increases, and the search efficiency improves.

[0099] The path planning algorithm is used to perform path smoothing processing on the path data of the evaluation points, and a path with the smallest elevation change is planned as the walking path of the unmanned excavator, including:

[0100] The evaluation point path data is extracted to form a path point set P, and smoothing is performed on the path turning points in the path point set P, and the expression is:

[0101] B(t)=(1-t) 3 P 0 +3t(1-t) 2 P 1 +3t 2 (1-t)P 2 +t 3 P 3

[0102] Among them, t represents the time parameter, which ranges from [0,1], P 0 , P 1 , P 2 , P 3 Represents any path turning point in the path point set P;

[0103] By controlling P 0 , P 1 , P 2 , P 3 The position of the control path is fitted at the inflection point;

[0104] A safety radius is set at the turning point of the path, and the elevation data change value of the turning point of the path is repeatedly calculated, and the path with the smallest elevation data change is used as the walking path of the drone excavator.

[0105] In this embodiment, the target point and its parent nodes are first placed in the total node set Q, and the starting point is marked as P. 0 The remaining nodes are marked as P 1 , P 2 , ..., P n , where P n As the target point, the starting point P 0 Connect the path point set P with straight lines in sequence to determine whether the straight line collides with the obstacle. If the straight line collides with the obstacle, the node is a must-pass node of the path and must be a key turning point. Establish a key turning point set O and move the node into it first; if the straight line does not collide with the obstacle, the node is considered redundant, and then P is used to 1 Restart as the next starting point and connect the remaining nodes in Q in sequence until P n , the screening ends. The nodes in set O are used as the intermediate guiding points of the fusion algorithm, and the walking path of the UAV excavator is obtained through curve fitting.

[0106] Adaptively and dynamically adjusting the walking path of the unmanned excavator includes:

[0107] The shortest distance between the unmanned excavator and the obstacle in the walking path of the unmanned excavator is recorded as D min , through the shortest distance D min Set the speed evaluation function v r , whose expression is:

[0108]

[0109] Among them, v 0 represents the starting speed weight, θ represents the speed parameter, D s represents the safety threshold, n represents the weight coefficient of the safety area, vmax represents the maximum speed of the unmanned excavator, and v represents the linear acceleration of the unmanned excavator;

[0110] Evaluate the function v according to the speed r The walking path of the unmanned excavator is adjusted in real time in combination with the safety distance to obtain dynamic walking path data.

[0111] In this embodiment, the speed evaluation function v r It indicates the speed of the mobile robot when it moves along the current simulated trajectory. The faster the unmanned excavator moves, the greater the value of the speed evaluation function, indicating that the mobile robot can approach the target point faster when moving along the current simulated trajectory. If the unmanned excavator moves too fast, although it can move to the target point faster when there are fewer obstacles, improving the efficiency of task completion, but when there are many obstacles or the environment is more complex, the too fast movement speed will increase the risk of collision between the mobile robot and the obstacles, and it is easy to miss the target point when approaching the target point, which will reduce the efficiency of task completion.

[0112] An evaluation function is established for the dynamic walking path data to obtain an optimal path trajectory, including:

[0113] The dynamic walking path data is input into the path evaluation function hed(v,w), the global path nodes are collected, and the local optimal path is planned. The path evaluation function hed(v,w) is expressed as:

[0114] hed(v,w)=v r *vel(v,w)+δ*dist_s(v,w)+w*dist_r(v,w)

[0115] Among them, δ and w represent the weight coefficients of the evaluation function, vel(v,w) represents the speed evaluation function, dist_s(v,w) represents the shortest distance evaluation function between the predicted trajectory endpoint and the environmental obstacle, and dist_r(v,w) represents the trajectory walking path azimuth evaluation function;

[0116] The local path is optimized according to the path evaluation function, and the global optimal path is obtained through an adaptive fusion algorithm.

[0117] In this embodiment, a speed evaluation function v is established according to the traveling speed of the unmanned excavator. r As well as the path evaluation function hed(v,w), the unmanned excavator can effectively avoid falling into the local optimal solution when performing path planning, thereby improving the success rate of the algorithm and the working efficiency of the unmanned excavator.

[0118] The adaptive fusion algorithm selects the optimal path, including:

[0119] The optimal local paths selected by the path evaluation function are merged into a global path set, key path nodes are extracted, trajectory prediction is performed on the key path nodes, and the optimal path is screened;

[0120] The optimal path is converted into an evaluation function based on linear velocity and angular velocity according to the walking speed of the unmanned excavator, the positions of obstacles around the trajectory are output, and the global optimal path is planned according to the safety distance.

[0121] In this embodiment, an adaptive fusion algorithm is used to make up for the lack of local path planning capabilities in a dynamic environment, and also solves the problem that the dynamic window does not have a good global guide, so that the unmanned excavator can plan a smooth path with global optimality while avoiding unknown obstacles and moving obstacles. By adaptively and dynamically adjusting the walking path of the unmanned excavator, the walking dynamics of the unmanned excavator can be grasped in real time. By dynamically adjusting the path planning, it is ensured that the unmanned excavator always remains stable during the walking process to avoid overturning and jamming.

[0122] A high-precision inertial navigation module combined with a laser radar is used to obtain the position information of the unmanned excavator in the cabin in real time. Through multi-sensor data fusion, the limitations of a single sensor are overcome to ensure positioning accuracy and stability, and a high-precision three-dimensional environmental map of the cabin is generated. By calculating the elevation data in the map, the terrain changes in the cabin are identified. The adaptive fusion algorithm is used to make up for the lack of local path planning capabilities in a dynamic environment and solve the problem that the dynamic window does not have a good global orientation. The unmanned excavator can plan a smooth path with global optimality while avoiding unknown obstacles and moving obstacles. At the same time, through adaptive dynamic adjustment of the unmanned excavator's walking path, the walking dynamics of the unmanned excavator can be grasped in real time. By dynamically adjusting the path planning, it is ensured that the unmanned excavator always remains stable during the walking process to avoid overturning and jamming.

[0123] The above embodiments are only exemplary embodiments of the present application and are not intended to limit the present application. The protection scope of the present application is defined by the claims. Those skilled in the art may make various modifications or equivalent substitutions to the present application within the essence and protection scope of the present application, and such modifications or equivalent substitutions shall also be deemed to fall within the protection scope of the present application.

Claims

1. A walking path planning method for an unmanned excavator in a cabin, characterized in that: The following steps are involved: An inertial navigation module is used in combination with a laser radar to scan the cabin environment information, and the cabin environment information is divided into a plurality of grid units. The position point of the unmanned excavator is used as the origin, and the length of the grid unit is used as the unit length to construct a cabin environment space coordinate system to form a three-dimensional environment map; Taking the grid unit in the spatial coordinate system as the basic unit, assigning values ​​to the obstacle position points, and obtaining the planning area path data; The planned area path data takes the origin as the initial node, calculates the elevation data from the initial node to the target node, and obtains the path data from the initial node to the target node; The path planning algorithm is used to smooth the path information of the path data, and a path with the smallest elevation change is planned as the walking path of the unmanned excavator; The walking path of the unmanned excavator is adaptively and dynamically adjusted, an evaluation function is established, and an optimal path trajectory is obtained.

2. A walking path planning method for an unmanned excavator in a cabin according to claim 1, characterized in that: The inertial navigation module is combined with a laser radar to scan the cabin environment information, and the cabin environment information is divided into multiple grid units, including: An inertial navigation system is installed inside the cabin to provide navigation information, and a laser radar fusion sensor is used to scan the cabin environment information in real time; The cabin environment information is rasterized one by one from left to right in a flat-top regular hexagonal pattern, and six vertices of the regular hexagon are connected with circles to form a horizontal grid; Each of the regular hexagons is numbered by row, and double numbered according to the odd and even row numbers, so as to obtain a plurality of grid units in sequence.

3. A walking path planning method for an unmanned excavator in a cabin according to claim 2, characterized in that: Taking the position of the unmanned excavator as the origin and the grid unit length as the unit length, a cabin environment space coordinate system is constructed to form a three-dimensional environment map, including: Taking the center of the regular hexagonal grid occupied by the position of the unmanned excavator as the origin and the length of the diagonal of the regular hexagon as the length of the grid unit, a spatial rectangular coordinate system is established; The regular hexagonal grid is extended in the positive direction of the spatial rectangular coordinate system by adopting a combination form of right shifting of even-numbered rows; The regular hexagonal grid at the obstacle point position is marked in the spatial coordinate system. If the obstacle area in the regular hexagonal grid is less than one tenth of the grid area, the regular hexagonal grid is not marked.

4. A walking path planning method for an unmanned excavator in a cabin according to claim 3, characterized in that: The grid unit in the spatial coordinate system is used as a basic unit, and the obstacle position point is assigned a value to obtain the planning area path data, including: The matrix composed of regular hexagonal grids is converted into an adjacency matrix in the spatial coordinate system, the location of the unmanned excavator is taken as the starting origin, and the path from the starting origin to the target point is planned in the adjacency matrix using an ant colony algorithm; The obstacle positions are assigned within the path from the starting point to the target point, and a safety distance is set around the starting point, the arrived path is recorded, and the path data of the planned area is obtained.

5. The walking path planning method for an unmanned excavator in a cabin according to claim 3 is characterized in that: The planning area path data takes the origin as the initial node, calculates the elevation data from the initial node to the target node, and obtains the path data from the initial node to the target node, including: Divide the path set S according to the planned area path data and calculate the path length matrix C i , for each unmanned excavator j, insert an empty path P into the path set S j , select the target node v0 and insert the target node into the corresponding empty path P j middle; The adjacency matrix is ​​used to continuously search for the next nearest point that meets the time window limit and the capacity limit, and inserts it into the empty path P. j Until no point that meets the time window limit or the empty path P is found j The maximum capacity reaches W; In the empty path P j Insert the target node v0 at the end of , and obtain a path S in the path set S j , for the path S j The path data in the path set S is obtained by calculating the elevation data.

6. A walking path planning method for an unmanned excavator in a cabin according to claim 5, characterized in that: Evaluating the path data to obtain evaluation point path data includes: Create Openlist and Closedlist for the path data to store evaluated path nodes and nodes to be evaluated, set cost values ​​for the path nodes in the Openlist to find the shortest path, and complete the path planning until the end point is reached; A cost function h(n,m) is established for the shortest path, and a dynamic weighting process is performed on the cost function h(n,m), and its expression is: h(n,m)=h(n)*exp[k(1-m)*h(n)] Wherein, h(n) represents the cost function from the current evaluation node to the end point, which is obtained by calculating the Euclidean distance from the current evaluation node to the end point, m represents the ratio of the obstacle area under the current regular hexagonal grid to the entire map area, and k represents the dynamic weighting constant; The minimum value of the cost function is updated to the Openlist, and the evaluated path nodes in the Openlist are horizontally expanded and diagonally expanded to find jump points; The cost value of the jump point is calculated and added to the Openlist to obtain the evaluation point path data.

7. A walking path planning method for a cabin unmanned excavator according to claim 6, characterized in that: The path planning algorithm is used to perform path smoothing processing on the path data of the evaluation points, and a path with the smallest elevation change is planned as the walking path of the unmanned excavator, including: The evaluation point path data is extracted to form a path point set P, and smoothing is performed on the path turning points in the path point set P, and the expression is: B(t)=(1-t) 3 P0+3t(1-t) 2 P1+3t 2 (1-t)P2+t 3 P3 Where t represents the time parameter, which ranges from [0,1], and P0, P1, P2, and P3 represent any path turning points in the path point set P; By controlling the positions of P0, P1, P2, and P3, the control path is fitted at the inflection point; A safety radius is set at the turning point of the path, and the elevation data change value of the turning point of the path is repeatedly calculated, and the path with the smallest elevation data change is used as the walking path of the drone excavator.

8. The walking path planning method for an unmanned excavator in a cabin according to claim 7 is characterized in that: Adaptively and dynamically adjusting the walking path of the unmanned excavator includes: The shortest distance between the unmanned excavator and the obstacle in the walking path of the unmanned excavator is recorded as D min , through the shortest distance D min Set the speed evaluation function v r , whose expression is: Among them, v0 represents the starting speed weight, θ represents the speed parameter, and D s represents the safety threshold, n represents the weight coefficient of the safety area, v max represents the maximum speed of the unmanned excavator, and v represents the linear acceleration of the unmanned excavator; Evaluate the function v according to the speed r The walking path of the unmanned excavator is adjusted in real time in combination with the safety distance to obtain dynamic walking path data.

9. A walking path planning method for a cabin unmanned excavator according to claim 8, characterized in that: An evaluation function is established for the dynamic walking path data to obtain an optimal path trajectory, including: The dynamic walking path data is input into the path evaluation function hed(v,w), the global path nodes are collected, and the local optimal path is planned. The path evaluation function hed(v,w) is expressed as: hed(v,w)=v r *vel(v,w)+δ*dist_s(v,w)+w*dist_r(v,w) Among them, δ and w represent the weight coefficients of the evaluation function, vel(v,w) represents the speed evaluation function, dist_s(v,w) represents the shortest distance evaluation function between the predicted trajectory endpoint and the environmental obstacle, and dist_r(v,w) represents the trajectory walking path azimuth evaluation function; The local path is optimized according to the path evaluation function, and the global optimal path is obtained through an adaptive fusion algorithm.

10. A walking path planning method applicable to a cabin unmanned excavator according to claim 9, characterized in that: The adaptive fusion algorithm selects the optimal path, including: The optimal local paths selected by the path evaluation function are merged into a global path set, key path nodes are extracted, trajectory prediction is performed on the key path nodes, and the optimal path is screened; The optimal path is converted into an evaluation function based on linear velocity and angular velocity according to the walking speed of the unmanned excavator, the positions of obstacles around the trajectory are output, and the global optimal path is planned according to the safety distance.

Images