Cluster storage robot system control method based on artificial intelligence

Abstract

The invention belongs to the technical field of logistics carrying, and discloses a cluster storage robot system control method based on artificial intelligence. Comprising the steps of collecting basic information data of a storage robot and basic information data of goods; a storage robot-cargo azimuth map is obtained according to the basic information data of the storage robot and the basic information data of the cargo, a linear distance and a carrying distance are obtained according to the storage robot-cargo azimuth map, and carrying tasks are distributed according to the linear distance, the carrying distance and the basic information data of the storage robot; according to the carrying task distribution result, path planning and path optimization are carried out, the optimized path is obtained, the storage robot is controlled to carry out goods carrying along the optimized path, the carrying distance is further shortened, the carrying efficiency of the storage robot is improved, and then the operation efficiency of a warehouse factory is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of logistics handling, and more specifically, to a control method for a cluster warehousing robot system based on artificial intelligence. Background Art

[0002] The patent with the application publication number CN108820664A discloses an intelligent warehousing system based on cluster warehousing robots, which is composed of cluster warehousing robots, a goods storage area, a robot passage area, and a shipping area; the goods storage area is used for storing goods and delivering goods to the cluster warehousing robots, the robot passage area is used for guiding the movement of the cluster warehousing robots, and the shipping area is used for the cluster warehousing robots to unload goods; the cluster warehousing robots go to the goods storage area through the robot passage area to carry goods, load the goods after reaching the goods storage area, go to the shipping area through the robot passage area to unload the goods, and then return to enter the next round of handling operations; this patent application optimizes the internal layout distribution and functional area distribution of the warehousing system in view of the characteristics of the cluster warehousing robot system such as strong scalability, low coupling degree, and emergence, so as to improve the efficiency of intelligent warehousing in many aspects, provide support for the normal operation of the system, this patent application has good economy, low maintenance cost, and high reliability, and has the prospect of popularization and application.

[0003] However, during the process of the warehousing robots carrying goods, the warehousing robots are distributed at various different positions, and the goods are also distributed at various different positions. The positions of the warehousing robots and the goods will have a great impact on the handling task allocation. If the handling task allocation is unreasonable, some warehousing robots will be idle for a long time, while some other warehousing robots will be overloaded with work. It will also cause the warehousing robots to repeat tasks or move invalidly, and will also cause some warehousing robots to travel a long distance to pick up goods, resulting in a decrease in handling efficiency, a decrease in the operating efficiency of the warehousing factory, and the transportation paths of the warehousing robots are numerous and complex. If the transportation paths of the warehousing robots are not optimized, it will cause an increase in transportation paths, resulting in an increase in transportation time and a decrease in the handling efficiency of the warehousing robots.

[0004] In view of this, the present invention proposes a control method for a cluster warehousing robot system based on artificial intelligence to solve the above problems. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art and to achieve the above object, the present invention provides the following technical solution: A control method for a cluster warehousing robot system based on artificial intelligence, including:

[0006] Step S1: Collect the basic information data of the warehousing robots and the basic information data of the goods;

[0007] Step S2: Obtain the storage robot - goods orientation map based on the basic information data of the storage robot and the basic information data of the goods. Obtain the straight - line distance and the handling distance according to the storage robot - goods orientation map. Optimize the ant colony algorithm by dynamically adjusting the pheromone evaporation rate and introducing a reinforcement learning reward factor. Perform handling task allocation using the optimized ant colony algorithm based on the straight - line distance, the handling distance, and the basic information data of the storage robot.

[0008] Step S3: According to the handling task allocation result, preliminarily plan the path between the storage robot and the initial position of the goods to obtain the preliminary path between the storage robot and the initial position of the goods.

[0009] Step S4: Optimize the preliminary path between the storage robot and the initial position of the goods to obtain the optimized path between the storage robot and the initial position of the goods. Preliminarily plan and optimize the path between the initial position of the goods and the handling destination position of the goods to obtain the optimized path between the initial position of the goods and the handling destination position of the goods. Control the storage robot to carry the goods along the optimized path between the storage robot and the initial position of the goods and the optimized path between the initial position of the goods and the handling destination position of the goods.

[0010] Further, the basic information data of the storage robot includes the current remaining power of the storage robot and the longitude and latitude of its position;

[0011] The basic information data of the goods includes the longitude and latitude of the initial position of the goods and the longitude and latitude of the handling destination position of the goods.

[0012] Further, the method for obtaining the storage robot - goods orientation map based on the basic information data of the storage robot and the basic information data of the goods includes:

[0013] Establish a blank two - dimensional rectangular coordinate system. Set the abscissa of the blank two - dimensional rectangular coordinate system as longitude and the ordinate as latitude;

[0014] Fill the position of the storage robot into the blank two - dimensional rectangular coordinate system according to the longitude and latitude of the position of the storage robot to obtain the storage robot position mapping point. Fill the initial position of the goods into the blank two - dimensional rectangular coordinate system according to the longitude and latitude of the initial position of the goods to obtain the initial position of the goods mapping point. Fill the handling destination position of the goods into the blank two - dimensional rectangular coordinate system according to the longitude and latitude of the handling destination position of the goods to obtain the handling destination position of the goods mapping point. Connect the initial position of the goods mapping point with the corresponding handling destination position of the goods mapping point to obtain the storage robot - goods orientation map.

[0015] Further, the method for obtaining the straight-line distance and the handling distance according to the warehousing robot-cargo orientation map includes:

[0016] Obtain the straight-line distance L between the mapping point of the warehousing robot's position and the mapping point of the initial cargo position according to the warehousing robot-cargo orientation map ik and the straight-line distance L between the mapping point of the initial cargo position and the mapping point of the cargo handling destination position kj ;

[0017] Among them, i is the index of the mapping point of the warehousing robot's position, k is the index of the mapping point of the initial cargo position, and L ik is the straight-line distance between the i-th mapping point of the warehousing robot's position and the k-th mapping point of the initial cargo position, x i is the abscissa of the i-th mapping point of the warehousing robot's position, x k is the abscissa of the k-th mapping point of the initial cargo position, y i is the ordinate of the i-th mapping point of the warehousing robot's position, y k is the ordinate of the k-th mapping point of the initial cargo position;

[0018] Among them, j is the index of the mapping point of the cargo handling destination position, and j = k, L kj is the straight-line distance between the k-th mapping point of the initial cargo position and the j-th mapping point of the cargo handling destination position, x j is the abscissa of the j-th mapping point of the cargo handling destination position, y j is the ordinate of the j-th mapping point of the cargo handling destination position;

[0019] Add the straight-line distance between the mapping point of the warehousing robot's position and the mapping point of the initial cargo position and the straight-line distance between the mapping point of the initial cargo position and the mapping point of the cargo handling destination position to obtain the handling distance L between the mapping point of the warehousing robot's position and the mapping point of the cargo handling destination position ij ;

[0020] Among them, L ij = L ik + L kj ; L ij is the handling distance between the i-th mapping point of the warehousing robot's position and the j-th mapping point of the cargo handling destination position.

[0021] Further, the method for performing handling task allocation using the optimized ant colony algorithm based on the straight-line distance, the handling distance, and the basic information data of the warehousing robot includes:

[0022] Step S51: Set the initial pheromone τ for each mapping point of the cargo handling destination positionij (0);

[0023] Among them, τ ij (0) represents the initial pheromone of the j-th cargo handling destination location mapping point when the i-th warehousing robot corresponding to the i-th warehousing robot location mapping point selects the cargo corresponding to the j-th cargo handling destination location mapping point for handling;

[0024] Step S52: Calculate the probability P that the i-th warehousing robot corresponding to the i-th warehousing robot location mapping point selects the cargo corresponding to the j-th cargo handling destination location mapping point for handling according to the task selection formula ij ;

[0025] Among them, the task selection formula is: η ij is the reciprocal of the handling distance between the i-th warehousing robot location mapping point and the j-th cargo handling destination location mapping point, m is the number of cargo handling destination location mapping points, α is the influence factor of the initial pheromone, the value range of α is [0, 1], β is the influence factor of the reciprocal of the handling distance between the i-th warehousing robot location mapping point and the j-th cargo handling destination location mapping point, the value range of β is [0, 1], R ij is the reinforcement learning reward factor when the i-th warehousing robot corresponding to the i-th warehousing robot location mapping point selects the cargo corresponding to the j-th cargo handling destination location mapping point for handling;

[0026] Step S53: Select the warehousing robot with the highest probability to simulate the handling of the cargo, and obtain the power consumption of the warehousing robot after the simulated handling;

[0027] Update the initial pheromone according to the power consumption of the warehousing robot after the simulated handling, and obtain the updated initial pheromone τ ij - (0);

[0028] Among them, τ ij - (0) = (1 - ρ ij ) × τ ij (0) + Δτ ij (0); ρ ij is the pheromone evaporation rate, ρ ij = ρ min + (ρ max - ρ min ) × e -γ×Iter ; ρ max is the maximum pheromone evaporation rate, ρ min is the minimum pheromone evaporation rate, e is the base of the natural logarithm, γ is the adjustment parameter, Iter is the current update times, Δτij (0) To increase pheromone, Q is a constant, is the power consumption of the i-th warehousing robot after simulating handling, and λ is the adjustment coefficient of;

[0029] Step S54: Repeat steps S52 and S53 until the set update times N are reached, and obtain the probability that the i-th warehousing robot corresponding to the position mapping point after N updates selects the goods corresponding to the j-th goods handling destination position mapping point for handling. For each good, obtain the warehousing robot with the highest probability after N updates. One good represents one handling task, and allocate the goods to the warehousing robot with the highest probability after N updates accordingly.

[0030] Furthermore, the reinforcement learning reward factor consists of the maximum power of the warehousing robot, the current remaining power, and the straight-line distance;

[0031] Among them, E max is the maximum power of the warehousing robot, and E i is the current remaining power of the warehousing robot corresponding to the i-th warehousing robot position mapping point, δ is the adjustment parameter of, and θ is the adjustment parameter of, and the value ranges of both δ and θ are [0, 1].

[0032] Furthermore, the method for initially planning the path between the warehousing robot and the initial position of the goods according to the handling task allocation result to obtain the initial path between the warehousing robot and the initial position of the goods includes:

[0033] Perform Cartesian grid division on the warehousing robot-goods orientation map, mark the center of each grid as a node, divide the warehousing robot position mapping point, the goods initial position mapping point, and the goods handling destination position mapping point into the corresponding grids, and obtain the positions of the obstacles during the handling process of the warehousing robot. Map the obstacles to the warehousing robot-goods orientation map after Cartesian grid division according to the positions of the obstacles during the handling process of the warehousing robot;

[0034] Mark the grid where the warehousing robot position mapping point is located as the initial grid, mark the grid where the goods initial position mapping point is located as the target grid, obtain the eight adjacent grids of the initial grid, and calculate the path cost D for moving from the initial grid to each adjacent grid;

[0035] where D = U1 + U2; U1 is the straight-line distance between the nodes of the initial grid and the nodes of the adjacent grid, and U2 is the straight-line distance between the nodes of the adjacent grid and the nodes of the target grid. If there is an obstacle in the adjacent grid, the path cost D from the initial grid to move to this adjacent grid is +∞;

[0036] Denote the adjacent grid with the minimum path cost as the first grid, use the first grid as the moving destination of the initial grid, obtain the eight adjacent grids of the first grid, calculate the path cost from the first grid to each adjacent grid, and obtain the minimum value among the path costs from the first grid to each adjacent grid. Denote the adjacent grid corresponding to the minimum value as the second grid, use the second grid as the moving destination of the first grid, and so on, repeating in a cycle until moving to the target grid to obtain the preliminary path between the warehousing robot and the initial position of the goods.

[0037] Further, the method for dividing the position mapping point of the warehousing robot, the initial position mapping point of the goods, and the handling destination position mapping point of the goods into the corresponding grids includes:

[0038] Obtain the node closest to the position mapping point of the warehousing robot, and divide the position mapping point of the warehousing robot into the grid corresponding to this node;

[0039] Obtain the node closest to the initial position mapping point of the goods, and divide the initial position mapping point of the goods into the grid corresponding to this node;

[0040] Obtain the node closest to the handling destination position mapping point of the goods, and divide the handling destination position mapping point of the goods into the grid corresponding to this node.

[0041] Further, the method for optimizing the preliminary path between the warehousing robot and the initial position of the goods to obtain the optimized path between the warehousing robot and the initial position of the goods includes:

[0042] During the process of moving from the initial grid to the first grid, if the first grid is directly above, directly below, directly to the left, or directly to the right of the initial grid, the process of moving the initial grid to the first grid is recorded as translation. If the first grid is in the upper left diagonal, lower left diagonal, upper right diagonal, or lower right diagonal of the initial grid, the process of moving the initial grid to the first grid is recorded as diagonal movement;

[0043] During the process of moving from the initial grid to the target grid, if there are two consecutive translations, obtain the grid before the first translation and the grid after the second translation. If the grid before the first translation directly reaches the grid after the second translation through a skew shift, optimize the path between the grid before the first translation and the grid after the second translation, delete the two translations, and directly perform a skew shift to obtain the optimized path between the warehousing robot and the initial position of the goods.

[0044] Furthermore, the method for initially planning and optimizing the path between the initial position of the goods and the destination position of the goods handling to obtain the optimized path between the initial position of the goods and the destination position of the goods handling includes:

[0045] Based on the initial planning and path optimization of the path between the warehousing robot and the initial position of the goods, initially plan and optimize the path between the initial position of the goods and the destination position of the goods handling to obtain the optimized path between the initial position of the goods and the destination position of the goods handling.

[0046] The technical effects and advantages of a control method for a cluster warehousing robot system based on artificial intelligence according to the present invention:

[0047] 1. Obtain a warehousing robot - goods orientation map based on the basic information data of the warehousing robot and the basic information data of the goods, clearly showing the position information of the warehousing robot and the goods, providing a solid foundation for subsequent handling task allocation;

[0048] 2. Obtain the straight-line distance and handling distance according to the warehousing robot - goods orientation map, and allocate handling tasks according to the straight-line distance, handling distance, and basic information data of the warehousing robot. While ensuring that the warehousing robot does not travel a long distance to pick up goods, it mobilizes the warehousing robot with more power to execute tasks, does not cause the warehousing robot to be overloaded, and also ensures that the warehousing robot has enough power to complete the handling task, thereby reducing the handling time and improving the handling efficiency;

[0049] 3. According to the result of the handling task assignment, preliminarily plan the path between the warehousing robot and the initial position of the goods to obtain the preliminary path between the warehousing robot and the initial position of the goods. Optimize the preliminary path between the warehousing robot and the initial position of the goods to obtain the optimized path between the warehousing robot and the initial position of the goods. Preliminarily plan and optimize the path between the initial position of the goods and the destination position of the goods handling to obtain the optimized path between the initial position of the goods and the destination position of the goods handling. Control the warehousing robot to carry out goods handling along the optimized path between the warehousing robot and the initial position of the goods and the optimized path between the initial position of the goods and the destination position of the goods handling. The preliminary path is optimized, ensuring that the warehousing robot carries out goods handling along the shortest path, thereby reducing the handling time, improving the handling efficiency, and improving the operation efficiency of the warehousing factory. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 Schematic diagram of a control method for an artificial intelligence-based cluster warehousing robot system according to the present invention;

[0051] Figure 2 Schematic diagram of a control system for an artificial intelligence-based cluster warehousing robot system according to the present invention;

[0052] Figure 3 Schematic diagram of the handling task assignment method according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0053] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0054] Embodiment 1

[0055] Please refer to [[ID=2`6]] Figure 1 and Figure 3 As shown, a control method for an artificial intelligence-based cluster warehousing robot system according to this embodiment includes:

[0056] Step S1: Collect the basic information data of the warehousing robot and the basic information data of the goods;

[0057] Step S2: Obtain the storage robot-cargo orientation map based on the basic information data of the storage robot and the basic information data of the cargo. Obtain the straight-line distance and handling distance according to the storage robot-cargo orientation map. Optimize the ant colony algorithm by dynamically adjusting the pheromone evaporation rate and introducing a reinforcement learning reward factor. Perform handling task allocation using the optimized ant colony algorithm based on the straight-line distance, handling distance, and the basic information data of the storage robot;

[0058] Step S3: According to the handling task allocation result, preliminarily plan the path between the storage robot and the initial position of the cargo to obtain the preliminary path between the storage robot and the initial position of the cargo;

[0059] Step S4: Optimize the preliminary path between the storage robot and the initial position of the cargo to obtain the optimized path between the storage robot and the initial position of the cargo. Preliminarily plan and optimize the path between the initial position of the cargo and the handling destination position of the cargo to obtain the optimized path between the initial position of the cargo and the handling destination position of the cargo. Control the storage robot to carry the cargo along the optimized path between the storage robot and the initial position of the cargo and the optimized path between the initial position of the cargo and the handling destination position of the cargo.

[0060] The process of collecting the position data of the storage robot and the basic information data of the cargo includes:

[0061] The basic information data of the storage robot includes the current remaining power of the storage robot and the longitude and latitude of the position. The basic information data of the cargo includes the longitude and latitude of the initial position of the cargo and the longitude and latitude of the handling destination position of the cargo;

[0062] Set up a cargo basic information data collection terminal and a storage robot basic information data collection terminal. The cargo basic information data collection terminal is used to collect the basic information data of the cargo, and the storage robot basic information data collection terminal is used to collect the basic information data of the storage robot;

[0063] Configure the cargo basic information data collection terminal and the storage robot basic information data collection terminal, and generate corresponding configuration channels. Through the generated corresponding configuration channels, connect to the server of the logistics center. The cargo basic information data collection terminal obtains the longitude and latitude of the initial position of the cargo and the longitude and latitude of the handling destination position of the cargo in the server of the logistics center through the corresponding configuration channel. The storage robot basic information data collection terminal obtains the current remaining power of the storage robot and the longitude and latitude of the position in the server of the logistics center through the corresponding configuration channel.

[0064] The process of obtaining the storage robot-cargo orientation map based on the basic information data of the storage robot and the basic information data of the cargo includes:

[0065] Establish a blank two-dimensional rectangular coordinate system, set the abscissa of the blank two-dimensional rectangular coordinate system as longitude, set the ordinate of the blank two-dimensional rectangular coordinate system as latitude, fill in the position of the storage robot into the blank two-dimensional rectangular coordinate system according to the longitude and latitude of the position of the storage robot to obtain the storage robot position mapping point, fill in the initial position of the cargo into the blank two-dimensional rectangular coordinate system according to the longitude and latitude of the initial position of the cargo to obtain the cargo initial position mapping point, fill in the handling destination position of the cargo into the blank two-dimensional rectangular coordinate system according to the longitude and latitude of the handling destination position of the cargo to obtain the cargo handling destination position mapping point, and connect the cargo initial position mapping point with the corresponding cargo handling destination position mapping point to obtain the storage robot-cargo orientation map;

[0066] The process of obtaining the straight-line distance and the handling distance according to the storage robot-cargo orientation map includes:

[0067] Obtain the straight-line distance L between the storage robot position mapping point and the cargo initial position mapping point according to the storage robot-cargo orientation map ik and the straight-line distance L between the cargo initial position mapping point and the cargo handling destination position mapping point kj ;

[0068] Among them, i is the index of the storage robot position mapping point, k is the index of the cargo initial position mapping point, L ik is the straight-line distance between the i-th storage robot position mapping point and the k-th cargo initial position mapping point, x i is the abscissa of the i-th storage robot position mapping point, x k is the abscissa of the k-th cargo initial position mapping point, y i is the ordinate of the i-th storage robot position mapping point, y k is the ordinate of the k-th cargo initial position mapping point;

[0069] Among them, j is the index of the cargo handling destination position mapping point, and j = k, L kj is the straight-line distance between the k-th cargo initial position mapping point and the j-th cargo handling destination position mapping point, x j is the abscissa of the j-th cargo handling destination position mapping point, y j is the ordinate of the j-th cargo handling destination position mapping point;

[0070] Add the straight-line distance between the mapped point of the warehousing robot's position and the mapped point of the initial position of the goods to the straight-line distance between the mapped point of the initial position of the goods and the mapped point of the destination position of the goods handling to obtain the handling distance L between the mapped point of the warehousing robot's position and the mapped point of the destination position of the goods handling ij ;

[0071] Among them, L ij = L ik + L kj ; L ij is the handling distance between the i-th mapped point of the warehousing robot's position and the j-th mapped point of the destination position of the goods handling;

[0072] The process of performing handling task allocation using the optimized ant colony algorithm based on the straight-line distance, handling distance, and basic information data of the warehousing robot includes:

[0073] Step S51: Set the initial pheromone τ ij (0) for each mapped point of the destination position of the goods handling;

[0074] It should be explained that τ ij (0) represents the initial pheromone of the j-th mapped point of the destination position of the goods handling when the warehousing robot corresponding to the i-th mapped point of the warehousing robot's position selects the goods corresponding to the j-th mapped point of the destination position of the goods handling. τ ij (0) is a constant value. The smaller the handling distance between the mapped point of the warehousing robot's position and the mapped point of the destination position of the goods handling, the larger the initial pheromone set for the mapped point of the destination position of the goods handling;

[0075] Step S52: Calculate the probability P that the warehousing robot corresponding to the i-th mapped point of the warehousing robot's position selects the goods corresponding to the j-th mapped point of the destination position of the goods handling according to the task selection formula ij ;

[0076] Among them, the task selection formula is: η ij is the reciprocal of the handling distance between the i-th mapped point of the warehousing robot's position and the j-th mapped point of the destination position of the goods handling, The larger the reciprocal of the handling distance between the storage robot position mapping point and the cargo handling destination position mapping point, the smaller the handling distance between the storage robot position mapping point and the cargo handling destination position mapping point, and the stronger the tendency of the storage robot to select the cargo for handling. m is the number of cargo handling destination position mapping points, and α is the influence factor of the initial pheromone, which determines the importance of the initial pheromone when the storage robot selects cargo for transportation. A larger α value means that the initial pheromone has a greater influence in the process of the storage robot selecting cargo for transportation. The value range of α is [0,1]. β is the influence factor of the reciprocal of the handling distance between the i-th storage robot position mapping point and the j-th cargo handling destination position mapping point, and the value range of β is [0,1];

[0077] Among them, R ij is the reinforcement learning reward factor when the storage robot corresponding to the i-th storage robot position mapping point selects the cargo corresponding to the j-th cargo handling destination position mapping point, which is used to further adjust the probability that the storage robot corresponding to the i-th storage robot position mapping point selects the cargo corresponding to the j-th cargo handling destination position mapping point. The reinforcement learning reward factor takes into account the battery power and path length of the storage robot when the storage robot selects cargo for handling, can enhance the intelligence and dynamics of cargo selection, encourage the selection of storage robots with more sufficient battery power and shorter paths for tasks, so as to optimize the scheduling efficiency, E max is the maximum battery power of the storage robot, and E i is the current remaining battery power of the storage robot corresponding to the i-th storage robot position mapping point. δ is the adjustment parameter of, and θ is the adjustment parameter of, and the value ranges of both δ and θ are [0,1];

[0078] Step S53: Select the storage robot with the highest probability to simulate handling the cargo, and obtain the battery power consumed by the storage robot after the simulation handling;

[0079] For example, there are two storage robots and two cargos. The probability that the first storage robot selects the first cargo for handling is 0.2, and the probability that the second storage robot selects the first cargo for handling is 0.3. Then select the second storage robot to handle the first cargo. The probability that the first storage robot selects the second cargo for handling is 0.5, and the probability that the second storage robot selects the second cargo for handling is 0.4. Then select the first storage robot to handle the second cargo;

[0080] Update the initial pheromone according to the battery power consumed by the storage robot after the simulation handling, and obtain the updated initial pheromone τ ij- (0);

[0081] Among them, τ ij -(0) = (1 - ρ ij ) × τ ij (0) + Δτ ij (0); ρ ij is the pheromone evaporation rate, indicating the decay rate of pheromone over time, used to prevent overfitting, avoid the algorithm falling into local optimal solutions, and let the old pheromone gradually disappear so that new high-quality task allocation schemes can dominate faster. ρ ij = ρ min + (ρ max - ρ min ) × e -γ×Iter ; ρ max is the maximum pheromone evaporation rate, that is, the upper limit of the pheromone evaporation rate, that is, the fastest speed at which pheromone can evaporate. ρ min is the minimum pheromone evaporation rate, that is, the lower limit of the pheromone evaporation rate, that is, the slowest speed at which pheromone can evaporate, and ρ max ∈ [0.7, 0.9], ρ min ∈ [0.1, 0.3], e is the base of the natural logarithm, γ is the adjustment parameter, and γ ∈ [0.01, 0.1]. Iter is the current update count, and Δτ ij (0) is the increased pheromone. Q is a constant used to control the overall amplitude of pheromone update. is the power consumption of the i-th warehousing robot after simulating the handling, and λ is the adjustment coefficient of, used to adjust the sensitivity of pheromone to the power consumption of the warehousing robot, and λ ∈ [0, 1];

[0082] Step S54: Repeat steps S52 and S53 until the set update count N is reached, obtain the probability that the i-th warehousing robot corresponding to the position mapping point selects the goods corresponding to the j-th goods handling destination position mapping point for handling after N updates. For each good, obtain the warehousing robot with the highest probability after N updates. One good represents one handling task, and allocate the goods to the corresponding warehousing robot with the highest probability after N updates;

[0083] It should be noted that during the process of repeating step S42, the initial pheromone in the task selection formula uses the updated initial pheromone;

[0084] It should be noted that in the traditional ant colony algorithm, the pheromone evaporation rate is static, which is prone to early convergence to the local optimal solution, resulting in unreasonable handling task allocation (such as selecting a warehousing robot farther from the initial position of the goods for handling), and only focusing on the path cost, which may cause the warehousing robot to run out of power during the handling process and be unable to complete the handling task. Therefore, it is necessary to improve the traditional ant colony algorithm. By dynamically adjusting the pheromone evaporation rate, the early convergence of the traditional ant colony algorithm to the local optimal solution is avoided, so that the handling task allocation is more reasonable. The reinforcement learning reward factor is introduced, so that the warehousing robot not only focuses on the path cost, but also can dynamically adjust the task allocation according to factors such as power and path length, thereby increasing the intelligence of task scheduling. After introducing the reinforcement learning reward factor, the selection of pheromone is more flexible and intelligent, and can dynamically respond to environmental changes, so that it can not only improve efficiency, but also better adapt to environmental changes, ensure the efficient operation of the robot cluster in a complex warehousing environment, reduce the transportation time, improve the handling efficiency, and improve the operation efficiency of the warehousing factory.

[0085] According to the handling task allocation result, the method for initially planning the path between the warehousing robot and the initial position of the goods and obtaining the initial path between the warehousing robot and the initial position of the goods includes:

[0086] Perform Cartesian grid division on the warehousing robot-goods orientation map, record the center of each grid as a node, divide the mapped point of the warehousing robot position, the mapped point of the initial position of the goods, and the mapped point of the goods handling destination position into the corresponding grids, and obtain the position of the obstacles during the handling process of the warehousing robot. Map the obstacles to the warehousing robot-goods orientation map after Cartesian grid division according to the position of the obstacles during the handling process of the warehousing robot;

[0087] The process of dividing the mapped point of the warehousing robot position, the mapped point of the initial position of the goods, and the mapped point of the goods handling destination position into the corresponding grids includes:

[0088] Obtain the node closest to the mapped point of the warehousing robot position, and divide the mapped point of the warehousing robot position into the grid corresponding to the node;

[0089] Obtain the node closest to the mapped point of the initial position of the goods, and divide the mapped point of the initial position of the goods into the grid corresponding to the node;

[0090] Obtain the node closest to the mapped point of the goods handling destination position, and divide the mapped point of the goods handling destination position into the grid corresponding to the node;

[0091] Denote the grid where the position mapping point of the warehousing robot is located as the initial grid, and denote the grid where the initial position mapping point of the goods is located as the target grid. Obtain the eight adjacent grids of the initial grid, and calculate the path cost D for moving from the initial grid to each adjacent grid;

[0092] Among them, D = U1 + U2; U1 is the straight-line distance between the node of the initial grid and the node of the adjacent grid, and U2 is the straight-line distance between the node of the adjacent grid and the node of the target grid. If there is an obstacle in the adjacent grid, the path cost D for moving from the initial grid to this adjacent grid is D = +∞;

[0093] Denote the adjacent grid with the minimum path cost as the first grid, use the first grid as the moving destination of the initial grid, obtain the eight adjacent grids of the first grid, calculate the path cost for moving from the first grid to each adjacent grid, and obtain the minimum value among the path costs for moving from the first grid to each adjacent grid. Denote the adjacent grid corresponding to the minimum value as the second grid, use the second grid as the moving destination of the first grid, and so on, repeating in a cycle until moving to the target grid to obtain the preliminary path between the warehousing robot and the initial position of the goods.

[0094] The process of optimizing the preliminary path between the warehousing robot and the initial position of the goods to obtain the optimized path between the warehousing robot and the initial position of the goods includes:

[0095] During the process of moving from the initial grid to the first grid, if the first grid is directly above, directly below, directly to the left, or directly to the right of the initial grid, then the process of moving from the initial grid to the first grid is recorded as a translation. If the first grid is in the upper left diagonal, lower left diagonal, upper right diagonal, or lower right diagonal of the initial grid, then the process of moving from the initial grid to the first grid is recorded as a diagonal shift. During the process of moving from the initial grid to the target grid, if there are two consecutive translations, then obtain the grid before the first translation and the grid after the second translation. If the grid before the first translation reaches the grid after the second translation directly through a diagonal shift, then optimize the path between the grid before the first translation and the grid after the second translation, delete the two translations, and directly perform a diagonal shift to obtain the optimized path between the warehousing robot and the initial position of the goods;

[0096] For example, if the first grid is directly to the right of the initial grid and the second grid is directly below the first grid, during the process of moving from the initial grid to the target grid, there are two consecutive translations, and the initial grid can reach the second grid directly through a diagonal shift, then delete the two translations and directly perform a diagonal shift, that is, directly reach the second grid from the initial grid through a diagonal shift without passing through the first grid, thereby greatly reducing the path length and increasing the rate of goods handling;

[0097] It should be noted that during the process of goods handling, there may be a situation where two consecutive translations are performed when only one skew translation is required. Therefore, it is necessary to optimize the preliminary path between the warehousing robot and the initial position of the goods, delete the two consecutive translations, and directly perform one skew translation, thereby reducing the length of the preliminary path, reducing the transportation time, improving the handling efficiency, and enhancing the operating efficiency of the warehousing factory;

[0098] The process of obtaining the optimized path between the initial position of the goods and the destination position of the goods handling by performing preliminary planning and path optimization on the path between the initial position of the goods and the destination position of the goods handling includes:

[0099] Based on the preliminary planning and path optimization of the path between the warehousing robot and the initial position of the goods, perform preliminary planning and path optimization on the path between the initial position of the goods and the destination position of the goods handling, obtain the optimized path between the initial position of the goods and the destination position of the goods handling, and control the warehousing robot to handle the goods along the optimized path between the warehousing robot and the initial position of the goods and the optimized path between the initial position of the goods and the destination position of the goods handling.[[ID=,7]]

[0100] In this embodiment, a warehousing robot-goods orientation map is obtained according to the basic information data of the warehousing robot and the basic information data of the goods, which clearly shows the position information of the warehousing robot and the goods, providing a solid foundation for subsequent handling task allocation; the straight-line distance and handling distance are obtained according to the warehousing robot-goods orientation map, and the handling task is allocated according to the straight-line distance, handling distance, and the basic information data of the warehousing robot, ensuring that the warehousing robot does not travel a long distance to pick up goods, while mobilizing the warehousing robot with more power to perform tasks, without causing overloading of the warehousing robot, and also ensuring that the warehousing robot has enough power to complete the handling task, thereby reducing the handling time and improving the handling efficiency; according to the handling task allocation result, perform preliminary planning on the path between the warehousing robot and the initial position of the goods to obtain the preliminary path between the warehousing robot and the initial position of the goods, perform path optimization on the preliminary path between the warehousing robot and the initial position of the goods to obtain the optimized path between the warehousing robot and the initial position of the goods, perform preliminary planning and path optimization on the path between the initial position of the goods and the destination position of the goods handling to obtain the optimized path between the initial position of the goods and the destination position of the goods handling, control the warehousing robot to handle the goods along the optimized path between the warehousing robot and the initial position of the goods and the optimized path between the initial position of the goods and the destination position of the goods handling, optimize the preliminary path, and ensure that the warehousing robot handles the goods along the shortest path, thereby reducing the handling time, improving the handling efficiency, and enhancing the operating efficiency of the warehousing factory.

[0101] Embodiment 2

[0102] Please refer to Figure 2 As shown, for the parts not described in detail in this embodiment, refer to the description of Embodiment 1. A control system for a cluster warehousing robot system based on artificial intelligence is provided, including:

[0103] A data acquisition module, configured to acquire the basic information data of the warehousing robot and the basic information data of the goods;

[0104] A task assignment module, configured to obtain a warehousing robot - goods orientation map according to the basic information data of the warehousing robot and the basic information data of the goods, obtain the straight-line distance and the handling distance according to the warehousing robot - goods orientation map, optimize the ant colony algorithm by dynamically adjusting the pheromone evaporation rate and introducing a reinforcement learning reward factor, and perform handling task assignment according to the straight-line distance, the handling distance and the basic information data of the warehousing robot by using the optimized ant colony algorithm;

[0105] A path planning module, configured to preliminarily plan the path between the warehousing robot and the initial position of the goods according to the handling task assignment result, and obtain the preliminary path between the warehousing robot and the initial position of the goods;

[0106] An optimized handling module, configured to optimize the path between the warehousing robot and the initial position of the goods to obtain the optimized path between the warehousing robot and the initial position of the goods, preliminarily plan and optimize the path between the initial position of the goods and the handling destination position of the goods to obtain the optimized path between the initial position of the goods and the handling destination position of the goods, and control the warehousing robot to carry the goods along the optimized path between the warehousing robot and the initial position of the goods and the optimized path between the initial position of the goods and the handling destination position of the goods.

[0107] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in the present invention can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0108] In several embodiments provided by the present invention, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only one type, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of the devices or units can be in electrical, mechanical, or other forms.

[0109] As described above, this is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered by the protection scope of the present invention.

[0110] Finally: The above description is only the preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A control method for a cluster storage robot system based on artificial intelligence, characterized in that, The control method of a cluster warehousing robot system based on artificial intelligence includes: Step S1: Collect the basic information data of the warehousing robot and the basic information data of the goods. Step S2: Obtain a warehousing robot-goods orientation map based on the basic information data of the warehousing robot and the basic information data of the goods. Obtain the straight-line distance and the handling distance according to the warehousing robot-goods orientation map. Optimize the ant colony algorithm by dynamically adjusting the pheromone evaporation rate and introducing a reinforcement learning reward factor. Perform handling task allocation using the optimized ant colony algorithm according to the straight-line distance, the handling distance, and the basic information data of the warehousing robot. Step S3: According to the handling task allocation result, preliminarily plan the path between the warehousing robot and the initial position of the goods to obtain the preliminary path between the warehousing robot and the initial position of the goods. Step S4: Optimize the preliminary path between the warehousing robot and the initial position of the goods to obtain the optimized path between the warehousing robot and the initial position of the goods. Preliminarily plan and optimize the path between the initial position of the goods and the handling destination position of the goods to obtain the optimized path between the initial position of the goods and the handling destination position of the goods. Control the warehousing robot to carry the goods along the optimized path between the warehousing robot and the initial position of the goods and the optimized path between the initial position of the goods and the handling destination position of the goods.

2. The control method of a cluster warehousing robot system based on artificial intelligence according to claim 1, characterized in that The basic information data of the warehousing robot includes the current remaining power of the warehousing robot and the longitude and latitude of the position. The basic information data of the goods includes the longitude and latitude of the initial position of the goods and the longitude and latitude of the handling destination position of the goods.

3. A control method for a cluster warehousing robot system based on artificial intelligence according to claim 2, characterized in that, The method for obtaining a warehousing robot-goods orientation map according to the basic information data of the warehousing robot and the basic information data of the goods includes: Establish a blank two-dimensional rectangular coordinate system, set the abscissa of the blank two-dimensional rectangular coordinate system as the longitude, and set the ordinate of the blank two-dimensional rectangular coordinate system as the latitude. Fill in the position of the warehousing robot in the blank two-dimensional rectangular coordinate system according to the longitude and latitude of the position of the warehousing robot to obtain the position mapping point of the warehousing robot. Fill in the initial position of the goods in the blank two-dimensional rectangular coordinate system according to the longitude and latitude of the initial position of the goods to obtain the initial position mapping point of the goods. Fill in the handling destination position of the goods in the blank two-dimensional rectangular coordinate system according to the longitude and latitude of the handling destination position of the goods to obtain the handling destination position mapping point of the goods. Connect the initial position mapping point of the goods with the corresponding handling destination position mapping point of the goods to obtain a warehousing robot-goods orientation map.

4. A control method for a cluster warehousing robot system based on artificial intelligence according to claim 3, characterized in that, The method for obtaining the straight-line distance and the handling distance according to the warehousing robot-goods orientation map includes: Obtain the straight-line distance between the position mapping point of the warehousing robot and the initial position mapping point of the goods and the straight-line distance between the initial position mapping point of the goods and the handling destination position mapping point of the goods according to the warehousing robot-goods orientation map. Add the straight-line distance between the mapped position point of the warehousing robot and the mapped position point of the initial position of the goods to the straight-line distance between the mapped position point of the initial position of the goods and the mapped position point of the handling destination position of the goods to obtain the handling distance between the mapped position point of the warehousing robot and the mapped position point of the handling destination position of the goods.

5. A control method for a cluster warehousing robot system based on artificial intelligence according to claim 4, characterized in that, The method for allocating handling tasks using the optimized ant colony algorithm based on the straight-line distance, handling distance, and basic information data of the warehousing robot includes: Step S51: Set the initial pheromone τ ij for each mapping point of the cargo handling destination location; ij (0); where \(i\) is the index of the location mapping point of the warehousing robot, \(j\) is the index of the location mapping point of the goods handling destination, and \(\tau\) ij (0) represents the initial pheromone of the \(j\)-th goods handling destination location mapping point when the warehousing robot corresponding to the \(i\)-th warehousing robot location mapping point selects the goods corresponding to the \(j\)-th goods handling destination location mapping point for handling; Step S52: Calculate the probability P that the i-th warehousing robot corresponding to the position mapping point of the warehousing robot selects the goods corresponding to the j-th goods handling destination position mapping point for handling according to the task selection formula ij ; Among them, the task selection formula is: η ij is the reciprocal of the handling distance between the i-th storage robot position mapping point and the j-th cargo handling destination position mapping point, m is the number of cargo handling destination position mapping points, α is the influence factor of the initial pheromone, the value range of α is [0, 1], β is the influence factor of the reciprocal of the handling distance between the i-th storage robot position mapping point and the j-th cargo handling destination position mapping point, the value range of β is [0, 1], R ij is the reinforcement learning reward factor when the storage robot corresponding to the i-th storage robot position mapping point selects the cargo corresponding to the j-th cargo handling destination position mapping point for handling; Step S53: Select the warehousing robot with the highest probability to simulate handling the goods, and obtain the power consumption of the warehousing robot after the simulated handling. Update the initial pheromone according to the power consumed by the warehousing robot after simulated handling to obtain the updated initial pheromone τ ij - (0); Among them, τ ij - (0) = (1 - ρ ij ) × τ ij (0) + Δτ ij (0); ρ ij is the pheromone evaporation rate, ρ ij = ρ min + (ρ max - ρ min ) × e -γ×Iter ; ρ max is the maximum information evaporation rate, ρ min is the minimum information evaporation rate, e is the base of the natural logarithm, γ is the adjustment parameter, Iter is the current update count, Δτ ij (0) is the increased pheromone, Q is a constant, is the power consumption of the i-th warehousing robot after simulating the handling, λ is the adjustment coefficient, L ij is the handling distance between the position mapping point of the i-th warehousing robot and the position mapping point of the j-th cargo handling destination; Step S54: Repeat steps S52 and S53 until the set number of updates N is reached, and obtain the probability that the i-th warehousing robot corresponding to the mapped position point after N updates selects the goods corresponding to the j-th handling destination position point of the goods for handling. For each good, obtain the warehousing robot with the highest probability after N updates. One good represents one handling task, and allocate the goods to the warehousing robot with the highest probability after N updates accordingly.

6. A control method for a cluster warehousing robot system based on artificial intelligence according to claim 5, characterized in that, The reinforcement learning reward factor consists of the maximum power of the warehousing robot, the current remaining power, and the straight-line distance. Among them, k is the index of the mapping point of the initial position of the goods, and L ik is the straight-line distance between the mapping point of the position of the i-th warehousing robot and the mapping point of the initial position of the k-th goods, and L kj is the straight-line distance between the mapping point of the initial position of the k-th goods and the mapping point of the handling destination position of the j-th goods, and E max is the maximum power of the warehousing robot, and E i is the current remaining power of the warehousing robot corresponding to the mapping point of the position of the i-th warehousing robot. δ is the adjustment parameter of, and θ is the adjustment parameter of, and the value ranges of both δ and θ are [0, 1].

7. A control method for a cluster warehousing robot system based on artificial intelligence according to claim 6, characterized in that, The method for initially planning the path between the warehousing robot and the initial position of the goods according to the handling task allocation result to obtain the initial path between the warehousing robot and the initial position of the goods includes: Perform Cartesian grid division on the warehousing robot-goods orientation map, mark the center of each grid as a node, divide the mapped position point of the warehousing robot, the mapped position point of the initial position of the goods, and the mapped position point of the handling destination position of the goods into the corresponding grids, and obtain the positions of the obstacles during the handling process of the warehousing robot. Map the obstacles to the warehousing robot-goods orientation map after Cartesian grid division according to the positions of the obstacles during the handling process of the warehousing robot. Mark the grid where the mapped position point of the warehousing robot is located as the initial grid, mark the grid where the mapped position point of the initial position of the goods is located as the target grid, obtain the eight adjacent grids of the initial grid, and calculate the path cost D for moving from the initial grid to each adjacent grid. Among them, D = U1 + U2; U1 is the straight-line distance between the node of the initial grid and the node of the adjacent grid, and U2 is the straight-line distance between the node of the adjacent grid and the node of the target grid. If there is an obstacle in the adjacent grid, the path cost D for moving from the initial grid to this adjacent grid is +∞. Mark the adjacent grid with the minimum path cost as the first grid, use the first grid as the moving destination of the initial grid, obtain the eight adjacent grids of the first grid, calculate the path cost for moving from the first grid to each adjacent grid, and obtain the minimum value among the path costs for moving from the first grid to each adjacent grid. Mark the adjacent grid corresponding to the minimum value as the second grid, use the second grid as the moving destination of the first grid, and so on, repeating in a cycle until moving to the target grid to obtain the initial path between the warehousing robot and the initial position of the goods.

8. A control method for a cluster warehousing robot system based on artificial intelligence according to claim 7, characterized in that, The method of dividing the mapped position points of the warehousing robot, the initial position mapped points of the goods, and the mapped destination position points of the goods into corresponding grids includes: Obtain the node closest to the mapped position point of the warehousing robot, and divide the mapped position point of the warehousing robot into the grid corresponding to this node; Obtain the node closest to the initial position mapped point of the goods, and divide the initial position mapped point of the goods into the grid corresponding to this node; Obtain the node closest to the mapped destination position point of the goods, and divide the mapped destination position point of the goods into the grid corresponding to this node.

9. A control method for a cluster warehousing robot system based on artificial intelligence according to claim 8, characterized in that, The method of optimizing the preliminary path between the warehousing robot and the initial position of the goods to obtain the optimized path between the warehousing robot and the initial position of the goods includes: During the process of moving from the initial grid to the first grid, if the first grid is directly above, directly below, directly to the left, or directly to the right of the initial grid, the process of moving the initial grid to the first grid is recorded as translation. If the first grid is in the upper left diagonal, lower left diagonal, upper right diagonal, or lower right diagonal of the initial grid, the process of moving the initial grid to the first grid is recorded as diagonal movement; During the process of moving from the initial grid to the target grid, if there are two consecutive translations, obtain the grid before the first translation and the grid after the second translation. If the grid before the first translation directly reaches the grid after the second translation through one diagonal movement, optimize the path between the grid before the first translation and the grid after the second translation, delete the two translations, and directly perform one diagonal movement to obtain the optimized path between the warehousing robot and the initial position of the goods.

10. A control method for a cluster warehousing robot system based on artificial intelligence according to claim 9, characterized in that, The method of preliminarily planning and optimizing the path between the initial position of the goods and the destination position of the goods to obtain the optimized path between the initial position of the goods and the destination position of the goods includes: Based on the preliminary planning and path optimization of the path between the warehousing robot and the initial position of the goods, preliminarily plan and optimize the path between the initial position of the goods and the destination position of the goods to obtain the optimized path between the initial position of the goods and the destination position of the goods.

Images