A logistics robot path planning method based on an intelligent storage environment

Abstract

The application relates to a logistics robot path planning method based on an intelligent storage environment and belongs to the field of intelligent storage logistics robot path planning. The method comprises the following steps: constructing an intelligent storage environment model; based on an ant colony algorithm, constructing a heuristic function which comprehensively considers path Manhattan distance and turning cost, and adopting an adaptive pheromone volatilization updating strategy to update pheromone concentration on the path, a transfer strategy to fuse a random selection strategy and a deterministic selection strategy, and finally obtaining an improved adaptive ant colony algorithm; based on the improved adaptive ant colony algorithm, performing logistics robot path planning, and outputting an optimized logistics robot path. The application can realize logistics robot path planning in an intelligent storage environment.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent warehousing logistics robot path planning, and specifically relates to a logistics robot path planning method based on an intelligent warehousing environment. Background Art

[0002] In intelligent warehouse scheduling, most operations rely on robotic "goods-to-person" picking. Path planning for logistics robots is a key factor limiting the overall operational efficiency of warehouse systems. Robotic path planning involves planning an optimal, collision-free path from a given starting point to a destination in an obstacle-filled work environment. Commonly used robotic path planning methods include the artificial potential field method, the A* algorithm, neural networks, genetic algorithms, and particle swarm algorithms. Ant colony algorithms, with their positive feedback and parallel computing capabilities, are often used to solve robotic path planning problems. However, ant colony algorithms are prone to localized search problems, path smoothness, and algorithmic instability. To address these issues, research both domestically and internationally has focused on improving the search capabilities and quality of optimal solutions through approaches such as heuristic information functions, search strategies, and pheromone initialization and update rules.

[0003] Regarding the application of ant colony algorithms in intelligent warehousing and logistics robot path planning, Gu Yong et al. proposed a warehouse logistics robot path planning method based on an improved ant colony algorithm. This method dynamically adapts and updates pheromones based on the number of iterations, achieving a balance between rapid convergence and escaping local optima. Ge Zhiyuan et al. proposed research on AGV path planning using an improved ant colony algorithm. They utilized a maximum-minimum ant system to constrain pheromones along the path and resolved deadlock issues in AGV path planning by executing regression actions. Song Yu et al. proposed a logistics robot path planning method based on an improved ant colony algorithm. This method used the A* algorithm to generate suboptimal paths and linearly initialize pheromones. The pheromones were then updated based on path length and the number of turns, resulting in a smoother path. Chen Yun et al. proposed a path planning method for intelligent warehousing robots based on an improved ant colony algorithm. This method considers the emergent factors of flexible intelligent warehousing environments, incorporates distance, turns, and safety factors into the heuristic function, and optimizes the pheromone update method. However, the map environments used in the validation of these techniques tend to be random. Smart warehouse environments are filled with shelves (arranged in double rows, back-to-back, neatly arranged and relatively regular). Logistics robots have limited available space (typically a single aisle) and a high probability of obstacles. When planning paths, the search space is smaller than in a random obstacle map environment. Therefore, path planning is more complex than in a typical random map. Summary of the Invention

[0004] The purpose of the present invention is to provide a logistics robot path planning method based on an intelligent warehousing environment, so as to realize the path planning of the logistics robot in the intelligent warehousing environment.

[0005] To achieve the above objectives, the technical solution of the present invention is: a logistics robot path planning method based on an intelligent warehousing environment, comprising:

[0006] Build an intelligent warehousing environment model;

[0007] Based on the ant colony algorithm, a heuristic function that comprehensively considers the Manhattan distance and turn cost of the path is constructed. An adaptive pheromone volatilization update strategy is used to update the pheromone concentration on the path. The transfer strategy integrates the random selection strategy and the deterministic selection strategy, ultimately obtaining an improved adaptive ant colony algorithm.

[0008] Logistics robot path planning is performed based on the improved adaptive ant colony algorithm, and the optimized logistics robot path is output.

[0009] Furthermore, the intelligent warehousing environment model is constructed as follows: with the logistics system KIVA as a reference, an intelligent warehousing environment model based on the grid method is constructed. The intelligent warehousing environment model includes picking workbenches, shelves, charging areas, and logistics robots. The shelves are arranged in double rows "back to back", neatly arranged, and single-aisle.

[0010] Furthermore, the heuristic function that comprehensively considers the path Manhattan distance and turn cost is expressed as follows:

[0011]

[0012] Where: is the weight of the path Manhattan distance; is the weight of the turning cost; d jE A is the Manhattan distance from node j to the target location E; ij The turning cost function required to select node j for the logistics robot; x is the horizontal coordinate of node j; y is the vertical coordinate of node j; E x is the horizontal coordinate of the target position E; E y is the ordinate of the target position E; log2( ) is the logarithmic function; Ang ij is the turning angle required for the logistics robot to reach node j from node (i-1) via node i.

[0013] Furthermore, the adaptive pheromone volatilization update strategy is an adaptive pheromone volatilization update strategy based on an inverse hyperbolic secant function model.

[0014] Furthermore, the adaptive pheromone volatilization update strategy based on the inverse hyperbolic secant function model is expressed as follows:

[0015]

[0016] In the formula, C is a constant, K is the total number of iterations, k represents the kth iteration, () is the secant function, and ρ represents the pheromone volatility coefficient.

[0017] Furthermore, the transfer strategy integrates the random selection strategy and the deterministic selection strategy that follows the Gaussian distribution. That is, when the ant selects the next node, if the random variable is less than or equal to the deterministic variable, the node with the highest probability is selected; otherwise, a node is randomly selected. This is expressed as follows:

[0018]

[0019] Where j is the next node to be visited, p ij m is the transition probability, q is a random variable uniformly distributed in the interval [0,1]; q0 is a variable that obeys the Gaussian distribution, 0≤q0≤1, and J is a random variable generated according to the probability distribution.

[0020] Furthermore, q0 is defined as follows:

[0021]

[0022] Where, represents a deterministic selection function that follows a Gaussian distribution, k represents the kth iteration, μ is the expected value, σ is the standard deviation, and C1 and C0 are constants.

[0023] Furthermore, the specific implementation steps of this method are as follows:

[0024] Step S1: constructing an intelligent storage environment model based on the grid method;

[0025] Step S2: Initialize the parameters of the improved adaptive ant colony algorithm;

[0026] Step S3, initialization state, setting the starting position and target position of the logistics robot (i.e., the ant), initializing the path matrix, initializing the heuristic information matrix according to the heuristic function, and initializing the initial pheromone concentration of all positions;

[0027] Step S4: Calculate the probability of going to each reachable node based on the pheromone concentration of each node, and select the next node based on the fused random selection strategy and deterministic selection strategy;

[0028] Step S5: Update the ant status;

[0029] Step S6: Repeat steps S4 and S5 until the ant reaches the target location or has no way to go;

[0030] Step S7: Repeat steps S4 to S6 until all ant searches in this iteration are completed.

[0031] Step S8: Adopting the adaptive pheromone volatilization update strategy to update the pheromone concentration on the path;

[0032] Step S9: Repeat steps S4 to S8 until the last iteration is completed, and output the optimized path.

[0033] The present invention also provides a computer-readable storage medium on which computer program instructions that can be executed by a processor are stored. When the processor executes the computer program instructions, the steps of any of the above methods can be implemented.

[0034] Compared with the existing technology, the present invention has the following beneficial effects: the method of the present invention comprehensively considers the heuristic function of the path Manhattan distance and the turning cost, the adaptive pheromone volatilization update strategy based on the inverse hyperbolic secant function model, and the fusion of the random selection strategy and the deterministic selection strategy obeying the Gaussian distribution, and finally obtains an improved adaptive ant colony algorithm. Compared with the existing path planning algorithm, it has obvious advantages in shortening the optimal path length, reducing the cumulative turning angle, and accelerating the convergence speed. It has strong robustness, thereby making the path planned for the logistics robot in the intelligent warehousing environment smoother and the operation result more stable. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 This is a flow chart of a logistics robot path planning method based on an intelligent warehousing environment in the present invention.

[0036] Figure 2 It is a smart warehousing environment model.

[0037] Figure 3 It is the adaptive volatility trend diagram of the inverse hyperbolic secant function.

[0038] Figure 4 The adaptive deterministic selection function q0 follows a Gaussian distribution.

[0039] Figure 5 The following are the results of basic ant colony algorithm path planning under three map environments.

[0040] Figure 6 The following are the convergence curves of the basic ant colony algorithm under three environments.

[0041] Figure 7 These are the path planning results of the three algorithms in a 20*20 environment.

[0042] Figure 8 The following are the convergence curves of the three algorithms in a 20*20 environment.

[0043] Figure 9 These are the path planning results of three algorithms in a 30*30 environment.

[0044] Figure 10 The following are the convergence curves of the three algorithms in a 30*30 environment. DETAILED DESCRIPTION

[0045] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings.

[0046] The present invention provides a logistics robot path planning method based on an intelligent warehousing environment, comprising:

[0047] Build an intelligent warehousing environment model;

[0048] Based on the ant colony algorithm, a heuristic function that comprehensively considers the Manhattan distance and turn cost of the path is constructed. An adaptive pheromone volatilization update strategy is used to update the pheromone concentration on the path. The transfer strategy integrates a random selection strategy and a deterministic selection strategy. That is, when the ant selects the next node, if the random variable is less than or equal to the deterministic variable, the node with the highest probability is selected; otherwise, a node is randomly selected. Ultimately, an improved adaptive ant colony algorithm is obtained.

[0049] Logistics robot path planning is performed based on the improved adaptive ant colony algorithm, and the optimized logistics robot path is output.

[0050] The following is a specific implementation process of the present invention.

[0051] 1. Intelligent warehousing environment model

[0052] Smart warehousing systems are generally arranged according to certain rules, primarily consisting of picking workstations, shelves, and a certain number of logistics robots. Shelves are arranged in double rows, back-to-back, and typically have a single aisle. The available space is limited, and the system can be reconfigured based on actual needs. To ensure the feasibility and universality of the research, the following simplifications and assumptions were made when modeling the smart warehousing environment:

[0053] (1) Add a safety distance to the actual boundary of obstacles such as picking workbenches and shelves;

[0054] (2) The logistics robot is regarded as a point mass and can move arbitrarily;

[0055] (3) During the path planning stage, the warehouse environment is static, and idle logistics robots are dispatched nearby.

[0056] like Figure 2 As shown, the present invention takes the classic logistics system - KIVA system as a reference to construct an intelligent warehousing environment model based on the grid method, including picking workbenches, shelves, charging areas, logistics robots, etc.

[0057] 2. Improved adaptive ant colony algorithm

[0058] 2.1. Heuristic function considering path Manhattan distance and turn cost

[0059] To accelerate algorithm convergence, the estimated cost function of the A* algorithm is borrowed, using the Manhattan distance from the current node to the target location as a path length heuristic. Furthermore, to improve the smoothness of the logistics robot's movement and minimize unnecessary turns, a turning angle cost is introduced based on the path length heuristic, and weights are assigned to path length and turning cost. This allows the ant to consider not only the path length but also the turning angle when selecting the next node. If the path is relatively short, the ant will choose a node with a smaller turning angle.

[0060]

[0061] Where: is the weight of the path Manhattan distance; is the weight of the turning cost; d jE A is the Manhattan distance from node j to the target location E; ij The turning cost function required to select node j for the logistics robot; x is the horizontal coordinate of node j; y is the vertical coordinate of node j; E x is the horizontal coordinate of the target position E; E y is the ordinate of the target position E; log2( ) is the logarithmic function; Ang ij is the turning angle required for the logistics robot to reach node j from node (i-1) via node i. On the grid map of the intelligent warehouse environment, the Manhattan distance between adjacent nodes is 1 or 2 units, while the turning angle can be 0°, 45°, 90°, or 135°. Once a turn occurs, the angle is significantly greater than the path length. Therefore, to balance path length and turning angle, the turning cost is calculated as the logarithm of the turning angle, resulting in a nonlinear positive correlation between the turning angle and the turning cost.

[0062] 2.2 Adaptive pheromone volatilization strategy

[0063] In the basic ant colony algorithm, the pheromone volatility coefficient ρ is generally a constant. In order to improve the global search capability of the algorithm, an adaptive pheromone volatility update strategy based on the inverse hyperbolic secant function model is proposed, namely

[0064]

[0065] In the formula, C is a constant, K is the total number of iterations, k represents the kth iteration, () is the secant function, and ρ represents the pheromone volatility coefficient.

[0066] like Figure 3As shown, the inverse hyperbolic secant function has the characteristic of decreasing rapidly in the early stages and slowly in the later stages. Pheromones can be adaptively updated based on the distribution of solutions, dynamically adjusting the pheromone concentration on each path. This ensures that ants are neither overly concentrated nor overly dispersed during the search process, avoiding premature convergence and localized convergence, and improving the algorithm's global search capability.

[0067] 2.3 Transfer strategy combines deterministic selection strategy and random selection strategy following Gaussian distribution

[0068] The basic ant colony algorithm uses a roulette wheel method to randomly select the next node during solution construction, which slows down the evolution and is prone to algorithm stagnation. Existing technologies primarily combine random selection and deterministic selection methods, dynamically adjusting the deterministic selection probability q0 during the search process. Existing technologies use piecewise functions to dynamically adjust the value of q0. However, the value of q0 in existing technologies often experiences step changes, potentially leading to significant perturbations in the solution space at a given moment. Furthermore, the piecewise function strategy relies on empirical experience to determine the segmentation method, segmentation intervals, and related numerical values. This requires continuous tuning of relevant parameters for different optimization problems, leaving room for improvement in applicability and flexibility.

[0069] The present invention proposes a transfer strategy that integrates a random selection strategy and a deterministic selection strategy, that is, when an ant selects the next node, if the random variable is less than or equal to the deterministic variable, the node with the highest probability is selected; otherwise, a node is randomly selected.

[0070]

[0071] Where j is the next node to be visited, p ij m is the transition probability, q is a random variable uniformly distributed in the interval [0,1]; q0 is a variable that obeys the Gaussian distribution, 0≤q0≤1, and J is a random variable generated according to the probability distribution.

[0072] Where q0 is defined as follows:

[0073]

[0074] Where, represents a deterministic selection function that follows a Gaussian distribution, k represents the number of iterations, μ is the expected value, σ is the standard deviation, and C1 and C0 are constants.

[0075] like Figure 4 As shown in the figure, the adaptive deterministic selection function obeying the Gaussian distribution can not only realize the rule of first getting bigger and then getting smaller, but also achieve smooth transition, simulating the actual search process of the ant colony, and better balancing the search ability and evolution speed.

[0076] In summary, the process of the logistics robot path planning method based on the intelligent warehousing environment of the present invention, that is, the intelligent warehousing robot path planning method based on the improved adaptive ant colony algorithm, is as follows: Figure 1 As shown, the following steps are included:

[0077] Step S1: Constructing an intelligent warehouse environment model based on the grid method, including picking workbenches, shelves, charging areas, logistics robots, etc.

[0078] Step S2, initializing the parameters of the improved adaptive ant colony algorithm, including the weights of the path length and the turning cost in the heuristic function, and the parameters related to the inverse hyperbolic secant function and the Gaussian distribution function;

[0079] Step S3, initialization state, setting the starting position and target position of the logistics robot (i.e., the ant), initializing the path matrix, initializing the heuristic information matrix according to the heuristic function, and initializing the initial pheromone concentration of all positions;

[0080] Step S4: Calculate the probability of going to each reachable node based on the pheromone concentration of each node, and select the next node based on the fused random selection strategy and deterministic selection strategy;

[0081] Step S5, update the ant status: update the path and its length, turning angle;

[0082] Step S6: Repeat steps S4 and S5 until the ant reaches the target location or has no way to go;

[0083] Step S7: Repeat steps S4 to S6 until all ant searches in this iteration are completed.

[0084] Step S8: Adopting the adaptive pheromone volatilization update strategy to update the pheromone concentration on the path;

[0085] Step S9: Repeat steps S4 to S8 until the last iteration is completed, and output the optimized path.

[0086] 3. Simulation experiment

[0087] To test the effectiveness of the improved adaptive ant colony algorithm (MAACO) in solving the path planning of logistics robots in intelligent warehousing environments, simulation verification was carried out on Matlab 2016a. The rationality and effectiveness of the improved adaptive ant colony algorithm in solving the path planning problem in intelligent warehousing environments were verified from the following three aspects: (1) Comparative analysis of the path planning of the basic ant colony algorithm (ACO) in different obstacle environments. The small random obstacle environment used in the robot path planning research based on the improved ant colony algorithm proposed by Wan Xiaofeng et al. and the large random obstacle environment used in the mobile robot path planning proposed by the adaptive ant colony algorithm proposed by Feng Shengfei et al. were selected for comparison with the intelligent warehousing environment; (2) Comparative experiments were conducted on the improved adaptive ant colony algorithm, the basic ant colony algorithm, and the elite ant colony algorithm (EACO) in a simple 20*20 warehousing environment; (3) Comparative experiments were conducted on the improved adaptive ant colony algorithm, the basic ant colony algorithm, and the elite ant colony algorithm in a complex 30*30 warehousing environment. All algorithms were run independently 20 times. The basic ant colony algorithm heuristic function is the inverse of the Euclidean distance from the current node to the target location. The parameters are set as follows: number of iterations K = 100, number of ants M = 50, ɑ = 1, β = 7, ρ = 0.3, and pheromone increase intensity coefficient Q = 1.

[0088] 3.1 Basic Ant Colony Algorithm Experiments in Different Environments

[0089] The basic ant colony algorithm path planning results in three environments: a small random obstacle map of 20*20 scale, a large random obstacle map, and an intelligent warehouse map (with obstacle rates of 22.75%, 28%, and 45% respectively) are as follows: Figure 5 As shown (where Figure 5 (a) is a small random obstacle map; Figure 5 Middle (b) is a large random obstacle map; Figure 5 (c) is the smart warehouse map).

[0090] The experimental performance parameters are shown in Table 1.

[0091] Table 1 Basic ant colony computing performance under three environments

[0092]

[0093] Table 1 shows that the basic ant colony algorithm (BACA) consistently found the optimal solution in both small and large random obstacle environments, with relatively stable optimal path lengths. However, the BACA only found the optimal solution seven times in the intelligent warehousing environment. Compared to the small and large random obstacle environments, the path planning results in the intelligent warehousing environment were less stable, primarily manifested in significant fluctuations in the optimal path length and cumulative turning angles. In particular, the standard deviation of the cumulative turning angles was significantly greater than in the random obstacle environment. Furthermore, the minimum convergence algebra values ​​of the BACA in both small and large random obstacle environments were larger than those in the intelligent warehousing environment, but the standard deviations were smaller.

[0094] The convergence curves of the basic ant colony algorithm under three environments are as follows: Figure 6 As shown. Figure 6 As can be seen, the basic ant colony algorithm's search space for path planning in a smart warehouse environment is significantly smaller than that in a random obstacle map environment. Preliminary analysis suggests this is due to the fact that smart warehouses are densely populated with shelves, have a high obstacle rate, and have limited feasible space. In summary, smart warehouses differ from random obstacle environments when planning paths for logistics robots.

[0095] 3.2 Experiment in 20*20 intelligent warehouse environment

[0096] Simulation environment parameter settings: intelligent warehouse scale 20*20, obstacle rate 45%, starting position S=184, target position E=79. Among them, the elite ant colony algorithm ζ=8, based on the experience of multiple control experiments, the adaptive ant colony algorithm weight is improved =0.9, =0.1.

[0097] The optimal path trajectory of basic ant colony algorithm, elite ant colony algorithm and improved adaptive ant colony algorithm is as follows Figure 7 As shown. Figure 7 It can be seen that the basic ant colony algorithm falls into the local optimum, while the elite ant colony algorithm and the improved adaptive ant colony algorithm find the global optimal solution.

[0098] The convergence curves of the three algorithms in the 20*20 environment are as follows Figure 8 As shown. Figure 8 It can be seen that in the 20*20 simple intelligent warehousing environment, the basic ant colony algorithm does not converge, while the elite ant colony algorithm and the improved adaptive ant colony algorithm can converge faster.

[0099] The performance parameters of the three algorithms are shown in Table 2.

[0100] Table 2 Performance of three algorithms in 20*20 environment

[0101]

[0102] As shown in Table 2, in a simple 20x20 intelligent warehouse environment, the average shortest path length of the elite ant colony algorithm and the improved adaptive ant colony algorithm is 13.5% shorter than that of the basic ant colony algorithm, and the cumulative turning angle is 17.4% lower. The improved adaptive ant colony algorithm uses an adaptive pheromone concentration update strategy based on the inverse hyperbolic secant model, which avoids premature algorithm maturation and can quickly find the global optimal solution. The convergence algebraic mean is 96% and 62.2% lower than that of the basic ant colony algorithm and the elite ant colony algorithm, respectively. Furthermore, the convergence algebraic standard deviation is small, indicating that the improved adaptive ant colony algorithm has relatively stable performance.

[0103] 3.3 Experiment in 30*30 intelligent warehouse environment

[0104] To verify the robustness of the improved adaptive ant colony algorithm for solving the logistics robot path planning problem in an intelligent warehouse environment, we further expanded the warehouse scale, added robot obstacles, and increased the environmental complexity. The simulation environment parameters were set as follows: intelligent warehouse scale 30*30, obstacle rate 46.4%, starting position S=544, target position E=839, and all other parameters remained unchanged.

[0105] In a 30*30 complex warehouse environment, the optimal paths of the three algorithms are as follows: Figure 9 As shown. Figure 9 It can be seen that the basic ant colony algorithm and the elite ant colony algorithm fall into local optimality, and the cumulative turning angle of the path is large. The improved adaptive ant colony algorithm adds the turning cost to the heuristic function and finds the global optimal path, which is smoother.

[0106] The convergence curves of the three algorithms in the 30*30 environment are as follows Figure 10 As shown. Figure 10 As can be seen, with the expansion of warehouse scale and increased environmental complexity, the basic ant colony algorithm searches more slowly, almost failing to converge when K = 100. The elite ant colony algorithm still converges quickly but struggles to escape the local optimal solution. The improved adaptive ant colony algorithm, on the other hand, exhibits a stable evolutionary direction and converges more quickly, quickly finding the global optimal solution.

[0107] In a 30*30 complex warehouse environment, the performance parameters of the three algorithms are shown in Table 3.

[0108] Table 3 Performance of three algorithms in 30*30 environment

[0109]

[0110] Table 3 shows that in complex intelligent warehousing environments, the search results of the basic ant colony algorithm and the elite ant colony algorithm are unstable, resulting in large optimal path lengths and cumulative turning angles. The improved adaptive ant colony algorithm provides stable path planning results, with average shortest path lengths reduced by 3.1% and 7% compared to the basic ant colony algorithm and the elite ant colony algorithm, respectively. The average cumulative turning angles are reduced by 24% and 27.7% compared to the basic ant colony algorithm and the elite ant colony algorithm, respectively. The improved adaptive ant colony algorithm, which combines a random selection strategy with a deterministic selection strategy based on a Gaussian distribution, achieves an average convergence number of 7.9, which is 91.9% and 55.2% less than the basic ant colony algorithm and the elite ant colony algorithm, respectively. Furthermore, the improved adaptive ant colony algorithm exhibits a small standard deviation and stable results, making it an advantage in logistics robot path planning in intelligent warehousing environments.

[0111] The present invention also provides a computer-readable storage medium on which computer program instructions that can be executed by a processor are stored. When the processor executes the computer program instructions, the steps of any of the above methods can be implemented.

[0112] The above are preferred embodiments of the present invention. Any changes made according to the technical solution of the present invention, as long as the resulting functions and effects do not exceed the scope of the technical solution of the present invention, shall fall within the scope of protection of the present invention.

Claims

1. A logistics robot path planning method based on an intelligent warehousing environment, characterized in that: include: Build an intelligent warehousing environment model; Based on the ant colony algorithm, a heuristic function that comprehensively considers the Manhattan distance and turn cost of the path is constructed. An adaptive pheromone volatilization update strategy is used to update the pheromone concentration on the path. The transfer strategy integrates the random selection strategy and the deterministic selection strategy, ultimately obtaining an improved adaptive ant colony algorithm. The adaptive pheromone volatilization update strategy is based on the inverse hyperbolic secant function model and is expressed as follows: In the formula, C is a constant, K is the total number of iterations, k represents the kth iteration, () is the secant function, ρ represents the pheromone volatility coefficient; The heuristic function that comprehensively considers the Manhattan distance of the path and the turn cost is expressed as follows: Where: is the weight of the path Manhattan distance; is the weight of the turning cost; d jE A is the Manhattan distance from node j to the target location E; ij The turning cost function required to select node j for the logistics robot; x is the horizontal coordinate of node j; y is the vertical coordinate of node j; E x is the horizontal coordinate of the target position E; E y is the ordinate of the target position E; log2( ) is the logarithmic function; Ang ij is the turning angle required for the logistics robot to go from node (i-1) to node j via node i; Logistics robot path planning is performed based on the improved adaptive ant colony algorithm, and the optimized logistics robot path is output.

2. A logistics robot path planning method based on an intelligent warehousing environment according to claim 1, characterized in that: The intelligent warehousing environment model is constructed as follows: taking the logistics system KIVA as a reference, a grid-based intelligent warehousing environment model is constructed. The intelligent warehousing environment model includes picking workbenches, shelves, charging areas, and logistics robots.

3. A logistics robot path planning method based on an intelligent warehousing environment according to claim 1, characterized in that: The transfer strategy combines a random selection strategy with a deterministic selection strategy that follows a Gaussian distribution. That is, when the ant selects the next node, if the random variable is less than or equal to the deterministic variable, the node with the highest probability is selected; otherwise, a node is randomly selected. This is expressed as follows: Where j is the next node to be visited, p ij m is the transition probability, q is a random variable uniformly distributed in the interval [0,1]; q0 is a variable that obeys the Gaussian distribution, 0≤q0≤1, and J is a random variable generated according to the probability distribution.

4. A logistics robot path planning method based on an intelligent warehousing environment according to claim 3, characterized in that: q0 is defined as follows: Where, represents a deterministic selection function that follows a Gaussian distribution, k represents the kth iteration, μ is the expected value, σ is the standard deviation, and C1 and C0 are constants.

5. A logistics robot path planning method based on an intelligent warehousing environment according to any one of claims 1 to 4, characterized in that: The specific implementation steps of this method are as follows: Step S1: constructing an intelligent storage environment model based on the grid method; Step S2: Initialize the parameters of the improved adaptive ant colony algorithm; Step S3, initialization state, setting the starting position and target position of the logistics robot (i.e., the ant), initializing the path matrix, initializing the heuristic information matrix according to the heuristic function, and initializing the initial pheromone concentration of all positions; Step S4: Calculate the probability of going to each reachable node based on the pheromone concentration of each node, and select the next node based on the fused random selection strategy and deterministic selection strategy; Step S5: Update the ant status; Step S6: Repeat steps S4 and S5 until the ant reaches the target location or has no way to go; Step S7: Repeat steps S4 to S6 until all ant searches in this iteration are completed. Step S8: Adopting the adaptive pheromone volatilization update strategy to update the pheromone concentration on the path; Step S9: Repeat steps S4 to S8 until the last iteration is completed, and output the optimized path.

6. A computer-readable storage medium storing computer program instructions that can be executed by a processor, wherein when the processor executes the computer program instructions, the steps of the method according to any one of claims 1 to 5 can be implemented.

Images