A method and device for planning material accumulation in a cabin of a cabin cleaning robot
By determining the optimal stacking point and path in the material stacking planning method in the cleaning robot cabin, combined with Hybrid A* and ant colony algorithm, the safety risks and low efficiency of manual operation in bulk cargo tank cleaning operations are solved. It realizes efficient and safe automatic cabin cleaning operations.
Patent Information
- Application Number
- CN202411235704.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-04
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-09-04
AI Technical Summary
The cleaning operation of bulk cargo ships relies on manual operations, which poses problems of high personal safety risks and low mechanical efficiency.
Provide a method for material stacking planning in the cabin clearing robot. By determining the optimal material stacking point and path, a directed graph is constructed to achieve full coverage. Combining the Hybrid A* planning algorithm and ant colony algorithm, the path and material stacking order are optimized. Improve cabin cleaning efficiency and safety.
It has achieved no blind spot coverage of the cabin cleaning operation area, significantly improved the efficiency, safety and automation level of cabin cleaning operation, and provided a systematic and intelligent cabin cleaning solution.
Smart Images

Figure CN118735419B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of intelligent ship technology, and in particular to a method and device for planning the accumulation of materials in a cabin of a cabin cleaning robot. Background Art
[0002] Bulk carriers play a very important role in shipping, and the unloading efficiency of bulk carriers directly affects the overall efficiency of shipping. The unloading of bulk carriers can be divided into two parts. On the one hand, when the bulk cargo volume is large, the bulk cargo is directly unloaded through the shore unloader. On the other hand, when the unloading is almost completed, the unloader cannot continue to unload, and it is necessary to use manual or human-driven tank cleaning machinery to enter the cabin for tank cleaning operations. Usually, when there is less remaining material, it is necessary to use a crane to put the tank cleaning machinery into the cabin, and then manually operate it to pile the bulk cargo around the cabin so that the unloader can grab and unload the remaining material.
[0003] At present, the tank cleaning operation of bulk carriers is still carried out by manual driving. The operator needs to go deep into the bottom of the cabin and operate the tank cleaning machinery to pile the cabin materials on the surrounding bulkheads, and then the unloader will grab and unload them.
[0004] Traditional tank cleaning methods rely on manual operations, which have high risks to personal safety and low mechanical efficiency. Therefore, tank cleaning robots based on artificial intelligence and robotics technology have become a hot topic in modern tank cleaning technology research. The key and difficulty of using tank cleaning robots to automatically clean materials in the hold lies in how to carry out efficient material stacking planning for the complex environment inside the bulk carrier. Summary of the invention
[0005] In view of this, it is necessary to provide a method for planning material accumulation in a cabin of a cabin cleaning robot to solve the problems of high personal safety risk and low mechanical efficiency in the prior art.
[0006] In order to solve the above problems, the present application provides a method for planning material accumulation in a cabin of a cabin cleaning robot, comprising:
[0007] Based on the material distribution information in the cabin, a material field for the robot to move is determined, and the optimal stacking points for achieving full coverage of the material field are determined with the goal of minimizing the number of stacking points on the boundary line of the material field;
[0008] The goal is to minimize the sub-path distance from one optimal stockpile point through the geometric center of the material field to another optimal stockpile point, and determine the optimal sub-path and its time cost;
[0009] The optimal stockpiling point and the geometric center of the material yard are taken as nodes, and time weights are assigned to the edges between nodes based on the time cost of the optimal sub-path to construct a directed graph. The optimal total path that achieves full coverage of the materials in the material yard is determined with the goal of minimizing the total path time passing through all nodes.
[0010] In some possible implementations, a material field for robot activity is determined based on the material distribution information in the cabin, and the optimal stacking points for achieving full coverage of the material field are determined with the goal of minimizing the number of stacking points on the boundary line of the material field, including:
[0011] Based on the material distribution information in the cabin, a material field for the robot to move is determined, the material field is abstracted as a rectangle, and a single boundary model corresponding to a certain boundary line of the material field is constructed. The single boundary model aims to minimize the number of stockpiling points on the certain boundary line and achieve full coverage of the materials within the range of the certain boundary line;
[0012] Solving the single boundary model to obtain an optimal stockpiling point on the certain boundary line;
[0013] The coordinates of the optimal stockpile point on the certain boundary line are rotated or translated to determine the optimal stockpile point on all boundary lines of the material yard.
[0014] In some possible implementations, the sub-paths include a first path from an optimal stockpile point to a geometric center of a material field and a second path from the geometric center of the material field to another optimal stockpile point;
[0015] Taking the minimum sub-path distance from one optimal material stacking point to another optimal material stacking point as the goal, determining the optimal sub-path includes:
[0016] Constructing a robot kinematic model, wherein the robot kinematic model adopts an articulated bicycle model, determining a minimum turning radius of the robot based on the robot kinematic model, and determining a minimum turning radius constraint of an RS curve of a heuristic function with kinematic constraints in a HybridA* planning algorithm based on the minimum turning radius;
[0017] Based on the Hybrid A* planning algorithm, performing path planning on the first path to determine an optimal first path;
[0018] A straight line path from the geometric center of the material yard to the other optimal stockpiling point is used as an optimal second path, and an optimal sub-path is determined based on the optimal first path and the optimal second path.
[0019] In some possible implementations, performing path planning on the first path based on the Hybrid A* planning algorithm to determine an optimal first path includes:
[0020] Constructing a material yard navigation map, processing the material yard navigation map based on a rasterization method, and determining a material yard map index coordinate system;
[0021] Determine the starting point and the end point of the first path, and create an openlist table and a closelist table for forward search and reverse search respectively, put the starting point into the closelist table of the forward search, and put the end point into the closelist table of the reverse search;
[0022] Based on the heuristic function of the Hybrid A* planning algorithm, the cost value of node expansion is calculated, and the nodes of the forward search and the reverse search are synchronously expanded based on the cost value;
[0023] When the nodes of the forward search and the reverse search are expanded synchronously, the node coordinates in the physical space where the robot is located are mapped to the grid coordinates in the material field map index coordinate system, and it is determined whether the grid coordinates of the expanded nodes of the forward search and the reverse search are equal;
[0024] If the grid coordinates of the forward search and the reverse search expansion nodes are not equal, then continue to synchronously expand the forward search and the reverse search nodes;
[0025] If the grid coordinates of the forward search and the reverse search expansion nodes are equal, the last child node is selected according to the terminal processing rule, and the closelist tables of the forward search and the reverse search are merged to determine the optimal first path.
[0026] In some possible implementations, the terminal processing rule includes:
[0027] The terminal processing rules include:
[0028] Determine the sum of the cost value from the parent node of the forward search to the child node of the forward search and the cost value from the parent node of the reverse search to the child node of the forward search as the first cost value, and determine the sum of the cost value from the parent node of the forward search to the child node of the reverse search and the cost value from the parent node of the reverse search to the child node of the reverse search as the second cost value;
[0029] Determine whether the first generation value is greater than the second generation value;
[0030] If the first generation value is not greater than the second generation value, retain the terminal child node of the forward search and clear the child node of the reverse search;
[0031] If the first generation value is not greater than the second generation value, the terminal child node of the reverse search is retained, and the child node of the forward search is cleared.
[0032] In some possible implementations, after determining the optimal first path, the method further includes:
[0033] Determine a conversion point where the vehicle speed in the optimal first path changes from positive to negative, and divide the optimal first path into a plurality of first segments based on the conversion point;
[0034] Based on the Bezier curve technology, smoothing is performed on each of the first segments to obtain a smooth path corresponding to each of the first segments;
[0035] The smoothed paths are concatenated to obtain a smoothed optimal first path.
[0036] In some possible implementations, calculating the time cost of the optimal subpath includes:
[0037] Segment the optimal first path according to the nodes on the optimal first path determined by the Hybrid A* planning algorithm to determine a plurality of second segments;
[0038] The path length between the parent node and its adjacent child node is used as the distance of the second segment, and the average speed between the parent node and its adjacent child node is used as the travel speed of the second segment;
[0039] Determine the travel time cost of the second segment based on the distance of the second segment and the travel speed of the second segment, concatenate the travel time costs of the second segment to obtain the travel time cost of the optimal first path, and determine the time cost of the optimal first path based on the travel time cost of the optimal first path;
[0040] The time cost of the optimal second path is determined, and the time cost of the optimal sub-path is determined based on the time cost of the optimal first path and the time cost of the optimal second path.
[0041] In some possible implementations, determining the time cost of the optimal first path based on the travel time cost of the optimal first path includes:
[0042] Determining a vehicle speed conversion time cost of the optimal first path;
[0043] The time cost of the optimal first path is obtained based on the travel time cost of the optimal first path and the vehicle speed conversion time cost of the optimal first path.
[0044] In some possible implementations, the optimal total path for achieving full coverage of materials in the material yard is determined with the goal of minimizing the total path time passing through all nodes, including: based on an ant colony algorithm, the directed graph is solved with the goal of minimizing the total path time passing through all nodes to determine the optimal total path for achieving full coverage of materials in the material yard.
[0045] The present application also provides a device for planning material accumulation in a cabin of a cabin cleaning robot, comprising:
[0046] An optimal stacking point determination unit determines a material field for the robot to move in based on the material distribution information in the cabin, and determines the optimal stacking point that fully covers the material field with the goal of minimizing the number of stacking points on the boundary line of the material field;
[0047] The optimal path planning unit aims to minimize the distance of the sub-path from one optimal stockpile point through the geometric center of the material field to another optimal stockpile point, and determines the optimal sub-path and its time cost;
[0048] The optimal stockpile point selection order determination unit takes the optimal stockpile point and the geometric center of the material yard as nodes, assigns time weights to the edges between nodes based on the time cost of the optimal sub-path to construct a directed graph, and determines the optimal total path for full coverage of materials in the material yard with the goal of minimizing the total path time passing through all nodes.
[0049] The beneficial effects of the present application are as follows: the material stacking planning method for a cabin cleaning robot provided in the present application has provided a matching optimal stacking point selection strategy, a fast path planning algorithm and a material stacking sequence optimization strategy for the complex cabin environment and the robot's characteristic information, and together constructed a full-coverage, systematic and intelligent cabin cleaning robot operation solution specifically for the cabin material field, which not only achieves no-dead-angle coverage of the entire operation area, but also significantly improves the efficiency, safety and automation level of the cabin cleaning operation, and has important practical significance. BRIEF DESCRIPTION OF THE DRAWINGS
[0050] Figure 1 A schematic flow chart of an embodiment of a method for planning the accumulation of materials in a cabin of a cabin cleaning robot provided in the present application;
[0051] Figure 2 Schematic diagram of the material space structure in the cabin provided for this application;
[0052] Figure 3 For this application Figure 1 A schematic flow chart of an embodiment of step S101;
[0053] Figure 4 Schematic diagram of the robot operation process provided for this application;
[0054] Figure 5 A schematic diagram of a plane rectangular coordinate system of the lower boundary model provided in this application;
[0055] Figure 6 A schematic diagram of the rectangular coordinate system of the material field provided for this application;
[0056] Figure 7 For this application Figure 1 A schematic flow chart of an embodiment of step S102;
[0057] Figure 8 Schematic diagram of the robot kinematic model provided for this application;
[0058] Fig. 9 For this application Figure 7 A schematic flow chart of an embodiment of step S702;
[0059] Fig.10 A schematic diagram of forward and reverse search nodes provided for this application;
[0060] Fig.11 A schematic diagram of an embodiment of the terminal processing rule provided in this application;
[0061] Fig.12 A schematic diagram of an embodiment of the optimal sub-path smoothing process provided by the present application;
[0062] Fig.13 A schematic diagram of a flow chart of an embodiment of determining the time cost of an optimal sub-path provided by the present application;
[0063] Fig.14 For this application Fig.13 A flowchart of an embodiment of step S1304;
[0064] Fig.15 A schematic diagram of an embodiment of the ant colony algorithm provided in this application;
[0065] Fig.16 This is a schematic structural diagram of an embodiment of a cabin material accumulation planning device for a cabin cleaning robot provided in the present application. DETAILED DESCRIPTION
[0066] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.
[0067] It should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of the present application. It should be understood that the operations of the flowcharts may be implemented out of order, and steps without logical contextual relationships may be reversed in order or implemented simultaneously. In addition, those skilled in the art may add one or more other operations to the flowcharts, or may remove one or more operations from the flowcharts, under the guidance of the content of this application. Some of the block diagrams shown in the accompanying drawings are functional entities, and do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor systems and / or microcontroller systems.
[0068] The descriptions of "first", "second", etc. involved in the embodiments of the present application are only used to describe the implicit purpose, and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the technical features defined as "first" and "second" may explicitly or implicitly include at least one of the features. "And / or" describes the association relationship of the associated objects, indicating that three relationships may exist, for example: A and / or B, which can represent: A exists alone, A and B exist at the same time, and B exists alone.
[0069] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.
[0070] The present application provides a method for planning material accumulation in a cabin of a cabin cleaning robot, which is described below.
[0071] Figure 1 A schematic diagram of an embodiment of a method for planning the accumulation of materials in a cabin of a cabin cleaning robot provided in this application is shown in FIG. Figure 1 As shown, the material accumulation planning method of the cabin cleaning robot includes:
[0072] S101, based on the material distribution information in the cabin, determine the material field for the robot to move, with the goal of minimizing the number of material stacking points on the boundary line of the material field, and determine the optimal material stacking points that fully cover the material field;
[0073] Specifically, the space structure of materials in the cabin is as follows: Figure 2As shown, the working space of the robot in the cabin can be regarded as a rectangular area, which is subdivided into four functional areas from the outside to the inside: the bulkhead area, the stacking area, the stacking line (the target stacking boundary line) and the working area (the area where the robot moves); among them, the bulkhead area, as the physical boundary of the entire cabin, plays a limiting role in the process of cabin cleaning operations, and the robot's stacking operation must not cross the boundary; the stacking area is the designated area where the robot actually places materials, and has a certain width to accommodate material stacking; the stacking line is the edge of the robot's stacking point, marking the boundary of the target point that the robot should reach, and also serves as a limiting boundary for the robot's movement; the working area covers all the space within the stacking line. In view of the thin material accumulation in this area, the robot can move flexibly in this area. This application will focus on the rectangular working area defined by the stacking line. This area is defined as the "material field" in this application, and all path planning, motion scheduling and optimization strategies for the robot will revolve around this material field.
[0074] It should be noted that the core goal of selecting the optimal stacking point is to minimize the total number of shovels required by the robot and maximize the material utilization efficiency of each shovel while meeting the following multiple conditions. Each shovelling by the robot corresponds to an optimal stacking point, which minimizes the total number of shovels required by the robot, that is, minimizes the number of optimal stacking points.
[0075] S102, determining an optimal sub-path and its time cost with the goal of minimizing the sub-path distance from an optimal stockpile point through the geometric center of the material field to another optimal stockpile point by the robot;
[0076] S103. The optimal stacking point and the geometric center of the material yard are taken as nodes, and time weights are assigned to the edges between the nodes corresponding to the optimal stacking point based on the time cost of the optimal sub-path to construct a directed graph. The goal is to minimize the total path time for the robot to pass through the nodes, and determine the optimal total path for achieving full coverage of the materials in the material yard.
[0077] Compared with the existing technology, this application has developed a matching optimal stacking point selection strategy, a fast path planning algorithm and a material stacking sequence optimization strategy for the complex environment inside the cabin and the characteristic information of the robot, and jointly constructed a full-coverage, systematic and intelligent cabin cleaning robot operation solution specifically for the cabin material field. It not only achieves no-dead-angle coverage of the entire operating area, but also significantly improves the efficiency, safety and automation level of the cabin cleaning operation, which has important practical significance.
[0078] In order to make full use of the symmetry of the rectangular material field model in the vertical and horizontal directions and simplify the complexity of selecting the optimal stockpile point, in some embodiments, as Figure 3 As shown, step S101 includes:
[0079] S301, based on the material distribution information in the cabin, determine the material field for the robot to move, abstract the material field into a rectangle, and construct a single boundary model corresponding to a certain boundary line of the material field. The single boundary model aims to minimize the number of material stacking points on a certain boundary line and achieve full coverage of the materials within a certain boundary line;
[0080] Robot operation flow chart Figure 4 As shown in the figure, the material field can be abstracted as a rectangular model, the robot is a forklift equipped with an intelligent navigation and control system, the bucket carried by the robot can be abstracted as a straight tool with a constant length, and the task goal of the robot in the single boundary model can be abstracted as pushing the material to the optimal stacking point on the boundary line of the material field in a straight line starting from the geometric center of the material field.
[0081] Specifically, take the lower boundary line as an example to explain, the plane rectangular coordinate system of the lower boundary model is as follows: Figure 5 As shown in the figure, the lower boundary line is set as line segment AB, the midpoint of AB is the coordinate origin O, AB is the horizontal axis, and the direction of the line connecting the origin O and the geometric center C of the material field is the vertical axis. A plane rectangular coordinate system is established, and the coordinates of the three points are recorded as A , B and C , assuming the length of the bucket is , on line segment AB The optimal stacking point, that is, the point where the bucket shovels (shovel point), has the following horizontal coordinates from left to right: , assuming The bucket center, shoveling point and the geometric center of the material field are collinear, ensuring that the robot can push the material to the predetermined stockpiling points on the four sides along a straight path each time it shovels, so as to achieve full coverage of the four sides of the rectangular material field by the robot. In addition, the line connecting the bucket center and the geometric center of the material field is perpendicular to the bucket.
[0082] like Figure 5 As shown, at point Department When used as a shovel point, the angle between the bucket push direction and AB is discussed in detail:
[0083] 1) When When , the robot shovels down to the left of the origin:
[0084]
[0085] In the formula, represents the first The angle between the shovel point and the line connecting point C to line segment AB, represents the first The horizontal coordinate of the shovel point, represents the abscissa of point C;
[0086] 2) When When , the robot shovels down at the origin:
[0087]
[0088] 3) When When , the robot shovels down to the right of the origin:
[0089]
[0090] In summary, the angle between the shovel point and the line connecting point C to AB is:
[0091]
[0092] In the formula, represents the first The angle between the shovel point and the line connecting point C to line segment AB, represents the first The horizontal coordinate of the shovel point, Represents the horizontal coordinate of point C.
[0093] like Figure 5 As shown, at point Department When used as shoveling points, the horizontal coordinates of the points swept by the leftmost and rightmost ends of the bucket on AB are discussed in different cases:
[0094] 1) When When the leftmost end of the bucket on AB is swept, the horizontal coordinate of the point and the horizontal coordinate of the point swept by the rightmost end of the bucket satisfy:
[0095]
[0096]
[0097] In the formula, Represents half of the bucket length, The first The horizontal coordinate of the point swept by the leftmost end of the bucket corresponding to the shovel point, represents the first The horizontal coordinate of the point swept by the rightmost end of the bucket corresponding to the shovel point;
[0098] 2) When When the leftmost end of the bucket on AB is swept, the horizontal coordinate of the point and the horizontal coordinate of the point swept by the rightmost end of the bucket satisfy:
[0099]
[0100]
[0101] In summary, no matter which direction the bucket is shoveled, the horizontal coordinates of the points swept by the leftmost and rightmost ends of the bucket on AB satisfy
[0102]
[0103]
[0104] In the formula, represents the first The angle between the shovel point and the line connecting point C to line segment AB, Represents half of the bucket length, represents the first The horizontal coordinate of the shovel point, represents the first The horizontal coordinate of the point swept by the leftmost end of the bucket corresponding to the shovel point, represents the first The horizontal coordinate of the point swept by the rightmost end of the bucket corresponding to the shovel point.
[0105] In order to ensure that there are no gaps in the material when the robot shovels in the material yard, at least one point of the area where the robot shovels twice in succession must overlap, that is:
[0106] 1) The point swept by the leftmost end of the bucket corresponding to the first shovel point on AB should cover point A, and ensure that the rightmost side of the bucket does not exceed the lower boundary line area:
[0107]
[0108] In the formula, It represents the horizontal coordinate of the point swept by the leftmost end of the bucket corresponding to the first shovel point on AB. represents the horizontal coordinate of point A;
[0109] 2) Starting from the second shovel point on line segment AB, the point swept by the leftmost end of the bucket corresponding to each shovel point must cover the point swept by the rightmost end of the bucket corresponding to the previous shovel point:
[0110]
[0111] In the formula, Indicates the first The horizontal coordinate of the point swept by the leftmost end of the bucket corresponding to the shovel point, represents the first The horizontal coordinate of the point swept by the rightmost end of the bucket corresponding to the shovel point;
[0112] 3) The point swept by the rightmost end of the bucket corresponding to the last shovel point on line segment AB should cover point B, and ensure that the rightmost side of the bucket does not exceed the lower boundary line area:
[0113]
[0114] In the formula, represents the horizontal coordinate of point B, It represents the horizontal coordinate of the point swept by the rightmost end of the bucket corresponding to the last shovel point on AB.
[0115] The final lower boundary model is:
[0116]
[0117] S302, solving the single boundary model to obtain the optimal stockpile point on a certain boundary line;
[0118] Specifically, the point swept by the leftmost end of the bucket corresponding to the first shovel point on AB coincides with point A, and the first shovel point on AB can be determined. , and then assume that there is only one point where the areas of two adjacent shovels overlap, and then by the recursive equation , the subsequent virtual shovel point can be calculated , until the last virtual shovel point The value of the right endpoint satisfies: , then it means is the minimum value that satisfies all constraints. If it does not meet the requirements, it means that there are not enough shovels. The value of continues to be verified until the first The value is the minimum number of stockpiling points, virtual shovel point With the original shovel point Compared to the length of the left move, that is, the length of the last two shovels repeated on the lower boundary line .
[0119] To ensure that the repeated coverage of two adjacent shovels is uniform, even if the repeated length Divide evenly between each adjacent shovel, starting from the second shovel point until the n -1 shovel point , all move a part to the left, so that the distance between adjacent shovels increases To cover the area of a unit (that is, each shovel evenly spreads the excess of the last shovel), then you need to move the first and last shovels to the left and move the second shovel to the left. Units, the third shovel moves left units, and so on until n -1 point, that is .
[0120] S303, rotating or translating the coordinates of the optimal stockpile point on a certain boundary line to determine the optimal stockpile point on all boundary lines of the material site.
[0121] Specifically, the rectangular coordinate system of the material field is as follows: Figure 6 As shown in the figure, a rectangular coordinate system is established with the geometric center of the material field as the origin. c , width is 2 d , at this time the geometric center of the material field is Figure 5 In (0, b ) Move to Figure 6 The (0,0) in the material field can be obtained by rotating or translating the coordinates of the optimal stockpile point obtained by the lower boundary model:
[0122] Lower boundary, move the optimal stockpile point coordinates solved by the lower boundary model downward d coordinate units;
[0123] Right boundary, move the optimal stockpiling point coordinates solved by the lower boundary model downward first d After the coordinate units, rotate 90° counterclockwise around the origin;
[0124] Upper boundary, move the optimal stockpiling point coordinates solved by the lower boundary model downward first d After the coordinate units, rotate 180° counterclockwise around the origin;
[0125] Left boundary, move the optimal stockpiling point coordinates solved by the lower boundary model downward first d After the coordinate units, rotate 270° counterclockwise around the origin;
[0126] Taken together, the optimal stockpiling points on the four boundary lines of the entire rectangular material yard are obtained, which can achieve full coverage of the material yard.
[0127] After the optimal stacking point is optimized and selected, not only the position coordinates of the optimal stacking point are obtained, but also the expected orientation angle of the robot at each optimal stacking point is obtained for subsequent planning of the stacking path that fully covers the materials in the material yard.
[0128] In the subsequent planning of the full-coverage stacking path for materials in the material yard, the path from an optimal stacking point through the geometric center of the material yard to another material will be considered as a sub-path, and the sub-path includes a first path from an optimal stacking point to the geometric center of the material yard and a second path from the geometric center of the material yard to another optimal stacking point; wherein the optimal second path is a straight line path determined by the starting point and the end point, and the path planning is relatively simple; and in order to ensure that the robot can follow the straight line from the geometric center of the material yard to the optimal stacking point at the bottom during operation, the robot's bucket needs to adjust its direction in advance during the first path process from the optimal stacking point back to the geometric center of the material yard so that it can push the material to the next preset optimal stacking point along the straight line at one time in the next trip. This path is relatively complex. Considering the kinematic characteristics of the robot, this application uses the Hybrid A* planning algorithm to accurately generate a first path trajectory planning scheme that is consistent with the robot vehicle kinematic model.
[0129] Therefore, in some embodiments, Figure 7 As shown, step S102 includes:
[0130] S701, constructing a robot kinematic model, the robot kinematic model adopts an articulated bicycle model, determining a minimum turning radius of the robot based on the robot kinematic model, and determining a minimum turning radius constraint of an RS curve of a heuristic function with kinematic constraints in a Hybrid A* planning algorithm based on the minimum turning radius;
[0131] It should be noted that the robot's kinematic characteristics in the specific environment of the cabin are that its chassis adopts an articulated forklift structure, and the steering action is mainly achieved by the hydraulic cylinder pushing the hinge pin between the front and rear bodies. This mechanism is very different from the steering method of ordinary cars, which makes its kinematic model unique. In view of the relatively simple material field environment and the fact that the robot usually operates at a low speed during operation, the dynamic effect can be reasonably ignored, and the front and rear bodies can be considered to be rigid bodies in the same plane. Therefore, this application simplifies the robot's kinematics into an articulated bicycle model to accurately describe its movement behavior in the cabin. The robot kinematic model diagram is shown in the figure. Figure 8 As shown, the kinematic constraints are:
[0132]
[0133] Transformed into the equation of state
[0134]
[0135] In the formula, is the center coordinate of the robot’s rear axle, is the center coordinate of the front axis of the cleaning robot, is the heading angle of the cleaning robot, is the hinge deflection angle, is the rear axle speed of the cleaning robot, and the lengths of the front and rear bodies are .
[0136] Based on the above robot kinematic model, the turning radius of the front and rear bodies of the robot can be further derived. Since the lengths of the front and rear bodies are equal, the front and rear turning radius The same, that is
[0137]
[0138] This formula shows that the turning radius of the robot depends on the hinge angle and vehicle length , and the articulation angle has certain upper and lower limits, based on which the minimum turning radius of the cleaning robot can be calculated , providing key parameters for Hybrid A* path planning.
[0139] S702, performing path planning on the first path based on the Hybrid A* planning algorithm to determine the optimal first path;
[0140] S703: taking a straight line path from the geometric center of the material yard to another optimal stockpiling point as an optimal second path, and determining an optimal sub-path based on the optimal first path and the optimal second path.
[0141] Furthermore, in order to enhance the performance and practicability of the Hybrid A* planning algorithm, in some embodiments, the forward and reverse synchronous search algorithm is deeply integrated with the Hybrid A* planning algorithm to significantly improve the overall pathfinding efficiency, such as Fig. 9 As shown, in step S702, based on the Hybrid A* planning algorithm, path planning is performed on the first path to determine the optimal first path, which specifically includes:
[0142] S901, constructing a material field navigation map, processing the material field navigation map based on a rasterization method, and determining a material field map index coordinate system;
[0143] Specifically, a detailed material yard navigation map was established through SLAM technology. On this basis, the rasterization method was used to subdivide the navigation map. According to the pre-set navigation map size and resolution parameters, the map was divided into a series of grid units of equal size. These grid units were numbered one by one from small to large in the form of rows and columns, and standardized coordinate indexes were given to the cabin grid map to form a grid network coordinate system. This is not only conducive to the accurate description and management of the location information of each grid unit, but also greatly facilitates the tracking and maintenance of the topological relationship between the expanded nodes and the nodes to be expanded, thereby achieving more efficient map data management and path planning in the scenario of cabin clearing operations.
[0144] It should be noted that necessary data preprocessing and initialization are required before path planning, including loading the original material site map data, setting the grid resolution, configuring various parameters of the cleaning robot (robot body length, width, robot deflection angle, etc.), and clarifying the initial state position information of the cleaning robot ( x , y , θ ),in x and y Represents the position coordinates of the robot in the actual coordinate system, θ Indicates its current heading angle.
[0145] S902, determine the starting point and the end point of the first path, and create an openlist table and a closelist table for the forward search and the reverse search respectively, put the starting point into the closelist table of the forward search, and put the end point into the closelist table of the reverse search;
[0146] S903, based on the heuristic function of the Hybrid A* planning algorithm, calculating the cost value of node expansion, and synchronously expanding the nodes of the forward search and the reverse search based on the cost value;
[0147] Specifically, during the expansion process, the nodes with the lowest cost are sequentially added to their respective closed lists as parent nodes for the next round of expansion.
[0148] S904, when synchronously expanding the nodes of the forward search and the reverse search, mapping the node coordinates in the physical space where the robot is located to the grid coordinates in the material field map index coordinate system, and determining whether the grid coordinates of the forward search and the reverse search expansion nodes are equal;
[0149] It should be noted that in the process of using the forward and reverse synchronous fast search strategy algorithm, given the phenomenon that the spatial positions of sub-nodes may not completely correspond to each other during the forward and reverse search processes, even if the actual coordinates of the forward and reverse sub-nodes in the same grid unit may be different, the schematic diagram of the forward and reverse search nodes is as follows: Fig.10 As shown, this incomplete overlap may cause the forward and reverse searches to continue to expand along their respective paths, eventually forming two independent candidate paths. To solve this problem, the present application proposes a node coordinate conversion method, which systematically maps the actual coordinates of the nodes in the physical space where the robot is located to the grid coordinate system. In this way, it is only necessary to compare whether the grid coordinates corresponding to the end sub-nodes of the forward and reverse search trajectories are the same to effectively determine whether the two search paths intersect. This strategy significantly improves the search efficiency and ensures the accuracy and consistency of the robot motion planning in the cabin cleaning operation environment. The actual coordinates of the node are converted into grid coordinates as shown in the following formula:
[0150]
[0151] In the formula, , and Respectively represent the row index, column index and orientation number of the grid coordinates, , and Respectively represent the horizontal coordinate, vertical coordinate, and vehicle heading angle of the current path end point. represents the path resolution, Indicates the vehicle heading angle resolution.
[0152] S905, if the grid coordinates of the nodes expanded by the forward search and the reverse search are not equal, then continue to synchronously expand the nodes of the forward search and the reverse search;
[0153] S906: If the grid coordinates of the forward search and reverse search extended nodes are equal, the last child node is selected according to the terminal processing rule, and the closelist tables of the forward search and reverse search are merged to determine the optimal first path.
[0154] In some embodiments, the terminal processing rule is based on a specific node retention rule, and the algorithm will select one of the terminal child nodes to retain and remove the other node, such as Fig.11 As shown, specifically including:
[0155] S1101, determine the sum of the cost value from the parent node of the forward search to the child node of the forward search and the cost value from the parent node of the reverse search to the child node of the forward search as the first generation value, and determine the sum of the cost value from the parent node of the forward search to the child node of the reverse search and the cost value from the parent node of the reverse search to the child node of the reverse search as the second generation value;
[0156] S1102, determining whether the first generation value is greater than the second generation value;
[0157] S1103, if the first generation value is not greater than the second generation value, retain the child node at the end of the forward search and clear the child node of the reverse search;
[0158] S1104. If the first generation value is not greater than the second generation value, the child node at the end of the reverse search is retained and the child node of the forward search is cleared.
[0159] Furthermore, considering that although the obtained optimal sub-path can meet the kinematic constraints of the vehicle, the complex optimal first path is composed of a combination of straight line segments and circular arcs, which inevitably contains curvature mutation points and operations that may cause the vehicle to over-steer, it is necessary to perform detailed smoothing on the optimal first path. As for the straight line path of the optimal second path, since its own geometric characteristics are relatively simple and there is no curvature change, no additional smoothing optimization steps are required.
[0160] When considering smoothing the optimal first path, since the Bezier curve has the advantage of multiple control point design, it can ensure the continuity of the curve in multiple derivatives, so it is suitable for the path smoothing task. Therefore, the Bezier curve is used to smooth the optimal first path in this application. However, when directly applied to the overall driving path, especially at the position where the vehicle driving direction undergoes a 180° reversal, that is, the positive and negative conversion area of the vehicle speed, a single continuous Bezier curve may aggravate the path mismatch phenomenon and cause the vehicle head direction to deviate from the expected trajectory. Therefore, in some embodiments, the continuous Bezier curve is interrupted at the key turning point, and the segmented Bezier curve method is used to smooth the optimal first path. Specifically, Fig.12 As shown, in some embodiments, after determining the optimal first path, the method further includes:
[0161] S1201, determining a conversion point where the vehicle speed in the optimal first path changes from positive to negative, and dividing the optimal first path into a plurality of first segments based on the conversion point;
[0162] S1202, based on Bezier curve technology, smoothing each first segment respectively to obtain a smooth path corresponding to each first segment;
[0163] S1203: Connect the smoothed paths to obtain the smoothed optimal first path.
[0164] Furthermore, in order to improve the overall efficiency of the cleaning operation, the speed planning of each optimal sub-path is refined. Specifically, the speed planning of the optimal sub-path is divided into three core stages:
[0165] First, the process of retreating from an optimal stockpiling point along a straight path to the geometric center of the material field in the optimal first path is the first stage of speed planning. In this stage, when the robot starts, it first accelerates to the maximum allowable safe speed using the maximum safe acceleration and maintains a uniform straight-line driving speed. As it approaches the turning point of the second stage, the robot begins to gradually decelerate according to the maximum safe deceleration to ensure that when entering the curved path, its instantaneous speed can be calculated according to the curvature of the curve and the minimum turning radius of the robot to adapt to the current curve.
[0166] Secondly, the small arc path in the optimal first path that transitions from the straight path to the geometric center of the material field constitutes the second stage of speed planning. Given that the path in this stage is relatively short, the planning requires the robot to strictly adjust the speed according to the curvature of the curve to ensure that it stops accurately and smoothly at the geometric center of the material field.
[0167] Finally, the process of the optimal second path starting from the geometric center of the material yard and driving in a straight line to the next optimal stacking point and finally stopping constitutes the third stage of speed planning. This stage is similar to the first stage. It also uses the maximum safe acceleration to accelerate to the maximum safe speed and maintain a constant straight line. When approaching the target optimal stacking point, the robot decelerates in advance again with the maximum safe deceleration to ensure that it stops completely when it reaches the optimal stacking point accurately.
[0168] Since the cabin cleaning robot adopts fully automated operation and does not require manual driving, during the implementation of speed planning, the patent mainly considers how to maximize operating efficiency while ensuring safety without having to worry about driving comfort. Through the above-mentioned speed planning scheme for each optimal sub-path, it can be ensured that the cabin cleaning robot can maximize time efficiency in each link, thereby improving the overall efficiency of the entire operation process.
[0169] Based on the path and speed planning of the optimal sub-path, the time cost of the optimal sub-path is determined. Specifically, in some embodiments, Fig.13 As shown, the time cost of determining the optimal sub-path in step S102 includes:
[0170] S1301, segmenting the optimal first path according to the nodes on the optimal first path determined by the Hybrid A* planning algorithm to determine a plurality of second segments;
[0171] S1302: taking the path length between the parent node and its adjacent child node as the distance of the second segment, and taking the average speed between the parent node and its adjacent child node as the travel speed of the second segment;
[0172] S1303: determining the travel time cost of the second segment based on the distance of the second segment and the travel speed of the second segment, and concatenating the travel time costs of the second segment to obtain the travel time cost of the optimal first path;
[0173] Specifically, the travel time cost calculation formula is:
[0174]
[0175] In the formula, are the coordinates of each node in the path, is the index of all nodes in the path, , each subprocess has nodes, The Euclidean distance for each segment is used to approximate the length of the segment curve. is the instantaneous velocity of each node, is the average driving speed for each segment, The time it takes for the cleaning robot to travel in each segment.
[0176] S1304, determining the time cost of the optimal first path based on the travel time cost of the optimal first path;
[0177] S1305: Determine the time cost of the optimal second path. Based on the time cost of the optimal first path and the time cost of the optimal second path, determine the time cost of the optimal sub-path.
[0178] Considering that during the execution of each optimal sub-path, the cleaning robot will experience two significant positive and negative speed conversion stages. The first conversion occurs at the starting optimal stockpiling point. At this time, the robot switches from the previous stockpiling action to retreat to the geometric center of the material field, which will trigger a reversal of the speed direction. The second conversion occurs after the robot completes the retreat to the geometric center of the material field and drives forward again to the next optimal stockpiling point. The time cost of these two stages cannot be ignored. Therefore, in some embodiments, not only the conventional time cost of path driving is considered, but also the additional time consumption caused by speed conversion is fully considered, so as to ensure that the evaluation of the total time cost of the sub-process is more comprehensive and accurate. Specifically, Fig.14 As shown, step S1304 includes:
[0179] S1401, determining the vehicle speed conversion time cost of the optimal first path;
[0180] It should be noted that there will inevitably be two positive and negative speed conversion states in each optimal first path, and the time losses in these two periods are roughly similar. Based on this understanding, a fixed speed conversion time cost can be set.
[0181] S1402: Obtain the time cost of the optimal first path based on the travel time cost of the optimal first path and the vehicle speed conversion time cost of the optimal first path.
[0182] When constructing a directed graph, the time cost weight from the optimal stockpile point A to B is usually different from that from B to A, that is, there is asymmetry. Therefore, the model is an asymmetric TSP model (ATSP). At the same time, the cleaning operation has a significant feature, that is, the path planning only considers other optimal stockpile points in its vicinity from the current optimal stockpile point, which significantly reduces the complexity of the planned path; however, this feature also brings a challenge, that is, not any two optimal stockpile points can directly establish an effective path connection, and only certain points in the neighborhood can maintain connectivity, which is difficult to convert into a standard This creates difficulties when modeling the Traveling Salesman Problem (TSP) because the TSP model requires that all optimal stockpiling points in the graph are connected to each other. To overcome this difficulty, the present application introduces virtual edges between the optimal stockpiling points that are not directly connected, and assigns larger weight values to these virtual edges to symbolize the high cost of reaching other optimal stockpiling points through indirect paths. In this way, by constructing a complete graph containing virtual edges, it is ensured that the problem can be adapted and transformed into a form that meets the requirements of the TSP model, thereby effectively solving the modeling difficulties of the path planning problem while maintaining the characteristics of the cabin cleaning operation.
[0183] For the cabin cleaning operation scenario, the ATSP model can be specifically described as follows: the cabin cleaning robot starts from the geometric center (starting point) of a material yard, and needs to visit all other optimal stockpiling points in sequence without repetition, and finally return to the starting point. The goal is to minimize the cumulative time cost of the entire operation process.
[0184] In some embodiments, step S103 includes: based on the ant colony algorithm, solving the directed graph with the goal of minimizing the total path time passing through all nodes, and determining the optimal total path for achieving full coverage of materials in the material yard.
[0185] like Fig.15 As shown in the figure, the specific process of the ant colony algorithm is as follows:
[0186] S1501, constructing input: determining a distance matrix based on the time cost of the optimal sub-path, and using the distance matrix as input data of the ant colony algorithm;
[0187] It should be noted that the matrix records the time cost between the optimal stockpiling points in detail, providing the necessary decision-making basis for the algorithm.
[0188] S1502, initializing ant colony parameters: setting key ant colony algorithm parameters, including the number of ants, the initial value of pheromone concentration, the pheromone volatility coefficient, and the heuristic factor;
[0189] It should be noted that the selection of these parameters must ensure the convergence of the algorithm while taking into account the global and local characteristics of the search. After completing the parameter setting, the ant colony system is initialized to prepare for the subsequent search process.
[0190] S1503, ants randomly initialize paths: following the probability rule, each ant builds a starting path between the optimal stacking points;
[0191] It should be noted that in this process, not only the random connection between the optimal stockpiling points is considered, but also the time cost weight is fully incorporated to ensure that the construction of the initial path has a certain degree of randomness and reflects the importance of time efficiency.
[0192] S1504, ant movement and path update: In each iteration, each ant selects the next optimal stockpile point to visit based on the current pheromone concentration, heuristic information (such as remaining time cost) and random disturbance;
[0193] It should be noted that through continuous iteration, ants constantly adjust and improve their respective paths, gradually approaching the global optimal solution.
[0194] S1505, pheromone update: after each round of iteration, the global pheromone distribution is updated according to the time cost of the path found by the ants;
[0195] It should be noted that for the edges passed by ants that find shorter paths (low time cost), the corresponding pheromone concentration is enhanced; conversely, for the edges passed by ants that walk longer paths (high time cost), the pheromone concentration is appropriately attenuated. This dynamic adjustment mechanism ensures that pheromones can guide excellent paths.
[0196] S1506, loop iteration and termination condition: repeat step S1404 and step S1405 until the preset maximum number of iterations is reached or the overall time cost meets the expected optimization threshold;
[0197] It should be noted that at this point, the algorithm enters the convergence stage and it can be considered that a nearly optimal excavation sequence has been found.
[0198] S1507, output the optimal path: when the algorithm terminates, the total path with the minimum time cost is selected from the numerous ant paths as the optimal solution, that is, the ideal stacking order of the cleaning robot when operating in the material yard is determined, and then the optimal total path for achieving full coverage of the materials in the material yard is determined.
[0199] In order to better implement a method for planning the accumulation of materials in a cabin of a cleaning robot in an embodiment of the present application, based on a method for planning the accumulation of materials in a cabin of a cleaning robot, correspondingly, Fig.16As shown, the embodiment of the present application also provides a cabin material accumulation planning device 1600 of a cabin cleaning robot, comprising:
[0200] The optimal stacking point determination unit 1601 is used to determine the material field for the robot to move based on the material distribution information in the cabin, and to determine the optimal stacking point that fully covers the material field with the goal of minimizing the number of stacking points on the boundary line of the material field;
[0201] The optimal path planning unit 1602 determines the optimal sub-path and its time cost with the goal of minimizing the sub-path distance of the robot from one optimal stockpile point through the geometric center of the material field to another optimal stockpile point;
[0202] The optimal stockpile point selection order determination unit 1603 takes the optimal stockpile point and the geometric center of the material yard as nodes, assigns time weights to the edges between the nodes corresponding to the optimal stockpile point based on the time cost of the optimal sub-path, and constructs a directed graph. The optimal total path that achieves full coverage of the materials in the material yard is determined with the goal of minimizing the total path time passing through all nodes.
[0203] The device for planning the accumulation of materials in the cabin of a tank cleaning robot provided in the above embodiment can realize the technical solution described in the embodiment of the method for planning the accumulation of materials in the cabin of a tank cleaning robot. The specific implementation principles of the above units can refer to the corresponding contents in the embodiment of the method for planning the accumulation of materials in the cabin of a tank cleaning robot, which will not be repeated here.
[0204] The above is a detailed introduction to a method for planning material accumulation in a cabin of a cabin cleaning robot provided by the present application. This article uses specific examples to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for technical personnel in this field, based on the ideas of the present application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.
[0205] The above is only a preferred specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed in the present application should be covered within the protection scope of the present application.
Claims
1. A method for planning material accumulation in a cabin of a cabin cleaning robot, characterized in that: include: Based on the material distribution information in the cabin, a material field for the robot to move is determined, and the optimal stacking points for achieving full coverage of the material field are determined with the goal of minimizing the number of stacking points on the boundary line of the material field; The goal is to minimize the sub-path distance from one optimal stockpile point through the geometric center of the material field to another optimal stockpile point, and determine the optimal sub-path and its time cost; The optimal stockpile point and the geometric center of the material yard are taken as nodes, and a directed graph is constructed by assigning time weights to the edges between the nodes based on the time cost of the optimal sub-path. The optimal total path that fully covers the materials in the material yard is determined with the goal of minimizing the total path time passing through all nodes. The material field for the robot to move is determined based on the material distribution information in the cabin, and the optimal stacking points for achieving full coverage of the material field are determined with the goal of minimizing the number of stacking points on the boundary line of the material field, including: Based on the material distribution information in the cabin, a material field for the robot to move is determined, the material field is abstracted as a rectangle, and a single boundary model corresponding to a certain boundary line of the material field is constructed. The single boundary model aims to minimize the number of stockpiling points on the certain boundary line and achieve full coverage of the materials within the range of the certain boundary line; Solving the single boundary model to obtain an optimal stockpiling point on the certain boundary line; The coordinates of the optimal stockpile point on the certain boundary line are rotated or translated to determine the optimal stockpile point on all boundary lines of the material yard.
2. The method for planning the accumulation of materials in a cabin of a cabin cleaning robot according to claim 1, characterized in that: The sub-paths include a first path from an optimal stockpile point to the geometric center of the material field and a second path from the geometric center of the material field to another optimal stockpile point; Taking the minimum sub-path distance from one optimal material stacking point to another optimal material stacking point as the goal, determining the optimal sub-path includes: Constructing a robot kinematic model, wherein the robot kinematic model adopts an articulated bicycle model, determining a minimum turning radius of the robot based on the robot kinematic model, and determining a minimum turning radius constraint of an RS curve of a heuristic function with kinematic constraints in a Hybrid A* planning algorithm based on the minimum turning radius; Based on the Hybrid A* planning algorithm, performing path planning on the first path to determine an optimal first path; A straight line path from the geometric center of the material yard to the other optimal stockpiling point is used as an optimal second path, and an optimal sub-path is determined based on the optimal first path and the optimal second path.
3. The method for planning the accumulation of materials in a cabin of a cleaning robot according to claim 2, characterized in that: Based on the Hybrid A* planning algorithm, performing path planning on the first path to determine an optimal first path includes: Constructing a material yard navigation map, processing the material yard navigation map based on a rasterization method, and determining a material yard map index coordinate system; Determine the starting point and the end point of the first path, and create an openlist table and a closelist table for forward search and reverse search respectively, put the starting point into the closelist table of the forward search, and put the end point into the closelist table of the reverse search; Based on the heuristic function of the Hybrid A* planning algorithm, the cost value of node expansion is calculated, and the nodes of the forward search and the reverse search are synchronously expanded based on the cost value; When the nodes of the forward search and the reverse search are expanded synchronously, the node coordinates in the physical space where the robot is located are mapped to the grid coordinates in the material field map index coordinate system, and it is determined whether the grid coordinates of the expanded nodes of the forward search and the reverse search are equal; If the grid coordinates of the forward search and the reverse search expansion nodes are not equal, then continue to synchronously expand the forward search and the reverse search nodes; If the grid coordinates of the forward search and the reverse search expansion nodes are equal, the last child node is selected according to the terminal processing rule, and the closelist tables of the forward search and the reverse search are merged to determine the optimal first path.
4. The method for planning the accumulation of materials in a cabin of a cabin cleaning robot according to claim 3, characterized in that: The terminal processing rules include: Determine the sum of the cost value from the parent node of the forward search to the child node of the forward search and the cost value from the parent node of the reverse search to the child node of the forward search as the first cost value, and determine the sum of the cost value from the parent node of the forward search to the child node of the reverse search and the cost value from the parent node of the reverse search to the child node of the reverse search as the second cost value; Determine whether the first generation value is greater than the second generation value; If the first generation value is not greater than the second generation value, retain the terminal child node of the forward search and clear the child node of the reverse search; If the first generation value is not greater than the second generation value, the terminal child node of the reverse search is retained, and the child node of the forward search is cleared.
5. The method for planning the accumulation of materials in a cabin of a cabin cleaning robot according to claim 2, characterized in that: After determining the optimal first path, the method further includes: Determine a conversion point where the vehicle speed in the optimal first path changes from positive to negative, and divide the optimal first path into a plurality of first segments based on the conversion point; Based on the Bezier curve technology, smoothing is performed on each of the first segments to obtain a smooth path corresponding to each of the first segments; The smoothed paths are concatenated to obtain a smoothed optimal first path.
6. The method for planning the accumulation of materials in a cabin of a cleaning robot according to claim 2, characterized in that: Calculating the time cost of the optimal subpath includes: Segment the optimal first path according to the nodes on the optimal first path determined by the Hybrid A* planning algorithm to determine a plurality of second segments; The path length between the parent node and its adjacent child node is used as the distance of the second segment, and the average speed between the parent node and its adjacent child node is used as the travel speed of the second segment; Determine the travel time cost of the second segment based on the distance of the second segment and the travel speed of the second segment, concatenate the travel time costs of the second segment to obtain the travel time cost of the optimal first path, and determine the time cost of the optimal first path based on the travel time cost of the optimal first path; The time cost of the optimal second path is determined, and the time cost of the optimal sub-path is determined based on the time cost of the optimal first path and the time cost of the optimal second path.
7. The method for planning the accumulation of materials in a cabin of a cabin cleaning robot according to claim 6, characterized in that: Determining the time cost of the optimal first path based on the travel time cost of the optimal first path includes: Determining a vehicle speed conversion time cost of the optimal first path; The time cost of the optimal first path is obtained based on the travel time cost of the optimal first path and the vehicle speed conversion time cost of the optimal first path.
8. The method for planning the accumulation of materials in a cabin of a cabin cleaning robot according to claim 1, characterized in that: With the goal of minimizing the total path time passing through all nodes, the optimal total path for achieving full coverage of materials in the material yard is determined, including: based on the ant colony algorithm, the directed graph is solved with the goal of minimizing the total path time passing through all nodes to determine the optimal total path for achieving full coverage of materials in the material yard.
9. A material accumulation planning device for a cabin cleaning robot, characterized in that: include: An optimal stacking point determination unit determines a material field for the robot to move in based on the material distribution information in the cabin, and determines the optimal stacking point that fully covers the material field with the goal of minimizing the number of stacking points on the boundary line of the material field; The optimal path planning unit aims to minimize the distance of the sub-path from one optimal stockpile point through the geometric center of the material field to another optimal stockpile point, and determines the optimal sub-path and its time cost; The optimal stockpile point selection order determination unit takes the optimal stockpile point and the geometric center of the material yard as nodes, assigns time weights to the edges between the nodes based on the time cost of the optimal sub-path to construct a directed graph, and determines the optimal total path that fully covers the materials in the material yard with the goal of minimizing the total path time passing through all nodes; The material field for the robot to move is determined based on the material distribution information in the cabin, and the optimal stacking points for achieving full coverage of the material field are determined with the goal of minimizing the number of stacking points on the boundary line of the material field, including: Based on the material distribution information in the cabin, a material field for the robot to move is determined, the material field is abstracted as a rectangle, and a single boundary model corresponding to a certain boundary line of the material field is constructed. The single boundary model aims to minimize the number of stockpiling points on the certain boundary line and achieve full coverage of the materials within the range of the certain boundary line; Solving the single boundary model to obtain an optimal stockpiling point on the certain boundary line; The coordinates of the optimal stockpile point on the certain boundary line are rotated or translated to determine the optimal stockpile point on all boundary lines of the material yard.
