Intelligent inventory robot path optimization method based on environment modeling and collaborative optimization

Abstract

The invention provides an intelligent inventory robot path optimization method based on environment modeling and collaborative optimization, and relates to the technical field of intelligent warehouse management, and the method comprises the steps: scanning a warehouse region through an unmanned plane platform, and generating three-dimensional warehouse point cloud data; after point cloud denoising, plane fitting, shelf boundary extraction and channel detection processing, a two-dimensional storage plane graph is generated and transmitted to a robot data base; generating an initial inventory planning path by adopting an A * algorithm, optimizing the initial path through a simulated annealing algorithm, and generating a target inventory planning path; based on the dynamic obstacle influence factors, the RFID signal intensity and the robot residual electric quantity are combined, and a target path is corrected through a mixed algorithm of multi-target optimization and PSO dynamic weight; and the operation end checks the checking progress through the AR equipment and manually adjusts the target path, so that the checking path of the intelligent checking robot can be optimized in real time in a dynamic environment, and the checking efficiency, the coverage rate and the cruising ability of the robot are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent warehouse management, and particularly to a path optimization method for an intelligent inventory robot with environmental modeling and collaborative optimization. Background Art

[0002] In the rapid development process of modern warehouse management, intelligent inventory robots, with their efficient and accurate characteristics, are increasingly widely used and have become an important tool for improving warehouse operation efficiency and reducing labor costs. These robots greatly improve the intelligence level of warehouse management through automatic inspection, data collection and processing.

[0003] However, with the increasing complexity and dynamism of the warehouse environment, most traditional path planning methods for inventory robots rely on pre-constructed static environment models, which often ignore the dynamic change factors that may occur in the warehouse environment during construction. In practical applications, the warehouse environment is dynamically changing, and there may be dynamically changing obstacles such as temporarily stacked goods, moving forklifts or personnel, which will have a significant impact on the path planning results of the robot. In addition, the change of RFID signal strength will also affect the accurate identification of the location of goods by the robot, and the real-time power consumption of the robot is directly related to its operation sustainability and efficiency.

[0004] Facing these dynamic change factors, traditional path planning methods often cannot be quickly adjusted according to real-time environmental information, which may lead to the robot being unable to avoid emergencies in time, and may even fall into an infinite loop or be unable to complete the task. This not only reduces the efficiency of the inventory operation, but also poses a potential threat to warehouse safety.

[0005] Therefore, it is necessary to provide a path optimization method for an intelligent inventory robot with environmental modeling and collaborative optimization to solve the above technical problems. Summary of the Invention

[0006] To solve the above technical problems, the present invention provides a path optimization method for an intelligent inventory robot with environmental modeling and collaborative optimization, which is used to solve the problem that traditional path planning methods for inventory robots are often based on static environment models, difficult to adapt to dynamic change factors in the warehouse environment, and unable to achieve efficient path planning and collaborative optimization of intelligent inventory robots.

[0007] A path optimization method for an intelligent inventory robot with environmental modeling and collaborative optimization provided by the present invention, the path optimization method includes: Scanning a target warehouse area by a drone platform to generate three-dimensional warehouse point cloud data, and performing point cloud denoising, plane fitting, shelf boundary extraction and channel detection processing on the three-dimensional warehouse point cloud data to generate a two-dimensional warehouse floor plan and transmit it to the robot data base; The A* algorithm is used to generate an initial inventory planning path based on the two-dimensional warehouse floor plan, and the simulated annealing algorithm is used to optimize the initial inventory planning path to generate a target inventory planning path; Based on the collected dynamic obstacle influence factors, combined with the RFID signal strength and the remaining battery power of the robot, the target inventory planning path is corrected by a hybrid algorithm of multi-objective optimization and PSO dynamic weight; The operation terminal views the inventory progress of the robot through the AR device and manually adjusts the target inventory planning path.

[0008] Preferably, the target warehouse area is scanned by a drone platform to generate three-dimensional warehouse point cloud data, and the three-dimensional warehouse point cloud data is subjected to point cloud denoising, plane fitting, shelf boundary extraction and channel detection processing to generate a two-dimensional warehouse floor plan and transmit it to the robot data base, specifically including: The drone platform is equipped with a 16-line lidar and an IMU inertial navigation system, and overlapping scans are performed along the preset grid route of the target warehouse area at a preset cruising speed to generate raw warehouse point cloud data; The voxel grid filtering technology is used to divide the three-dimensional space corresponding to the raw warehouse point cloud data into voxels of 0.08m³ cubes, and through the formula Perform voxel indexing on each piece of the raw warehouse point cloud data, and retain the centroid points within each voxel to generate the three-dimensional warehouse point cloud data; Among them, represents the calculation point in the three-dimensional warehouse point cloud data of the voxel; respectively represent the side lengths of the voxel in the x, y, and z directions; respectively represent the voxel indices to which the calculation point p belongs in the x, y, and z directions; W represents the total width of the voxel grid in the x direction, that is, the number of voxels in the x direction; H represents the total width of the voxel grid in the y direction, that is, the number of voxels in the y direction.

[0009] Preferably, the process of shelf boundary extraction and channel detection processing is as follows: In the three-dimensional coordinate system corresponding to the three-dimensional warehouse point cloud data, for each plane, with the centroid of the plane as the origin, a local coordinate system orthogonal to the plane normal vector is constructed, and the three-dimensional warehouse point cloud data is projected onto the local coordinate system; Perform Delaunay triangulation on the projected point cloud, generate a closed shelf boundary polygon by screening the triangle sides with the smallest circumcircle radius, and perform boundary error control on the closed shelf boundary polygon through the Douglas-Peucker algorithm to generate a shelf boundary contour; Convert the ground point cloud into a bird's-eye view raster, fill the pixel gaps in the 3D warehouse point cloud data through morphological closing operation, and calculate the pixel occupancy probability ; Obtain the shelf passage discrimination probability and the preset width limit, and the pixel occupancy probability All pixel points with the pixel occupancy probability greater than the shelf passage discrimination probability form shelf elements, and the pixel occupancy probability All pixel points with the pixel occupancy probability less than the shelf passage discrimination probability form passage elements, and at the same time mark the entrance and exit elements, and the width between the pixel points corresponding to both ends of the entrance and exit elements is greater than the preset width limit; Attach a shelf ID, a passage width, and an entrance and exit direction to the shelf elements, the passage elements, and the entrance and exit elements respectively.

[0010] Preferably, use the A* algorithm to generate an initial inventory planning path based on the 2D warehouse floor plan, and optimize the initial inventory planning path through the simulated annealing algorithm to generate a target inventory planning path, specifically including: Determine the inventory shelf positions based on the 2D warehouse floor plan, obtain the types of goods assets and the inventory position priorities corresponding to the inventory shelf positions from the warehouse management database corresponding to the target warehouse area, and set the robot inventory movement parameters, namely the maximum walking speed, the minimum safety distance, and the single-shelf single-layer scanning time; Use the A* algorithm to search for the shortest path from the starting inventory position to the ending shelf position, and combine the types of goods assets and the inventory position priorities to generate the initial inventory planning path; Perform local optimization on the initial inventory planning path, that is, optimize the initial inventory planning path through the simulated annealing algorithm to generate the target inventory planning path.

[0011] Preferably, use the A* algorithm to search for the shortest path from the starting inventory position to the ending shelf position, and combine the types of goods assets and the inventory position priorities to generate the initial inventory planning path, specifically including: Set the node cost function as follows: Among them, f(n) represents the total estimated cost of the inventory shelf position node n; g(n) represents the actual movement cost from the starting inventory position node to the inventory shelf position node n; Represents the actual movement cost from the starting inventory position node to the previous inventory shelf position node Of the actual movement cost; move_cost(n) represents the basic movement cost from the previous inventory shelf position node To the inventory shelf position node n, if from the previous inventory shelf position node If there is a horizontal or vertical movement to the inventory shelf position node n, then move_cost(n) = 1. If there is a diagonal movement from the previous inventory shelf position node parent(n) to the inventory shelf position node n, then ; turn_penalty(n) represents the turning penalty of the inventory shelf position node n. If a turn is required from the previous inventory shelf position node parent(n) to the inventory shelf position node n, then , represents the turning penalty coefficient. Otherwise, turn_penalty(n) = 0; width_penalty(n) represents the aisle width penalty of the inventory shelf position node n, , represents the width penalty coefficient, represents the standard aisle width, w(n) represents the aisle width where the inventory shelf position node n is located; h(n) represents the heuristic estimated cost from the inventory shelf position node n to the termination inventory position node; represents the coordinates of the inventory shelf position node n; represents the coordinates of the termination inventory position node; Expand and select the inventory shelf position node with the minimum total estimated cost through the priority queue until all the inventory shelf position nodes are traversed, generate a continuous walking sequence including turning points and straight path segments, that is, the shortest path from the starting inventory position to the termination shelf position, and combine the goods asset type and the inventory position priority to generate the initial inventory planning path.

[0012] Preferably, locally optimize the initial inventory planning path, that is, optimize the initial inventory planning path through the simulated annealing algorithm to generate the target inventory planning path, which specifically includes: Set the key parameters of the simulated annealing algorithm, namely the initial temperature T, the cooling rate , the termination temperature and the number of perturbation times Q at each temperature; Define the path perturbation operations including randomly selecting two path nodes in the initial inventory planning path P, deleting the intermediate nodes of these two path nodes and reconnecting the initial inventory planning path P, randomly swapping the order of two non - adjacent path nodes in the initial inventory planning path P, and randomly selecting a sub - path in the initial inventory planning path P and reversing the direction of the sub - path; Implement the path perturbation operation on the initial inventory planning path P to generate a new inventory planning path , calculate the change in path length , where L represents the initial path length of the initial inventory planning path P, Indicates the new inventory count planning path of the new path length; If , then ; If , then calculate the path acceptance probability based on the Metropolis acceptance criterion . If the path acceptance probability q is greater than the random number r, , then , otherwise ; The initial temperature T converges step by step according to the cooling rate , and the path perturbation operation is performed Q times at each temperature until the temperature is less than the termination temperature , terminate the simulated annealing algorithm, and output the target inventory count planning path.

[0013] Preferably, based on the collected dynamic obstacle influence factor, combined with the RFID signal strength and the remaining battery power of the robot, the target inventory count planning path is corrected by a hybrid algorithm of multi-objective optimization and PSO dynamic weight, specifically including: Normalize the RFID signal strength and the remaining battery power of the robot, and construct a multi-objective optimization function based on the target path length of the target inventory count planning path, the normalized RFID signal strength and the remaining battery power of the robot, and the dynamic obstacle influence factor as follows: In the formula, F represents the comprehensive cost function; represents the target weight coefficient; represents the target path length of the target inventory count planning path; represents the normalized RFID signal strength; represents the normalized remaining battery power of the robot; O represents the dynamic obstacle influence factor, , represents the distance from the intelligent inventory count robot to the nearest obstacle, represents the safety distance threshold of the intelligent inventory count robot, represents the dynamic obstacle influence coefficient; Each particle represents a set of target weight vectors , and the particle velocity and position are updated iteratively by PSO to determine the optimal combination of the target weight coefficients ; Based on the optimal combination of the target weight coefficients , adjust the node cost function to obtain an improved node cost function, and correct the target inventory count planning path based on the improved node cost function. The expression of the improved node cost function is as follows: In the formula, represents the Manhattan distance from the inventory shelf position node n to the termination inventory position node; represents the signal heuristic; represents the power heuristic; represents the obstacle heuristic.

[0014] An intelligent inventory robot path optimization system for environmental modeling and collaborative optimization, the path optimization system includes: A drone scanning and data transmission module, which is used to scan the target warehouse area through a drone platform to generate three-dimensional warehouse point cloud data, perform point cloud denoising, plane fitting, shelf boundary extraction and channel detection processing on the three-dimensional warehouse point cloud data to generate a two-dimensional warehouse floor plan and transmit it to the robot data base; An inventory path generation and optimization module, which is used to generate an initial inventory planning path based on the two-dimensional warehouse floor plan by using the A* algorithm, and optimize the initial inventory planning path through the simulated annealing algorithm to generate a target inventory planning path; A multi-objective optimization and PSO dynamic weight correction module, which is used to correct the target inventory planning path through a hybrid algorithm of multi-objective optimization and PSO dynamic weight based on the collected dynamic obstacle influence factor, combined with the RFID signal strength and the remaining battery power of the robot; An inventory progress viewing and manual adjustment module, which is used for the operation terminal to view the robot inventory progress through the AR device and manually adjust the target inventory planning path.

[0015] An electronic device includes a memory and a processor. A computer program is stored in the memory. When the processor runs the computer program stored in the memory, the processor executes the steps of an intelligent inventory robot path optimization method as described in any one of the above.

[0016] A readable storage medium stores a computer program, and when the computer program is executed by a processor, it is used to implement the steps of an intelligent inventory robot path optimization method as described in any one of the above.

[0017] Compared with the related technology, an intelligent inventory robot path optimization method provided by the present invention has the following beneficial effects: The present invention scans a target storage area through a drone platform to generate three-dimensional storage point cloud data, performs point cloud denoising, plane fitting, shelf boundary extraction, and channel detection on the three-dimensional storage point cloud data to generate a two-dimensional storage floor plan and transmits it to the robot data base; uses the A* algorithm to generate an initial inventory planning path based on the two-dimensional storage floor plan, and optimizes the initial inventory planning path through the simulated annealing algorithm to generate a target inventory planning path; based on the collected dynamic obstacle influence factors, combined with the RFID signal strength and the remaining battery power of the robot, corrects the target inventory planning path through a hybrid algorithm of multi-objective optimization and PSO dynamic weight; the operation terminal views the inventory progress of the robot through an AR device and manually adjusts the target inventory planning path, so that the inventory path of the intelligent inventory robot can be optimized in real time in a dynamic environment, improving the inventory efficiency, coverage rate, and battery life of the robot.

[0018] The present invention can deeply integrate the RFID signal strength, the remaining battery power of the robot, and the dynamic obstacle distribution data, dynamically adjust the path priority through real-time data collection and intelligent analysis technology, and can automatically plan to bypass preferentially for weak signal coverage areas and low battery threshold ranges to ensure the continuity and accuracy of the inventory task and improve the task execution ability in complex environments. The method of the present invention can use the PSO algorithm to flexibly allocate the weight ratios of path length, signal coverage density, energy consumption, and obstacle avoidance safety according to the real-time needs of warehousing operations, dynamically coordinate the signal integrity and energy management requirements while ensuring the path efficiency, realize intelligent strategy adaptation in different operation scenarios, and significantly improve the inventory efficiency and the battery life performance of the device. The present invention constructs a multi-level obstacle avoidance protection mechanism through dynamic obstacle trajectory prediction and segmented path planning technology, can capture the dynamic obstacle information in the environment in real time, predict its movement trajectory and generate a safe avoidance path, effectively reducing the collision risk and ensuring the stable operation of the robot in a warehousing environment with high-frequency personnel flow and equipment movement. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a flowchart of an intelligent inventory robot path optimization method for environment modeling and collaborative optimization provided by an embodiment of the present invention; Figure 2 It is a system block diagram of an intelligent inventory robot path optimization system for environment modeling and collaborative optimization provided by an embodiment of the present invention; Figure 3 It is a schematic hardware structure diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part rather than all of the embodiments of the present invention. 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.

[0021] As Figure 1 shown, it is a flowchart of a path optimization method for an intelligent inventory-taking robot for environmental modeling and collaborative optimization provided by an embodiment of the present invention. Figure 1 The execution subject of the method shown can be a software and / or hardware device. The execution subject of the present application can include but is not limited to at least one of the following: user equipment, network equipment, etc. Among them, the user equipment can include but is not limited to computers, smartphones, personal digital assistants (Personal Digital Assistant, abbreviated as: PDA), and the above-mentioned electronic devices, etc. The network equipment can include but is not limited to a single network server, a server group composed of multiple network servers, or a cloud composed of a large number of computers or network servers based on cloud computing. Among them, cloud computing is a type of distributed computing, which consists of a group of loosely coupled computers forming a super virtual computer. This embodiment does not make any restrictions. It includes steps S1 to S4, specifically as follows: S1, scan the target warehouse area through a drone platform and generate three-dimensional warehouse point cloud data, perform point cloud denoising, plane fitting, shelf boundary extraction, and channel detection processing on the three-dimensional warehouse point cloud data to generate a two-dimensional warehouse floor plan and transmit it to the robot data base; Among them, the drone platform refers to an unmanned aerial vehicle equipped with sensors such as lidar and cameras, which can scan the target warehouse area in all directions according to a preset route, collect three-dimensional space data in a large range and with high efficiency, and is especially suitable for warehouse scenarios with high shelves or complex layouts. The three-dimensional warehouse point cloud data refers to a massive set of discrete points formed by the reflection of laser beams emitted by lidar in space. Each point contains x, y, z coordinates and reflection intensity information, which can completely represent the geometric shape of the warehouse environment and accurately restore the spatial positions of shelves, channels, and obstacles. The two-dimensional warehouse floor plan refers to a grid map formed by projecting the three-dimensional point cloud data onto the horizontal plane, which is marked with information such as shelf positions, channel widths, and entrances and exits. The data base of the intelligent inventory-taking robot usually integrates a wireless communication module and a charging facility, which can receive, store, and distribute two-dimensional map data.

[0022] In practical applications, through an unmanned aerial vehicle (UAV) platform equipped with multiple sensors, a target warehousing area can be scanned comprehensively. Then, by utilizing the point cloud acquisition ability of lidar and the pose solution function of the IMU inertial navigation system, three-dimensional warehousing point cloud data containing spatial coordinates and reflection intensity information can be generated.

[0023] Specifically, first, the voxel grid filtering technique can be adopted to divide the continuous three-dimensional space into uniform cubic units. By retaining the centroid points of the voxels, point cloud downsampling and noise suppression can be achieved. Subsequently, the Random Sample Consensus (RANSAC) algorithm can be applied to fit planar structures, such as regular geometric elements like the ground and shelf laminates, from the point cloud. And through Delaunay triangulation and Alpha Shapes algorithm, the shelf boundary polygons can be extracted, and the Douglas-Peucker algorithm can be used for contour simplification to reduce the data complexity while maintaining geometric features. Finally, through bird's-eye view rasterization and morphological closing operation, the shelf occupancy area and the passable channels can be distinguished, and a two-dimensional warehousing floor plan marked with shelf layout, channel width, and entrance / exit positions can be generated and transmitted to the robot data base via a wireless communication link to provide a static environment model for subsequent path planning.

[0024] S2. Use the A* algorithm to generate an initial inventory planning path based on the two-dimensional warehousing floor plan, and optimize the initial inventory planning path through the simulated annealing algorithm to generate a target inventory planning path. It can be understood that the A* algorithm (A-Star Algorithm) is a heuristic path search algorithm that can quickly and accurately determine the shortest path in a two-dimensional grid map. The initial inventory planning path refers to the initial path planned based on the two-dimensional warehousing floor plan for inventorying all shelves. The simulated annealing algorithm (Simulated Annealing, SA) is a global optimization algorithm that allows accepting poor solutions in the search to jump out of local optima by simulating the physical annealing process, thereby optimizing the initial path, reducing repeated scans and ineffective walks, and improving path smoothness. The target inventory planning path refers to the path for the intelligent inventory robot to conduct shelf inventory after being optimized by the simulated annealing algorithm.

[0025] Specifically, based on the two-dimensional warehousing floor plan, the A* algorithm evaluation function can be used to balance the actual cost and the target heuristic estimate of the path. Among them, the actual cost covers the moving distance, the number of turns, and the channel width constraint, and the heuristic function uses the Manhattan distance to quickly estimate the shortest path from the node to the end point. Through the priority queue expansion mechanism, the A* algorithm can ensure that the node with the minimum comprehensive cost is selected for expansion each time, so that in the grid map, a shortest path sequence covering all inventory shelves can be generated, and this sequence contains continuous turning points and straight path segments.

[0026] To further improve the path quality, the simulated annealing algorithm can be used to locally optimize the initial path. This algorithm generates new paths through random perturbation operations and dynamically accepts inferior solutions during the cooling process according to the Metropolis acceptance criterion to jump out of the local optimum. The corresponding specific operations include: performing multiple path perturbations within each temperature iteration cycle, calculating the change in path length, directly accepting the new path if it is better, and otherwise accepting the inferior solution with a probability related to the temperature. As the temperature parameter decays exponentially, the algorithm gradually converges to a globally better solution, generating a target inventory planning path without repeated scans and low redundancy.

[0027] S3. Based on the collected dynamic obstacle influence factors, combined with the RFID signal strength and the remaining battery power of the robot, correct the target inventory planning path through a hybrid algorithm of multi-objective optimization and PSO dynamic weight. It should be noted that the dynamic obstacle influence factor is a factor used to quantify the influence degree of dynamic obstacles on the path. The RFID signal strength refers to the signal strength value of the ultra-high frequency radio frequency identification tag, which can reflect the communication quality between the tag and the robot. The remaining battery power of the robot refers to the remaining battery power of the intelligent inventory robot at a certain moment during the shelf inventory process. The AR device refers to an augmented reality terminal based on SLAM technology, which superimposes and displays the virtual path and the physical environment through spatial mapping and supports gesture interaction and voice commands.

[0028] During the operation of the robot, it is necessary to respond to the changes in the warehouse environment in real time. Collect the position data of dynamic obstacles through lidar and cameras, combine the RFID signal strength and the remaining battery power of the robot to construct multi-dimensional constraint conditions, and the RFID signal strength can reflect the shielding effect of metal shelves. The multi-objective optimization function can integrate conflicting objectives such as path length, signal coverage quality, energy consumption, and obstacle influence, and dynamically adjust the weight coefficients of each objective through the particle swarm optimization (PSO) algorithm to achieve strategy adaptation in different operation scenarios.

[0029] The PSO algorithm simulates the cooperation mechanism of bird flocks, regards each group of weight coefficients as particles, and guides the particles to search for the optimal solution in the weight space through individual optimal and global optimal. Then, the optimized weights can be injected into the improved A* algorithm to adjust the heuristic function and balance the multi-objective requirements. For example, increasing the scanning priority in areas with weak signals, preferentially planning charging paths when the battery power is insufficient, or generating obstacle avoidance detour trajectories for dynamic obstacles.

[0030] This hybrid algorithm can achieve closed-loop control of static planning and dynamic correction, significantly improve the path adaptability of the robot in dynamic scenarios such as personnel movement and shelf adjustment, and reduce the path redundancy rate.

[0031] S4. The operating terminal views the inventory progress of the robot through the AR device and manually adjusts the target inventory planning path.

[0032] Among them, the operating terminal refers to the personnel responsible for task scheduling, path monitoring, and exception handling during the robot inventory process. The robot inventory progress is the core indicator used to quantify and monitor the robot's operation status, which can reflect the completion progress of the robot's scanning, identification, and data verification of warehouse goods within a specified time, and can thus serve as an important basis for human-machine collaborative intervention, providing visual task execution feedback to the operator.

[0033] Through the AR device, a virtual-real fusion monitoring interface can be constructed to achieve the visual display of the robot inventory progress. The AR device aligns the virtual path information with the physical warehouse environment through spatial registration technology, so that the operator can directly view key indicators such as the robot's position, the area where the inventory has been completed, and the RFID reading status.

[0034] In addition, when signal reading anomalies, equipment failures, or emergency tasks are detected, the operator can mark a temporary no-go area, adjust the shelf inventory priority, or issue an immediate obstacle avoidance instruction through the AR interface. After receiving the instruction, the robot triggers the path replanning mechanism and updates the operation strategy in real time.

[0035] This human-machine collaborative mechanism makes up for the lack of flexibility of a pure automated system in complex abnormal scenarios, forming a two-way feedback closed loop of manual monitoring and intelligent execution. For example, when encountering a temporary stacking obstacle, the operator can manually draw a detour path through the AR device, and the robot quickly generates a new path based on the improved A* algorithm and executes it, thereby ensuring the continuity of the task and improving the reliability of the inventory operation.

[0036] In the specific implementation process, scanning the target warehouse area by the drone platform and generating three-dimensional warehouse point cloud data, and performing point cloud denoising, plane fitting, shelf boundary extraction, and channel detection processing on the three-dimensional warehouse point cloud data to generate a two-dimensional warehouse floor plan and transmit it to the robot data base specifically includes: The drone platform is equipped with a 16-line lidar and an IMU inertial navigation system, and performs overlapping scans along the preset grid route of the target warehouse area at a preset cruise speed to generate the original warehouse point cloud data; Using the voxel grid filtering technology to divide the three-dimensional space corresponding to the original warehouse point cloud data into voxels of 0.08m³ cubes, and through the formula Perform voxel indexing on each of the original warehouse point cloud data and retain the centroid points within each voxel to generate the three-dimensional warehouse point cloud data; Among them, represents the calculation point in the three-dimensional warehouse point cloud data Voxels; respectively represent the side lengths of the voxel in the x, y, and z directions; respectively represent the voxel indices to which the calculation point p belongs in the x, y, and z directions; W represents the total width of the voxel grid in the x direction, i.e., the number of voxels in the x direction; H represents the total width of the voxel grid in the y direction, i.e., the number of voxels in the y direction.

[0037] Among them, the UAV platform undertakes the core tasks of environmental perception and data collection. It is equipped with a 16-line lidar and an IMU inertial navigation system, and can perform overlapping scans on the target storage area according to the preset cruising speed and grid route. This process realizes the three-dimensional coverage of the storage space through the collaborative operation of multiple sensors. And the 16-line lidar can construct the original point cloud data containing three-dimensional coordinates and reflection intensity information by emitting laser beams and receiving reflected signals, while the IMU inertial navigation system can real-time calculate the pose parameters of the UAV, i.e., the translation and rotation matrices, providing the reference alignment of the spatial coordinate system for the point cloud data to ensure the accurate splicing of data in different scanning areas.

[0038] Furthermore, first, the voxel grid filtering technology can be used to discretize the three-dimensional space and map the original point cloud data into the voxel grid composed of 0.08m³ cubes. Then, according to the three-dimensional coordinates of each original point cloud data point, it can be assigned to the corresponding voxel unit. By calculating the centroid coordinates of all points in each voxel and retaining the representative points that can characterize the spatial characteristics of the voxel, the three-dimensional warehouse point cloud data can be generated, so that while removing redundant data, the geometric characteristics of the environment can be retained. This voxelization process not only reduces the scale of the point cloud data, but also suppresses the influence of random noise through the statistical characteristics of the centroid points, improving the reliability of the data.

[0039] The core mechanism of voxel indexing is to realize the grouped management of point clouds through the quantization mapping of spatial coordinates. For any calculation point, its position in the three-dimensional space can be decomposed into three axial coordinate components. By rounding the ratio of each component to the voxel side length, the index position of the point in the voxel grid can be determined. Specifically, the voxel indices of the calculation point in the x, y, and z directions are obtained by rounding the ratio of its coordinate value to the voxel side length in the corresponding axial direction, so that the three-dimensional space point cloud can be converted into a discrete voxel unit index sequence. This indexing method not only facilitates the storage and retrieval of spatial data by the computer, but also provides a structured data basis for subsequent operations such as plane fitting and boundary extraction.

[0040] The process of the shelf boundary extraction and channel detection is as follows: In the three-dimensional coordinate system corresponding to the three-dimensional warehousing point cloud data, for each plane, a local coordinate system orthogonal to the plane normal vector is constructed with the plane centroid as the origin, and the three-dimensional warehousing point cloud data is projected onto the local coordinate system; Perform Delaunay triangulation on the projected point cloud, generate a closed shelf boundary polygon by screening the triangle sides with the smallest circumcircle radius, and perform boundary error control on the closed shelf boundary polygon through the Douglas-Peucker algorithm to generate a shelf boundary contour; Convert the ground point cloud into a bird's-eye view grid, fill the pixel gaps in the three-dimensional warehousing point cloud data through morphological closing operations, and calculate the pixel occupancy probability ; Obtain the shelf channel discrimination probability and the preset width limit, and the pixel occupancy probability All pixel points with a pixel occupancy probability greater than the shelf channel discrimination probability form shelf elements, and all pixel points with a pixel occupancy probability less than the shelf channel discrimination probability form channel elements. At the same time, mark the entrance and exit elements, and the width between the pixel points corresponding to both ends of the entrance and exit elements is greater than the preset width limit; Attach a shelf ID, a channel width, and an entrance and exit direction to the shelf elements, the channel elements, and the entrance and exit elements respectively.

[0041] It can be understood that in the processing framework of three-dimensional warehousing point cloud data, for each fitted plane, such as a shelf board, a wall, etc., a local coordinate system is constructed with the plane centroid as the origin. The coordinate axes of this local coordinate system are orthogonal to the plane normal vector, that is, the normal vector serves as a certain main axis of the local coordinate system, usually the z-axis, and the other two axes extend along the plane tangent direction. Through this transformation, the three-dimensional point cloud data is projected onto the two-dimensional plane of the local coordinate system, such as the xy plane, so that the dimensionality reduction processing of the point cloud data can be realized through coordinate transformation, and the efficiency and accuracy of feature extraction can be improved by using the geometric prior knowledge of the plane.

[0042] Furthermore, Delaunay triangulation can be performed on the point cloud projected onto the local coordinate system to construct a triangular grid covering all points. This grid has the empty circle property, that is, no other points are contained within the circumcircle of any triangle, to ensure the geometric rationality of the grid. By screening the sides of the triangle with the smallest circumcircle radius less than the threshold, the boundary edges that can represent the shelf contour are extracted, and a closed shelf boundary polygon can be generated, so that the redundant edges caused by noise or discrete points can be excluded, and the contour line closely following the shelf point cloud distribution can be retained.

[0043] To further enhance the practicality of the boundary, the Douglas-Peucker algorithm can be used to optimize the closed polygon. This algorithm recursively evaluates the contribution of vertices to the boundary shape, deletes vertices whose distance from the fitted line is less than the preset error threshold, and reduces the number of vertices while maintaining the overall shape of the boundary. This process effectively reduces the complexity of the boundary data, avoids computational redundancy caused by excessive details, and ensures that the geometric accuracy of the shelf outline in the warehouse floor plan meets the path planning requirements.

[0044] Then, the ground point cloud data can be converted into a bird's-eye view raster, where each raster cell corresponds to a fixed size in the physical space. Through morphological closing operations, that is, dilation followed by erosion, the raster data is processed to fill pixel gaps caused by sparse point clouds or noise and smooth the boundaries between shelves and aisles. The application of the closing operation can effectively connect adjacent shelf point cloud regions, avoid misjudgment of aisles due to missing point clouds, and maintain the geometric connectivity of the shelf regions.

[0045] Based on the rasterization, the occupancy probability of each pixel can be calculated. This probability reflects the likelihood that the position is occupied by a shelf and is usually generated by statistically calculating the point cloud density or reflection intensity within the raster. The value range of the occupancy probability is [0, 1], and the higher the value, the greater the likelihood of the presence of a shelf.

[0046] Based on a preset probability threshold for differentiating shelves and aisles, the raster pixels can be binarized. Pixels with an occupancy probability greater than the threshold are classified as shelf elements, and pixels with a probability less than the threshold are classified as aisle elements. And the setting of this threshold needs to combine prior knowledge of the warehouse scenario. For example, in areas with dense metal shelves, the threshold can be set to 0.7 to ensure the integrity of the shelf area. At the same time, by detecting connected regions in the aisle elements with a width greater than a preset limit and marking the entrance and exit elements, it can be ensured that the entrances and exits meet the requirements for two-way passage or turning operations of the robot, and this limit is usually set according to the robot size and passage requirements.

[0047] To achieve semantic management of warehouse elements, attribute information can be attached to shelf elements, aisle elements, and entrance and exit elements respectively: shelf ID, which is used to identify the physical location and storage attributes of the shelf, such as the location number of the storage space and the type of goods; aisle width, which is used to record the width parameter of the aisle and provide a traffic capacity constraint for path planning; entrance and exit direction, which is used to mark the direction attributes of the entrance and exit, such as north-south or east-west, to assist in path selection during robot navigation. These attribute information are embedded in the two-dimensional warehouse floor plan in the form of metadata, forming a structured semantic map, which provides multi-dimensional data support for subsequent path planning, task scheduling, and conflict avoidance.

[0048] The method of generating an initial inventory planning path based on the two-dimensional warehouse floor plan by using the A* algorithm and optimizing the initial inventory planning path by using the simulated annealing algorithm to generate a target inventory planning path specifically includes: Determine the inventory shelf positions based on the two-dimensional warehouse floor plan, obtain the types of goods assets corresponding to the inventory shelf positions and the inventory position priorities from the warehouse management database corresponding to the target warehouse area, and set the movement parameters for the robot to perform inventory, namely the maximum walking speed, the minimum safety distance, and the single-shelf single-layer scanning time; Use the A* algorithm to search for the shortest path from the starting inventory position to the ending shelf position, and combine the types of goods assets and the inventory position priorities to generate the initial inventory planning path; Perform local optimization on the initial inventory planning path, that is, optimize the initial inventory planning path by using the simulated annealing algorithm to generate the target inventory planning path.

[0049] In practical applications, all the shelf positions that need to be inventoried can be identified based on the two-dimensional warehouse floor plan, and the types of goods assets and the inventory priority data corresponding to each shelf can be retrieved from the warehouse management database. Among them, the types of goods assets determine the scanning strategy. For example, high-value assets need to be repeatedly verified, while the inventory priorities are used for path sorting and optimization. At the same time, the movement parameters of the intelligent inventory robot can be set, including the maximum walking speed, the minimum safety distance, and the single-shelf single-layer scanning time. These parameters, as constraints, can ensure the physical executability of the generated path.

[0050] Subsequently, the A* algorithm can be used to generate the initial inventory path. This algorithm uses the grid map of the two-dimensional floor plan as the search space, calculates the node cost through a comprehensive evaluation function. Among them, the actual movement cost covers the distance, the number of turns, and the influence of the aisle width, and the heuristic estimated cost is based on the spatial distance to the target position. The algorithm iteratively expands the node with the minimum cost through a priority queue, sorts the path nodes by combining the types of goods assets and the inventory priorities, and generates a shortest path sequence from the starting position to all target shelves. This sequence contains continuous straight path segments and turning points, ensuring that all inventory tasks are covered and the passage constraints are met.

[0051] To further improve the path quality, the simulated annealing algorithm can be used to optimize the initial path. This algorithm generates new paths through random perturbation operations and dynamically accepts inferior solutions during the cooling process according to the Metropolis acceptance criterion to jump out of the local optimum. The specific operations include: performing multiple path transformations in each temperature iteration cycle, calculating the changes in path length and constraint satisfaction. If the new path is better, it is directly adopted; otherwise, a worse solution is accepted with a probability related to the current temperature. As the temperature parameter decays exponentially, the algorithm gradually converges to a globally better solution and generates a target inventory planning path with no redundant detours and the fewest number of turns. This path is theoretically shorter than the initial path by a certain proportion and has better motion smoothness and task execution efficiency.

[0052] The shortest path from the starting inventory position to the ending shelf position is searched using the A* algorithm, and the initial inventory planning path is generated by combining the type of goods assets and the priority of the inventory positions, specifically including: Set the node cost function as follows: Among them, f(n) represents the total estimated cost of the inventory shelf position node n; g(n) represents the actual movement cost from the starting inventory position node to the inventory shelf position node n; Represents the actual movement cost from the starting inventory position node to the previous inventory shelf position node Of; move_cost(n) represents the basic movement cost from the previous inventory shelf position node To the inventory shelf position node n. If a horizontal or vertical movement occurs from the previous inventory shelf position node To the inventory shelf position node n, then move_cost(n)=1. If a diagonal movement occurs from the previous inventory shelf position node parent(n) to the inventory shelf position node n, then ; turn_penalty(n) represents the turning penalty of the inventory shelf position node n. If a turn is required from the previous inventory shelf position node parent(n) to the inventory shelf position node n, then , Represents the turning penalty coefficient, otherwise turn_penalty(n)=0; width_penalty(n) represents the aisle width penalty of the inventory shelf position node n, , Represents the width penalty coefficient, Represents the standard aisle width, w(n) represents the aisle width where the inventory shelf position node n is located; h(n) represents the heuristic estimated cost from the inventory shelf position node n to the ending inventory position node; Represents the coordinates of the inventory shelf position node n; Represents the coordinates of the termination inventory position node; Expand and select the inventory shelf position node with the minimum total estimated cost through the priority queue until all the inventory shelf position nodes are traversed, generating a continuous walking sequence including turning points and straight path segments, that is, the shortest path from the starting inventory position to the termination shelf position, and combining the goods asset type and the inventory position priority to generate the initial inventory planning path.

[0053] Among them, the node cost function adopts a two-component model, and the total estimated cost is composed of the actual movement cost and the heuristic estimated cost. The calculation of the actual movement cost includes four parts: the cumulative cost of the previous node, the basic movement cost, the turning penalty, and the channel width penalty. Among them, the basic movement cost sets different weights according to the difference in movement directions, that is, the horizontal or vertical movement corresponds to the unit cost, and the diagonal movement sets a higher weight because of the longer distance. The turning penalty mechanism is used to suppress frequent turning. When the movement direction between the current node and the previous node changes, the penalty coefficient is superimposed; the channel width penalty is calculated by the ratio of the standard width to the channel width where the current node is located to guide the path to preferentially select a wider channel to reduce the collision risk.

[0054] The heuristic estimated cost adopts the Manhattan distance model to quickly estimate the remaining path length and ensure that the algorithm converges to the optimal solution in polynomial time. The priority control mechanism obtains the goods asset type and inventory priority data through the warehouse management system, and the nodes corresponding to the high-priority shelves have a higher expansion weight in the priority queue, so as to ensure that the intelligent inventory robot preferentially accesses the key areas during the inventory process.

[0055] Then, through the priority queue, the node with the minimum total estimated cost can be dynamically expanded, maintaining a closed list containing the visited nodes and an open list of the nodes to be expanded. In each iteration, the node with the minimum cost is selected from the open list for neighborhood expansion, the total estimated cost of the adjacent nodes is calculated and the priority queue is updated until all the inventory shelf nodes are traversed. The finally generated path consists of continuous turning points and straight path segments, which not only satisfies the geometric shortestness but also realizes the path sorting guided by business requirements through the priority weight, forming an initial inventory planning path that conforms to the warehouse operation specifications. On the premise of ensuring no collision, this path balances the movement efficiency, operation smoothness and channel safety through the multi-factor weighting of the cost function.

[0056] Performing local optimization on the initial inventory planning path, that is, optimizing the initial inventory planning path through the simulated annealing algorithm to generate the target inventory planning path, specifically including: Setting the key parameters of the simulated annealing algorithm, that is, the initial temperature T, the cooling rate and the termination temperature and the number of disturbances Q at each temperature; Define the path perturbation operation to include randomly selecting two path nodes in the initial inventory planning path P, deleting the intermediate nodes of these two path nodes and reconnecting the initial inventory planning path P, randomly swapping the order of two non-adjacent path nodes in the initial inventory planning path P, and randomly selecting a sub-path in the initial inventory planning path P and reversing the direction of the sub-path; Perform the path perturbation operation on the initial inventory planning path P to generate a new inventory planning path and calculate the change in path length where L represents the initial path length of the initial inventory planning path P, represents the new path length of the new inventory planning path ; If , then ; If , then calculate the path acceptance probability based on the Metropolis acceptance criterion. If the path acceptance probability q is greater than the random number r, , then , otherwise ; The initial temperature T converges step by step according to the cooling rate , and the path perturbation operation is performed Q times at each temperature until the temperature is less than the termination temperature , terminate the simulated annealing algorithm, and output the target inventory planning path.

[0057] It should be noted that first, algorithm parameter initialization is required, setting the initial temperature, cooling rate, termination temperature, and the number of disturbances per temperature. Among them, the initial temperature determines the breadth of the search space. A higher initial temperature value allows the algorithm to explore more path variants in the early stage; the cooling rate controls the temperature decay speed, usually a constant close to 1 to balance exploration and convergence; the termination temperature serves as the algorithm stop condition to ensure the stability of the solution; the number of disturbances per temperature limits the local search depth at each temperature level.

[0058] The path perturbation operation realizes path structure variation through three basic transformations: randomly selecting two nodes in the path, deleting the intermediate nodes of the two nodes and directly connecting them to shorten the path; randomly swapping the order of non-adjacent nodes to adjust the access priority; randomly reversing the direction of the sub-path to explore the local optimal arrangement. These operations are combined and applied with a certain probability to generate diverse new path variants, breaking through the local structure limitations of the initial path.

[0059] After each perturbation, the algorithm can calculate the change in the length of the old and new paths. If the new path is shorter, the new path is directly accepted; if the new path is longer, the acceptance probability is calculated according to the Metropolis acceptance criterion, which has a negative exponential relationship with the temperature and the path length difference, allowing the algorithm to accept inferior solutions with a higher probability at the high-temperature stage, thus jumping out of the local optimal trap. By generating a random number between 0 and 1 and comparing it with the acceptance probability, it can be determined whether to retain the new path, forming a probabilistic optimization mechanism.

[0060] The temperature decay mechanism is the core of the convergence of the simulated annealing algorithm, and the initial temperature decays exponentially according to the cooling rate. A preset number of perturbation operations are performed at each temperature level. As the temperature decreases, the acceptance probability of the algorithm for inferior solutions gradually decreases, and finally converges to the local optimal solution at the low-temperature stage. When the temperature is lower than the termination temperature, the algorithm terminates and outputs the optimized target inventory planning path, which can significantly reduce the total travel length and improve the path smoothness by reducing redundant turns and merging the access order of adjacent shelves.

[0061] Based on the collected dynamic obstacle influence factors, combined with the RFID signal strength and the remaining battery power of the robot, the target inventory planning path is corrected by a hybrid algorithm of multi-objective optimization and PSO dynamic weights, specifically including: Normalize the RFID signal strength and the remaining battery power of the robot, and construct a multi-objective optimization function based on the target path length of the target inventory planning path, the normalized RFID signal strength and the remaining battery power of the robot, and the dynamic obstacle influence factor as follows: In the formula, F represents the comprehensive cost function; represents the target weight coefficient; represents the target path length of the target inventory planning path; represents the normalized RFID signal strength; represents the normalized remaining battery power of the robot; O represents the dynamic obstacle influence factor, , represents the distance from the intelligent inventory robot to the nearest obstacle, represents the safety distance threshold of the intelligent inventory robot, represents the dynamic obstacle influence coefficient; Each particle represents a set of target weight vectors , and the particle velocity and position are iteratively updated by PSO to determine the optimal combination of the target weight coefficients ; Based on the optimal combination of the target weight coefficients , the node cost function is adjusted to obtain an improved node cost function, and the target inventory planning path is corrected based on the improved node cost function. The expression of the improved node cost function is as follows: In the formula, represents the Manhattan distance from the inventory shelf position node n to the termination inventory position node; represents the signal heuristic; represents the power heuristic; represents the obstacle heuristic.

[0062] In practical applications, first, the RFID signal strength and the remaining battery power of the robot collected in real time can be normalized to convert physical quantities into dimensionless indicators in the [0,1] interval, so as to facilitate the unified measurement and fusion calculation of multi-source data.

[0063] Then, combined with the inherent length attribute of the target inventory planning path and the spatial influence factor of dynamic obstacles, a multi-objective optimization function including path length, signal quality, equipment endurance, and obstacle avoidance safety can be constructed. This function integrates the optimization objectives of each dimension into a comprehensive cost evaluation index in the form of weighted summation.

[0064] As the core mechanism for dynamic weight adjustment, the particle swarm optimization algorithm maps each group of target weight coefficients to individuals in the particle swarm. Through iterative search simulating the cooperation of bird flocks, the priority allocation of each target is dynamically optimized. The velocity and position update rules of the particles integrate the two-way guidance of the individual historical optimal solution and the global optimal solution of the group, ensuring that the weight vector efficiently searches for the optimal combination in the multi-dimensional space. After several iterations, the optimal weight coefficient combination can be obtained to balance the current environmental constraints. For example, in areas with weak signals, the weight of the signal quality target is automatically increased, or the priority of the endurance target is enhanced when the battery power is insufficient.

[0065] Based on the optimized weight coefficients, the original node cost function can be reconstructed, introducing signal heuristic, power heuristic, and obstacle heuristic components. The signal heuristic is negatively correlated with the normalized signal strength, so as to guide the path to preferentially cover the reliable signal area and improve the inventory accuracy; the power heuristic is negatively correlated with the remaining battery power to ensure that the low battery risk is avoided in advance during path planning; the obstacle heuristic is positively correlated with the dynamic obstacle influence factor, forcing the path to stay away from dangerous areas. Through the improved node cost function, the priority of the path nodes can be recalculated and local path replanning can be implemented to generate a corrected path adapted to the real-time environmental changes. This path can dynamically balance multiple objectives such as efficiency, accuracy, endurance, and safety on the premise of ensuring the completion of the inventory task, significantly improving the operation robustness of the robot in the dynamic warehouse environment.

[0066] Such asFigure 2 As shown in the figure, it is a system block diagram of a path optimization system for an intelligent inventory robot for environmental modeling and collaborative optimization provided by an embodiment of the present invention. The path optimization system includes: An unmanned aerial vehicle (UAV) scanning and data transmission module, configured to scan a target warehousing area through a UAV platform to generate three-dimensional warehousing point cloud data, perform point cloud denoising, plane fitting, shelf boundary extraction, and channel detection processing on the three-dimensional warehousing point cloud data to generate a two-dimensional warehousing floor plan, and transmit it to a robot data base; An inventory path generation and optimization module, configured to generate an initial inventory planning path based on the two-dimensional warehousing floor plan by using the A* algorithm, and optimize the initial inventory planning path through a simulated annealing algorithm to generate a target inventory planning path; A multi-objective optimization and PSO dynamic weight correction module, configured to correct the target inventory planning path based on the collected dynamic obstacle influence factors, in combination with RFID signal strength and the remaining battery power of the robot, through a hybrid algorithm of multi-objective optimization and PSO dynamic weight; An inventory progress viewing and manual adjustment module, configured to view the inventory progress of the robot through an AR device at an operation end and manually adjust the target inventory planning path.

[0067] Figure 2 The device of the illustrated embodiment can correspondingly be used to execute Figure 1 the steps in the method embodiment shown. The implementation principle and technical effects are similar and will not be elaborated here.

[0068] An electronic device includes a memory and a processor. A computer program is stored in the memory. When the processor runs the computer program stored in the memory, the processor executes the steps of a method for optimizing the path of an intelligent inventory robot for environmental modeling and collaborative optimization as described in any one of the above.

[0069] As Figure 3 shown, it is a schematic hardware structure diagram of an electronic device provided by an embodiment of the present invention. The electronic device 30 includes: a processor 31, a memory 32, and a computer program; where The memory 32 is used to store the computer program. The memory can also be a flash memory. The computer program is, for example, an application program or a functional module for implementing the above method.

[0070] The processor 31 is configured to execute the computer program stored in the memory to implement each step executed by the device in the above method. Specifically, reference can be made to the relevant descriptions in the previous method embodiment.

[0071] Optionally, the memory 32 can be either independent or integrated with the processor 31.

[0072] When the memory 32 is a device independent of the processor 31, the device may further include: A bus 33 for connecting the memory 32 and the processor 31.

[0073] A readable storage medium stores a computer program, and when the computer program is executed by a processor, it is used to implement the steps of the path optimization method of the intelligent inventory robot for environment modeling and collaborative optimization described in any one of the above.

[0074] Among them, the readable storage medium may be a computer storage medium or a communication medium. The communication medium includes any medium that facilitates the transmission of a computer program from one place to another. The computer storage medium may be any available medium that can be accessed by a general or special computer. For example, the readable storage medium is coupled to the processor so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium may also be a component of the processor. The processor and the readable storage medium may be located in an application specific integrated circuit (ASIC). In addition, the ASIC may be located in the user equipment. Of course, the processor and the readable storage medium may also exist as discrete components in the communication device. The readable storage medium may be a read only memory (ROM), a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0075] The present invention also provides a program product, which includes execution instructions stored in a readable storage medium. At least one processor of the device can read the execution instructions from the readable storage medium, and the execution of the execution instructions by at least one processor enables the device to implement the methods provided by the above various embodiments.

[0076] In the embodiments of the above device, it should be understood that the processor may be a central processing unit (CPU for short), and may also be other general processors, digital signal processors (DSP for short), application specific integrated circuits (ASIC for short), etc. The general processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the present invention may be directly embodied as being completed by the execution of the hardware processor, or completed by the combination of the hardware and software modules in the processor.

[0077] Through the introduction of the above embodiments, the intelligent inventory robot path optimization method of the present invention through environmental modeling and collaborative optimization scans the target warehouse area through a drone platform and generates three-dimensional warehouse point cloud data. The three-dimensional warehouse point cloud data is subjected to point cloud denoising, plane fitting, shelf boundary extraction, and channel detection processing to generate a two-dimensional warehouse floor plan and transmit it to the robot data base; the A* algorithm is used to generate an initial inventory planning path based on the two-dimensional warehouse floor plan, and the initial inventory planning path is optimized through the simulated annealing algorithm to generate a target inventory planning path; based on the collected dynamic obstacle influence factors, combined with the RFID signal strength and the remaining battery power of the robot, the target inventory planning path is corrected through a hybrid algorithm of multi-objective optimization and PSO dynamic weight; the operation terminal views the inventory progress of the robot through an AR device and manually adjusts the target inventory planning path, so that the inventory path of the intelligent inventory robot can be optimized in real time in a dynamic environment, improving the inventory efficiency, coverage rate, and battery life of the robot.

[0078] The present invention can deeply integrate RFID signal strength, remaining battery power of the robot, and dynamic obstacle distribution data, dynamically adjust the path priority through real-time data collection and intelligent analysis technology, and can automatically plan to bypass preferentially for weak signal coverage areas and low battery threshold ranges to ensure the continuity and accuracy of the inventory task and improve the task execution ability in complex environments. The method of the present invention can use the PSO algorithm to flexibly allocate the weight ratios of path length, signal coverage density, energy consumption, and obstacle avoidance safety according to the real-time requirements of warehouse operations, dynamically coordinate the signal integrity and energy management requirements while ensuring the efficiency of the path, realize intelligent strategy adaptation in different operation scenarios, and significantly improve the inventory efficiency and equipment battery life performance. The present invention constructs a multi-level obstacle avoidance protection mechanism through dynamic obstacle trajectory prediction and segmented path planning technology, can real-time capture dynamic obstacle information in the environment, predict its movement trajectory and generate a safe avoidance path, effectively reduce the collision risk, and ensure the stable operation of the robot in a warehouse environment with high-frequency personnel flow and equipment movement.

[0079] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent inventory robot path optimization method for environmental modeling and collaborative optimization, characterized in that, The path optimization method includes the following steps: Scanning the target warehouse area by a drone platform to generate three-dimensional warehouse point cloud data, performing point cloud denoising, plane fitting, shelf boundary extraction, and channel detection on the three-dimensional warehouse point cloud data to generate a two-dimensional warehouse floor plan and transmitting it to the robot data base; Using the A* algorithm to generate an initial inventory planning path based on the two-dimensional warehouse floor plan, and optimizing the initial inventory planning path through the simulated annealing algorithm to generate a target inventory planning path; Based on the collected dynamic obstacle influence factors, combining the RFID signal strength and the remaining battery power of the robot, correcting the target inventory planning path through a hybrid algorithm of multi-objective optimization and PSO dynamic weight; The operation terminal views the robot inventory progress through an AR device and manually adjusts the target inventory planning path.

2. The path optimization method of the intelligent inventory robot for environmental modeling and collaborative optimization according to claim 1, wherein, The step of scanning the target warehouse area by a drone platform to generate three-dimensional warehouse point cloud data, performing point cloud denoising, plane fitting, shelf boundary extraction, and channel detection on the three-dimensional warehouse point cloud data to generate a two-dimensional warehouse floor plan and transmitting it to the robot data base specifically includes: The drone platform is equipped with a 16-line lidar and an IMU inertial navigation system, and performs overlapping scans along the preset grid route of the target warehouse area at a preset cruise speed to generate the original warehouse point cloud data; The three-dimensional space corresponding to the original warehouse point cloud data is divided into voxels of 0.08 m³ cubes by using the voxel grid filtering technique, and through the formula perform voxel indexing on each of the original warehouse point cloud data, and retain the centroid points within each voxel to generate the three-dimensional warehouse point cloud data; Among them, represents the calculation point in the 3D warehousing point cloud data voxel; respectively represent the side lengths of the voxel in the x, y, and z directions; respectively represent the voxel indices to which the calculation point p belongs in the x, y, and z directions; W represents the total width of the voxel grid in the x direction, that is, the number of voxels in the x direction; H represents the total width of the voxel grid in the y direction, that is, the number of voxels in the y direction.

3. The path optimization method of the intelligent inventory robot for environmental modeling and collaborative optimization according to claim 1, wherein The process of shelf boundary extraction and channel detection is as follows: In the three-dimensional coordinate system corresponding to the three-dimensional warehouse point cloud data, for each plane, taking the plane centroid as the origin, constructing a local coordinate system orthogonal to the plane normal vector, and projecting the three-dimensional warehouse point cloud data onto the local coordinate system; Performing Delaunay triangulation on the projected point cloud, generating a closed shelf boundary polygon by screening the triangle sides with the smallest circumcircle radius, and controlling the boundary error of the closed shelf boundary polygon through the Douglas-Peucker algorithm to generate the shelf boundary contour; Convert the ground point cloud into a bird's-eye view raster, fill the pixel gaps in the 3D warehouse point cloud data through morphological closing operations, and calculate the pixel occupancy probability ; Obtain the shelf passage discrimination probability and the preset width limit, and the pixel occupancy probability All pixel points with a pixel occupancy probability greater than the shelf passage discrimination probability form shelf elements, and the pixel occupancy probability All pixel points with a pixel occupancy probability less than the shelf passage discrimination probability form passage elements, and at the same time mark the entrance and exit elements, and the width between the pixel points corresponding to both ends of the entrance and exit elements is greater than the preset width limit value; Attaching a shelf ID, a channel width, and an entrance / exit direction to the shelf element, the channel element, and the entrance / exit element respectively.

4. The path optimization method of the intelligent inventory robot for environmental modeling and collaborative optimization according to claim 1, characterized in that, The step of using the A* algorithm to generate an initial inventory planning path based on the two-dimensional warehouse floor plan, and optimizing the initial inventory planning path through the simulated annealing algorithm to generate a target inventory planning path specifically includes: Determining the inventory shelf positions based on the two-dimensional warehouse floor plan, obtaining the types of goods assets and the priority of inventory positions corresponding to the inventory shelf positions from the warehouse management database corresponding to the target warehouse area, and setting the robot inventory movement parameters, namely the maximum walking speed, the minimum safety distance, and the single-shelf single-layer scanning time; Using the A* algorithm to search for the shortest path from the starting inventory position to the ending shelf position, and combining the types of goods assets and the priority of inventory positions to generate the initial inventory planning path; Performing local optimization on the initial inventory planning path, that is, optimizing the initial inventory planning path through the simulated annealing algorithm to generate the target inventory planning path.

5. An intelligent inventory robot path optimization method for environmental modeling and collaborative optimization according to claim 4, characterized in that The shortest path from the starting inventory position to the ending shelf position is searched using the A* algorithm, and the initial inventory planning path is generated by combining the types of goods assets and the priorities of the inventory positions, specifically including: Set the node cost function as follows: Among them, f(n) represents the total estimated cost of inventorying the shelf location node n; g(n) represents the actual movement cost from the starting inventory location node to the shelf location node n; represents the actual movement cost from the starting inventory location node to the previous shelf location node ; move_cost(n) represents the basic movement cost from the previous shelf location node to the shelf location node n. If there is a horizontal or vertical movement from the previous shelf location node to the shelf location node n, then move_cost(n) = 1. If there is a diagonal movement from the previous shelf location node parent(n) to the shelf location node n, then ; turn_penalty(n) represents the turning penalty of the shelf location node n. If a turn is required from the previous shelf location node parent(n) to the shelf location node n, then , represents the turning penalty coefficient, otherwise turn_penalty(n) = 0; width_penalty(n) represents the aisle width penalty of the shelf location node n, , represents the width penalty coefficient, represents the standard aisle width, w(n) represents the aisle width where the shelf location node n is located; h(n) represents the heuristic estimated cost from the shelf location node n to the ending inventory location node; represents the coordinates of the shelf location node n; represents the coordinates of the ending inventory location node; Expand and select the inventory shelf position node with the minimum total estimated cost through the priority queue until all the inventory shelf position nodes are traversed, generating a continuous walking sequence including turning points and straight path segments, that is, the shortest path from the starting inventory position to the ending shelf position, and generating the initial inventory planning path by combining the types of goods assets and the priorities of the inventory positions.

6. The path optimization method of the intelligent inventory robot for environmental modeling and collaborative optimization according to claim 4, characterized in that, The local optimization of the initial inventory planning path, that is, optimizing the initial inventory planning path through the simulated annealing algorithm to generate the target inventory planning path, specifically including: Set the key parameters of the simulated annealing algorithm, namely the initial temperature T, the cooling rate , the termination temperature and the number of perturbations Q at each temperature; Define the path perturbation operations including randomly selecting two path nodes in the initial inventory planning path P, deleting the intermediate nodes between these two path nodes and reconnecting the initial inventory planning path P, randomly swapping the order of two non-adjacent path nodes in the initial inventory planning path P, and randomly selecting a sub-path in the initial inventory planning path P and reversing the direction of the sub-path; Perform the path perturbation operation on the initial inventory planning path P to generate a new inventory planning path , calculate the change in path length , where L represents the initial path length of the initial inventory planning path P, represents the new inventory planning path 's new path length; If , then ; If , then calculate the path acceptance probability based on the Metropolis acceptance criterion . If the path acceptance probability q is greater than the random number r , then , otherwise ; The initial temperature T decreases step by step according to the cooling rate and the path perturbation operation is performed Q times at each temperature until the temperature is less than the termination temperature , terminate the simulated annealing algorithm and output the target inventory planning path.

7. The path optimization method for an intelligent inventory-taking robot with environmental modeling and collaborative optimization according to claim 5, characterized in that Based on the collected dynamic obstacle influence factor, combined with the RFID signal strength and the remaining battery power of the robot, correct the target inventory planning path through a hybrid algorithm of multi-objective optimization and PSO dynamic weight, specifically including: Normalize the RFID signal strength and the remaining battery power of the robot, and based on the target path length of the target inventory planning path, the normalized RFID signal strength and the remaining battery power of the robot, and the dynamic obstacle influence factor, construct the multi-objective optimization function as follows: In the formula, F represents the comprehensive cost function; represents the target weight coefficient; represents the target path length of the target inventory planning path; represents the RFID signal strength after normalization; represents the remaining battery power of the robot after normalization; O represents the dynamic obstacle influence factor, , represents the distance from the intelligent inventory robot to the nearest obstacle, represents the safety distance threshold of the intelligent inventory robot, represents the dynamic obstacle influence coefficient; Each particle represents a set of target weight vectors , and the particle velocity and position are updated iteratively by PSO to determine the optimal combination of the target weight coefficients ; Optimal combination based on the target weight coefficient , adjust the node cost function to obtain an improved node cost function, and correct the target inventory planning path based on the improved node cost function. The expression of the improved node cost function is as follows: In the formula, represents the Manhattan distance from the inventory shelf position node n to the termination inventory position node; represents the signal heuristic; represents the power heuristic; represents the obstacle heuristic.

8. An intelligent inventory robot path optimization system for environmental modeling and collaborative optimization, which is applied to an intelligent inventory robot path optimization method for environmental modeling and collaborative optimization as described in any one of claims 1-7, and is characterized in that The path optimization system includes: The UAV scanning and data transmission module is used to scan the target storage area through the UAV platform and generate three-dimensional storage point cloud data, perform point cloud denoising, plane fitting, shelf boundary extraction and channel detection on the three-dimensional storage point cloud data to generate a two-dimensional storage floor plan and transmit it to the robot data base; The inventory path generation and optimization module is used to generate the initial inventory planning path based on the two-dimensional storage floor plan using the A* algorithm, and optimize the initial inventory planning path through the simulated annealing algorithm to generate the target inventory planning path; The multi-objective optimization and PSO dynamic weight correction module is used to correct the target inventory planning path through a hybrid algorithm of multi-objective optimization and PSO dynamic weight based on the collected dynamic obstacle influence factor, combined with the RFID signal strength and the remaining battery power of the robot; The inventory progress viewing and manual adjustment module is used for the operation terminal to view the inventory progress of the robot through the AR device and manually adjust the target inventory planning path.

9. An electronic device, comprising a memory and a processor, wherein a computer program is stored in the memory, characterized in that, When the processor runs the computer program stored in the memory, the processor executes the steps of an intelligent inventory robot path optimization method for environment modeling and collaborative optimization as described in any one of claims 1-7.

10. A readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it is used to implement the steps of an intelligent inventory robot path optimization method for environmental modeling and collaborative optimization as described in any one of claims 1-7.

Images