A method and system for intelligent collaborative dispatching of transport forklifts for warehouse management

Abstract

The present invention relates to the technical field of forklift scheduling, and specifically to an intelligent collaborative scheduling method and system for transport forklifts for warehouse management. First, the present invention calculates the offset angle between the centroid of the pallet and the centroid of the object, detects the offset between the pallet and the object in real time, realizes early warning of the risk of cargo dumping, and greatly improves the safety of the transportation process; secondly, the present invention combines the dumping risk index and the operating status of the forklift, constructs a constraint function and uses a genetic algorithm for optimization, comprehensively considers the key parameters such as the operating speed, acceleration, turning speed and turning radius of the forklift, thereby realizing the global optimization of the operating status of individual forklifts and significantly improving transportation efficiency; finally, the present invention utilizes the optimal path of individual forklifts and distributed optimization algorithms to achieve global optimal scheduling in dynamic task allocation and path planning, minimize path conflicts and resource waste, and further ensure the efficiency and reliability of the system.

Description

Technical Field

[0001] The present invention relates to the technical field of forklift dispatching, and in particular to an intelligent collaborative dispatching method and system for transport forklifts oriented to warehouse management. Background Art

[0002] Transport forklifts are core equipment in warehouse management and are widely used in the handling, loading and unloading, and storage of goods. With the rapid development of the logistics industry and the continuous expansion of warehousing scale, the automation and intelligence of transport forklifts have become a key way to improve warehousing efficiency. At present, many companies use automated guided vehicles (AGVs) or unmanned forklifts to replace traditional manual operations to achieve efficient cargo transportation and dynamic scheduling. These intelligent forklifts can perform tasks in complex warehousing environments, significantly reducing labor costs, and improving the accuracy and consistency of operations, injecting new vitality into the modern warehousing system.

[0003] However, there are still many problems in the scheduling and management of transport forklifts at this stage. Due to the deviation of the centroid of the goods or improper placement, the goods are prone to tipping over during transportation, threatening safety; when multiple forklifts work together, collisions between forklifts often occur due to path conflicts, uneven task distribution or insufficient environmental perception; in addition, the scheduling algorithm's lack of dynamic adjustment and global optimization capabilities often leads to delays in forklift transportation tasks, further reducing warehouse management efficiency. These problems urgently need to be solved through more advanced intelligent scheduling methods and more sophisticated operation control strategies to achieve more efficient and safer warehouse management.

[0004] To this end, an intelligent collaborative scheduling method and system for transport forklifts for warehouse management is proposed. Summary of the invention

[0005] The purpose of the present invention is to provide a method and system for intelligent collaborative scheduling of transport forklifts for warehouse management. First, the present invention detects the degree of offset between the centroid of the pallet and the centroid of the object in real time through the offset angle calculation method, realizes early warning of the risk of cargo dumping, and effectively improves the safety of the transportation process; secondly, the present invention constructs a constraint function through the dumping risk index and the operating state of the forklift and adopts genetic algorithm optimization, comprehensively considering factors such as the operating speed, acceleration, turning speed and turning radius of the forklift, and realizes the global optimization of the operating state of the individual forklift, thereby significantly improving the transportation efficiency; finally, the present invention combines the optimal path of the individual forklift and the distributed optimization algorithm to realize the global optimal scheduling in the dynamic task allocation and path planning, minimizes the path conflict and resource waste, and ensures the efficiency and reliability of the system.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A method for intelligent collaborative dispatching of transport forklifts for warehouse management, comprising:

[0008] Acquire a pallet image and an image of an object above the pallet, perform centroid detection on the pallet image and the object image, and obtain the pallet centroid and the object centroid;

[0009] Calculating the vertical offset angle between the centroid of the tray and the centroid of the object, and if the offset angle is greater than a first offset threshold, issuing an alarm and uploading alarm information;

[0010] The vibration amplitude and vibration frequency of the forklift during operation are obtained by a vibration sensor under the fork arm, and a tipping risk index is calculated according to the vibration amplitude, the vibration frequency and the relative height of the centroid of the object;

[0011] A constraint function is established by combining the tipping risk index and the operation state of the forklift, and the constraint function is optimized by a genetic algorithm to obtain the optimal state of the individual forklift; the operation state of the forklift includes movement speed, acceleration, turning speed and turning radius;

[0012] Combining the SLAM map and the optimal state of the individual forklift, the path planning is performed through the A* algorithm to obtain the optimal path of the individual forklift from the starting point to the target point;

[0013] The individual forklift broadcasts its own optimal path and status information to other individual forklifts, and all individual forklifts coordinately adjust their own optimal paths to obtain a global optimal path; the status information includes path information, task information and the operating status of the forklift.

[0014] Furthermore, obtaining the centroid of the tray and the centroid of the object includes:

[0015] Acquire the pallet image and the object image by using an RGB-D camera in front of the transport forklift;

[0016] Performing contour extraction on the pallet image and the object image to obtain pallet contour points and object contour points;

[0017] The centroid of the pallet and the centroid of the object are obtained by calculation based on the pallet contour points and the object contour points.

[0018] Furthermore, when the centroid of the tray is offset from the centroid of the object in the vertical direction, the calculation formula of the offset angle is:

[0019] ;

[0020] in, represents the offset angle, represents the inverse tangent function, , and Respectively represent the components of the object's centroid on the x-axis, y-axis, and z-axis, , and They respectively represent the components of the centroid of the tray on the x-axis, y-axis and z-axis.

[0021] Furthermore, the calculation formula of the dumping risk index is:

[0022] ;

[0023] Wherein, TRI represents the tipping risk index, A represents the vibration amplitude, f represents the vibration frequency, h represents the relative height of the centroid of the object, and lg represents a logarithmic function with base 10.

[0024] Furthermore, the constraint function is:

[0025] ;

[0026] in, represents the constraint function, represents the speed of the movement, represents the acceleration, represents the turning radius, represents the steering speed, W represents the load weight, Indicates the area of ​​the pallet, represents the dumping risk index, and α, β, γ, δ and η represent the constraint coefficients of the constraint function.

[0027] Further, obtaining the optimal state of the individual forklift includes:

[0028] S401: Generate a population including N candidate solutions, and initialize the solution space of the candidate solutions;

[0029] S402: Calculate the fitness of each candidate solution through the constraint function, and select K candidate solutions with the largest fitness values ​​as parent individuals;

[0030] S403: exchanging variables between the two parent individuals using a single-point crossover method to generate offspring individuals with parent characteristics, and randomly adjusting the variable values ​​of the offspring individuals with a specific probability to obtain mutant offspring individuals;

[0031] S404: Incorporating the parent individual, the offspring individual and the mutant offspring individual into the new generation population, recalculating the fitness values ​​of all individuals in the new generation population, and selecting K individuals with the largest fitness values ​​to be retained for the next generation;

[0032] S405: Determine whether the algorithm meets the convergence condition. If not, return to S403; the convergence condition is that the change amplitude of the optimal solution in T consecutive generations is less than the change threshold or reaches the maximum number of iterations;

[0033] S406: Return the optimal solution and the objective function value corresponding to the optimal solution, and use the optimal solution as the optimal state of the individual forklift.

[0034] Further, obtaining the global optimal path includes:

[0035] The individual forklift broadcasts the optimal path and status information of its own individual forklift to other individual forklifts;

[0036] The individual forklift receiving the broadcast compares the optimal path of the individual forklift to which it belongs to detect potential time and space conflicts;

[0037] All forklifts collaborate and iteratively adjust the optimal path of their individual forklifts through a distributed optimization algorithm to achieve global path optimization;

[0038] When the distributed optimization algorithm converges, all forklifts obtain new individual forklift optimal paths, and finally obtain the global optimal path.

[0039] The present invention also proposes an intelligent collaborative dispatching system for transport forklifts for warehouse management, comprising:

[0040] A centroid acquisition module is used to acquire a pallet image and an image of an object above the pallet, and perform centroid detection on the pallet image and the object image to obtain a pallet centroid and an object centroid;

[0041] An offset calculation module, used to calculate an offset angle between the centroid of the tray and the centroid of the object in the vertical direction, and if the offset angle is greater than a first offset threshold, an alarm is issued and an alarm message is uploaded;

[0042] A tipping assessment module, used to obtain the vibration amplitude and vibration frequency of the forklift during operation through a vibration sensor under the fork arm, and calculate the tipping risk index according to the vibration amplitude, the vibration frequency and the relative height of the centroid of the object;

[0043] An individual forklift optimization module is used to establish a constraint function in combination with the dumping risk index and the forklift operating state, and optimize the constraint function through a genetic algorithm to obtain the optimal state of the individual forklift; the forklift operating state includes movement speed, acceleration, turning speed and turning radius;

[0044] The individual forklift path selection module combines the SLAM map and the optimal state of the individual forklift, performs path planning through the A* algorithm, and obtains the optimal path of the individual forklift from the starting point to the target point;

[0045] The global forklift path selection module is used for individual forklifts to broadcast their own individual forklift optimal path and status information to other individual forklifts, and all individual forklifts distribute and coordinate to adjust their own individual forklift optimal paths to obtain the global optimal path; the status information includes path information, task information and the forklift operating status.

[0046] Furthermore, the calculation formula of the dumping risk index is:

[0047] ;

[0048] Wherein, TRI represents the tipping risk index, A represents the vibration amplitude, f represents the vibration frequency, h represents the relative height of the centroid of the object, and lg represents a logarithmic function with base 10.

[0049] Furthermore, the constraint function is:

[0050] ;

[0051] in, represents the constraint function, represents the speed of the movement, represents the acceleration, represents the turning radius, represents the steering speed, W represents the load weight, Indicates the area of ​​the pallet, represents the dumping risk index, and α, β, γ, δ and η represent the constraint coefficients of the constraint function.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] 1. The present invention calculates the offset angle between the centroid of the pallet and the centroid of the object to monitor the stability of the cargo in the transportation process in real time. It can not only accurately determine whether the cargo is at risk of dumping, but also quickly issue an alarm and notify the system to take intervention measures when the risk first appears, thereby effectively avoiding the occurrence of cargo dumping accidents. By combining the dynamic monitoring of the offset angle and the threshold warning mechanism, the operation safety of the forklift in a complex operating environment is ensured.

[0054] 2. The present invention combines the dumping risk index and the operating status of the forklift, constructs a constraint function and uses a genetic algorithm for optimization, comprehensively considering key parameters such as the forklift's operating speed, acceleration, steering speed, and steering radius, not only realizing the global optimization of the operating status of individual forklifts, but also effectively reducing the decline in transportation efficiency and safety hazards caused by improper setting of operating parameters. At the same time, through the iterative evolution mechanism of the genetic algorithm, it can dynamically adapt to complex operating environments and task requirements, so that the forklift can maximize its transportation efficiency while ensuring safety.

[0055] 3. The present invention uses the optimal path of individual forklifts and distributed optimization algorithms to coordinate the collaborative operation of multiple forklifts in a dynamic environment, ensuring the efficiency and safety of global task allocation and path planning. Through the distributed optimization algorithm, each forklift can dynamically adjust its own path based on shared status information, avoiding path conflicts and resource preemption problems to the greatest extent, while reducing transportation delays and safety accidents caused by improper scheduling. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 A flow chart of a method for intelligent collaborative scheduling of transport forklifts for warehouse management provided by an embodiment of the present invention;

[0057] Figure 2 A flow chart of a method for obtaining the optimal state of an individual forklift provided by an embodiment of the present invention;

[0058] Figure 3 A structural diagram of an intelligent collaborative scheduling system for transport forklifts for warehouse management provided in an embodiment of the present invention. DETAILED DESCRIPTION

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

[0060] The first embodiment provided by the present invention is as follows:

[0061] A smart collaborative dispatching method for transport forklifts for warehouse management, such as Figure 1 As shown, including:

[0062] S100: Acquire a pallet image and an image of an object above the pallet, perform centroid detection on the pallet image and the object image, and obtain the pallet centroid and the object centroid;

[0063] Furthermore, obtaining the centroid of the tray and the centroid of the object includes:

[0064] Acquire the pallet image and the object image by using an RGB-D camera in front of the transport forklift;

[0065] Performing contour extraction on the pallet image and the object image to obtain pallet contour points and object contour points;

[0066] The centroid of the pallet and the centroid of the object are obtained by calculation based on the pallet contour points and the object contour points.

[0067] Specifically, an RGB-D camera is used to simultaneously capture RGB-D images of the pallet and the object to obtain three-dimensional information; the image coordinates are converted into a three-dimensional point cloud coordinate system through the camera internal parameters (focal length, principal point coordinates, etc.), and the three-dimensional point cloud data of the pallet and the object are generated. The edge contours of the pallet and the object are extracted using an edge detection algorithm, and the edge points are mapped into the depth map to obtain the corresponding three-dimensional contour point set, where the three-dimensional contour point set can be expressed as:

[0068] ;

[0069] Among them, P represents the 3D contour point set, n represents the number of contour points; the centroid coordinates of the pallet and the object are calculated through the 3D contour point data , the centroid coordinate calculation can be expressed as:

[0070] ;

[0071] ;

[0072] ;

[0073] in, and Represent the centroid coordinates of the x-axis, y-axis, and z-axis respectively, and Represent the coordinates of contour point i on the x-axis, y-axis, and z-axis respectively.

[0074] By acquiring images of pallets and objects through the RGB-D camera in front of the transport forklift and performing contour extraction on these images, the shape features of the pallet and objects can be accurately identified, thereby obtaining the contour points of the pallet and objects. Further calculation of the positions of these contour points can accurately determine the centroid positions of the pallet and objects, providing an important basis for subsequent object stability analysis, tilt risk assessment, and intelligent path planning, thereby improving the operating accuracy and transportation safety of the forklift.

[0075] S200: Calculating the vertical offset angle between the centroid of the tray and the centroid of the object, and if the offset angle is greater than a first offset threshold, issuing an alarm and uploading alarm information;

[0076] Furthermore, when the centroid of the tray is offset from the centroid of the object in the vertical direction, the calculation formula of the offset angle is:

[0077] ;

[0078] in, represents the offset angle, represents the inverse tangent function, , and Respectively represent the components of the object's centroid on the x-axis, y-axis, and z-axis, , and They respectively represent the components of the centroid of the tray on the x-axis, y-axis and z-axis.

[0079] Specifically, the offset angle in the vertical direction can be expressed as the angle between the vector between the centroid of the pallet and the centroid of the object and the z-axis. In a feasible real-time method, the offset angle data between the two sets of pallets and the object are shown in Table 1.

[0080] Table 1 Data table of the offset angle between the pallet and the object

[0081]

[0082] By calculating the vertical offset angle between the centroid of the pallet and the centroid of the object, the stability of the object on the pallet can be effectively monitored. When the offset angle exceeds the set threshold, it means that the object may be at risk of tilting or instability, so an alarm is issued and the alarm information is uploaded. This method can provide early warning of potential tipping risks, ensure that the forklift can predict the risk of cargo tipping before transportation, prevent object damage or operational accidents during transportation, and improve the safety and reliability of transportation operations.

[0083] S300: obtaining the vibration amplitude and vibration frequency of the forklift during operation through a vibration sensor under the fork arm, and calculating the tipping risk index according to the vibration amplitude, the vibration frequency and the relative height of the centroid of the object;

[0084] Furthermore, the calculation formula of the dumping risk index is:

[0085] ;

[0086] Wherein, TRI represents the tipping risk index, A represents the vibration amplitude, f represents the vibration frequency, h represents the relative height of the centroid of the object, and lg represents a logarithmic function with base 10.

[0087] Specifically, the relative height of the object's centroid can be expressed as:

[0088] ;

[0089] in, represents the component of the object's centroid in the z-axis direction, Represents the component of the centroid of the tray in the z-axis direction.

[0090] The vibration sensor under the fork arm monitors the vibration amplitude and frequency of the forklift in real time during operation, and calculates the tipping risk index based on the relative height of the object's centroid, which can effectively assess the risk of objects tilting or falling during the operation of the forklift. This real-time monitoring and risk assessment mechanism can provide early warning for forklift operation, adjust the operating status or take safety measures in a timely manner, and ensure the safety and stability of transportation operations.

[0091] S400: establishing a constraint function in combination with the tipping risk index and the forklift operation state, optimizing the constraint function by a genetic algorithm, and obtaining an optimal state of an individual forklift; the forklift operation state includes a movement speed, an acceleration, a turning speed, and a turning radius;

[0092] Furthermore, the constraint function is:

[0093] ;

[0094] in, represents the constraint function, represents the speed of the movement, represents the acceleration, represents the turning radius, represents the steering speed, W represents the load weight, Indicates the area of ​​the pallet, represents the dumping risk index, and α, β, γ, δ and η represent the constraint coefficients of the constraint function.

[0095] By combining the tipping risk index and the operating status of the forklift to establish a constraint function, the stability and safety of the forklift under different working conditions can be dynamically evaluated. This constraint function can combine the tipping risk with the motion characteristics of the forklift to ensure that the potential risk of object tipping is taken into account during path planning and scheduling, thereby limiting the operating speed and operation mode of the forklift in high-risk states.

[0096] Furthermore, the method for obtaining the optimal state of the individual forklift is as follows: Figure 2 As shown, including:

[0097] S401: Generate a population including N candidate solutions, and initialize the solution space of the candidate solutions;

[0098] S402: Calculate the fitness of each candidate solution through the constraint function, and select K candidate solutions with the largest fitness values ​​as parent individuals;

[0099] S403: exchanging variables between the two parent individuals using a single-point crossover method to generate offspring individuals with parent characteristics, and randomly adjusting the variable values ​​of the offspring individuals with a specific probability to obtain mutant offspring individuals;

[0100] S404: Incorporating the parent individual, the offspring individual and the mutant offspring individual into the new generation population, recalculating the fitness values ​​of all individuals in the new generation population, and selecting K individuals with the largest fitness values ​​to be retained for the next generation;

[0101] S405: Determine whether the algorithm meets the convergence condition. If not, return to S403; the convergence condition is that the change amplitude of the optimal solution in T consecutive generations is less than the change threshold or reaches the maximum number of iterations;

[0102] S406: Return the optimal solution and the objective function value corresponding to the optimal solution, and use the optimal solution as the optimal state of the individual forklift.

[0103] Specifically, N candidate solutions are randomly generated, each of which is a vector containing the optimization variables , the initialized solution space is shown in Table 2; using the constraint function As the basis of the fitness function, and calculate the fitness value of each candidate solution x, the fitness value is defined as:

[0104] ;

[0105] in, represents the fitness value function, Represents the constraint function; selects the K candidate solutions with the largest fitness values ​​as parent individuals; uses the single-point crossover method to exchange variables between two parent individuals to generate offspring individuals with parent characteristics, and randomly adjusts the variable values ​​of the offspring individuals with a specific probability to obtain mutant offspring individuals. In a feasible implementation scheme, the parent individuals , parent individual , for variables Adjust to get new offspring individuals and offspring individuals Then, the variable value of the offspring is randomly adjusted with a probability of 50% to obtain the mutant offspring individual and variant offspring individuals ; Include the parent individuals, offspring individuals and mutant offspring individuals into the new generation population, recalculate the fitness values ​​of all individuals in the new generation population, and select the K individuals with the largest fitness values ​​to be retained for the next generation; when the individual with the largest fitness value in the population changes very little in several consecutive generations (such as 10 generations), or reaches the maximum number of generations 20, the algorithm stops the algorithm iteration, otherwise re-execute S403; when the algorithm stops iterating, output the optimal solution And the objective function .

[0106] Table 2 Initialized solution space setting table

[0107]

[0108] The genetic algorithm simulates the process of natural selection and gradually adjusts the operating parameters of the forklift, such as speed, acceleration, steering angle and radius, to minimize the risk of tipping and other constraints. Through this intelligent optimization method, the forklift can automatically adapt to different work requirements in a complex transportation environment, achieve more accurate path planning and task execution, and significantly improve work efficiency and safety.

[0109] S500: combining the SLAM map and the optimal state of the individual forklift, performing path planning through the A* algorithm, and obtaining the optimal path of the individual forklift from the starting point to the target point;

[0110] S600: The individual forklift broadcasts its own optimal path and status information to other individual forklifts, and all individual forklifts coordinately and distributedly adjust their own optimal paths to obtain a global optimal path; the status information includes path information, task information and the forklift operation status.

[0111] Further, obtaining the global optimal path includes:

[0112] The individual forklift broadcasts the optimal path and status information of its own individual forklift to other individual forklifts;

[0113] The individual forklift receiving the broadcast compares the optimal path of the individual forklift to which it belongs to detect potential time and space conflicts;

[0114] All forklifts collaborate and iteratively adjust the optimal path of their individual forklifts through a distributed optimization algorithm to achieve global path optimization;

[0115] When the distributed optimization algorithm converges, all forklifts obtain new individual forklift optimal paths, and finally obtain the global optimal path.

[0116] Specifically, each forklift broadcasts its own path and status information through a wireless communication module (such as Wi-Fi, ZigBee); each forklift receives the broadcast information of other forklifts, checks the time conflict (two vehicles arrive at the same location at the same time) and space conflict (collision risk when the two vehicles turn) of the path; each forklift uses a distributed optimization algorithm (such as a multi-agent path planning algorithm or Nash equilibrium) to adjust the path to ensure that the conflict is minimized and the global total cost is optimized. After the algorithm converges, all forklifts use the final adjusted path to ensure the global path is optimal. In a feasible embodiment, the optimal path and status information of the two individual forklifts are shown in Table 3. Forklift 1 and forklift 2 will meet at point (10,12) after 8 seconds. In order to avoid collision between the two vehicles, the path of forklift 2 is updated after collaborative adjustment through the distributed optimization algorithm, and the risk of collision is avoided at this time.

[0117] Table 3 Optimal path and status information of individual forklifts

[0118]

[0119] By broadcasting its own optimal path and status information to other forklifts, all forklifts can share their own paths and operating status in real time, thereby making distributed collaborative adjustments on a global scale. This information sharing mechanism enables each forklift to optimize its own path selection based on the operating data and path planning of other forklifts, avoid potential path conflicts and traffic congestion, and ultimately obtain the global optimal path planning. Through this collaborative adjustment, not only the overall operating efficiency is improved, but also the conflicts and delays in the transportation process are effectively reduced, ensuring that multiple forklifts can work together efficiently and safely in a warehousing environment.

[0120] The second embodiment provided by the present invention is as follows:

[0121] A logistics company A has a large warehouse scale, with an average daily cargo throughput of thousands of tons. As the business scope increases, some goods cannot be packed in standard cubic cartons. The current unmanned forklift dispatching system is difficult to achieve safe and effective cargo transportation, resulting in frequent dumping during transportation. In order to cope with complex warehousing and transportation needs, Company A has adopted an intelligent collaborative dispatching system for transport forklifts for warehousing management, such as Figure 3 As shown, including:

[0122] A centroid acquisition module is used to acquire a pallet image and an image of an object above the pallet, and perform centroid detection on the pallet image and the object image to obtain a pallet centroid and an object centroid;

[0123] An offset calculation module, used to calculate an offset angle between the centroid of the tray and the centroid of the object in the vertical direction, and if the offset angle is greater than a first offset threshold, an alarm is issued and an alarm message is uploaded;

[0124] A tipping assessment module, used to obtain the vibration amplitude and vibration frequency of the forklift during operation through a vibration sensor under the fork arm, and calculate the tipping risk index according to the vibration amplitude, the vibration frequency and the relative height of the centroid of the object;

[0125] An individual forklift optimization module is used to establish a constraint function in combination with the dumping risk index and the forklift operating state, and optimize the constraint function through a genetic algorithm to obtain the optimal state of the individual forklift; the forklift operating state includes movement speed, acceleration, turning speed and turning radius;

[0126] The individual forklift path selection module combines the SLAM map and the optimal state of the individual forklift, performs path planning through the A* algorithm, and obtains the optimal path of the individual forklift from the starting point to the target point;

[0127] The global forklift path selection module is used for individual forklifts to broadcast their own individual forklift optimal path and status information to other individual forklifts, and all individual forklifts distribute and coordinate to adjust their own individual forklift optimal paths to obtain the global optimal path; the status information includes path information, task information and the forklift operating status.

[0128] Furthermore, the calculation formula of the dumping risk index is:

[0129] ;

[0130] Wherein, TRI represents the tipping risk index, A represents the vibration amplitude, f represents the vibration frequency, h represents the relative height of the centroid of the object, and lg represents a logarithmic function with base 10.

[0131] Furthermore, the constraint function is:

[0132] ;

[0133] in, represents the constraint function, represents the speed of the movement, represents the acceleration, represents the turning radius, represents the steering speed, W represents the load weight, Indicates the area of ​​the pallet, represents the dumping risk index, and α, β, γ, δ and η represent the constraint coefficients of the constraint function.

[0134] As shown in Table 4, after adopting the intelligent collaborative scheduling system for transport forklifts for warehouse management proposed by the present invention, the number of cargo dumping has been reduced from the original 19 / month to the current 4 / month, and the transportation time per 100 tons of cargo has been shortened from the original 1.5 hours to the current 1.2 hours. In addition, the number of forklift accidents per month has been reduced from the previous 5 / month to the current 2 / month. These achievements are mainly attributed to the multiple innovative designs of the system. First, through the RGB-D camera and vibration sensor, the cargo centroid deviation angle and dumping risk index are detected in real time, and an alarm can be issued and the forklift operation status can be adjusted when the dumping risk first appears, which greatly improves the stability and safety of the transportation process; secondly, the operation status of individual forklifts is optimized through genetic algorithms to ensure that the forklifts achieve the optimal configuration in key parameters such as operating speed, acceleration, turning radius and turning speed, so as to balance efficiency and safety; finally, combined with the path planning of SLAM maps and A* algorithms, and the global scheduling capability of distributed collaborative optimization algorithms, the system effectively reduces delays and accidents caused by path conflicts or resource preemption during the transportation of forklifts.

[0135] By obtaining the centroid of pallets and objects in real time, calculating the offset angle and combining the vibration sensor data to calculate the tipping risk index, it is possible to issue early warnings and adjust the operating status in a timely manner, significantly improving the safety and stability of the transportation process. At the same time, the operating status of individual forklifts is optimized based on genetic algorithms to achieve a dynamic balance of speed, acceleration and steering parameters. Combining the path planning of SLAM maps and A* algorithms, as well as the global scheduling function of distributed collaborative optimization algorithms, the system not only effectively reduces delays and accidents caused by path conflicts and resource preemption, but also greatly improves the transportation efficiency of warehouse management, achieving efficient collaboration from individual optimization to global optimization.

[0136] Table 4 Comparison of the intelligent collaborative dispatching system for transport forklifts before and after application

[0137]

[0138] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent collaborative dispatching of transport forklifts for warehouse management, characterized in that: include: Acquire a pallet image and an image of an object above the pallet, perform centroid detection on the pallet image and the object image, and obtain the pallet centroid and the object centroid; Calculating the vertical offset angle between the centroid of the tray and the centroid of the object, and if the offset angle is greater than a first offset threshold, issuing an alarm and uploading alarm information; The vibration amplitude and vibration frequency of the forklift during operation are obtained by a vibration sensor under the fork arm, and a tipping risk index is calculated according to the vibration amplitude, the vibration frequency and the relative height of the centroid of the object; A constraint function is established by combining the tipping risk index and the operation state of the forklift, and the constraint function is optimized by a genetic algorithm to obtain the optimal state of the individual forklift; the operation state of the forklift includes movement speed, acceleration, turning speed and turning radius; Combining the SLAM map and the optimal state of the individual forklift, the path planning is performed through the A* algorithm to obtain the optimal path of the individual forklift from the starting point to the target point; The individual forklift broadcasts its own optimal path and status information to other individual forklifts, and all individual forklifts coordinately adjust their own optimal path to obtain the global optimal path; the status information includes path information, task information and the operation status of the forklift; When the centroid of the tray is offset from the centroid of the object in the vertical direction, the calculation formula of the offset angle is: Wherein, θ represents the deviation angle, arctan represents the inverse tangent function, and x obj ,y obj and z obj Respectively represent the components of the object's centroid on the x-axis, y-axis, and z-axis, x tray ,y tray and z tray Respectively represent the components of the centroid of the tray on the x-axis, y-axis and z-axis; The calculation formula of the dumping risk index is: Wherein, TRI represents the tipping risk index, A represents the vibration amplitude, f represents the vibration frequency, h represents the relative height of the centroid of the object, and lg represents a logarithmic function with base 10.

2. The intelligent collaborative dispatching method for transport forklifts for warehouse management according to claim 1 is characterized in that: Getting the centroid of the tray and the object includes: Acquire the pallet image and the object image by using an RGB-D camera in front of the transport forklift; Performing contour extraction on the pallet image and the object image to obtain pallet contour points and object contour points; The centroid of the pallet and the centroid of the object are obtained by calculation based on the pallet contour points and the object contour points.

3. The intelligent collaborative dispatching method for transport forklifts for warehouse management according to claim 1 is characterized in that , the constraint function is: Wherein, F represents the constraint function, v represents the movement speed, a represents the acceleration, R represents the turning radius, ω represents the turning speed, W represents the load weight, and A s represents the pallet area, TRI represents the tipping risk index, and α, β, γ, δ and η represent the constraint coefficients of the constraint function.

4. The intelligent collaborative dispatching method for transport forklifts for warehouse management according to claim 1 is characterized in that: Obtaining the optimal state of the individual forklift includes: S401: Generate a population including N candidate solutions, and initialize the solution space of the candidate solutions; S402: Calculate the fitness of each candidate solution through the constraint function, and select K candidate solutions with the largest fitness values ​​as parent individuals; S403: exchanging variables between the two parent individuals using a single-point crossover method to generate offspring individuals with parent characteristics, and randomly adjusting the variable values ​​of the offspring individuals with a specific probability to obtain mutant offspring individuals; S404: Incorporating the parent individual, the offspring individual and the mutant offspring individual into the new generation population, recalculating the fitness values ​​of all individuals in the new generation population, and selecting K individuals with the largest fitness values ​​to be retained for the next generation; S405: Determine whether the algorithm meets the convergence condition. If not, return to S403; the convergence condition is that the change amplitude of the optimal solution in T consecutive generations is less than the change threshold or reaches the maximum number of iterations; S406: Return the optimal solution and the objective function value corresponding to the optimal solution, and use the optimal solution as the optimal state of the individual forklift.

5. The intelligent collaborative dispatching method for transport forklifts for warehouse management according to claim 1 is characterized in that: Obtaining the global optimal path includes: The individual forklift broadcasts the optimal path and status information of its own individual forklift to other individual forklifts; The individual forklift receiving the broadcast compares the optimal path of the individual forklift to which it belongs to detect potential time and space conflicts; All forklifts collaborate and iteratively adjust the optimal path of their individual forklifts through a distributed optimization algorithm to achieve global path optimization; When the distributed optimization algorithm converges, all forklifts obtain new individual forklift optimal paths, and finally obtain the global optimal path.

6. A transport forklift intelligent collaborative scheduling system for warehouse management, used to implement a transport forklift intelligent collaborative scheduling method for warehouse management as described in any one of claims 1 to 5, characterized in that: include: A centroid acquisition module is used to acquire a pallet image and an image of an object above the pallet, and perform centroid detection on the pallet image and the object image to obtain a pallet centroid and an object centroid; An offset calculation module, used to calculate an offset angle between the centroid of the tray and the centroid of the object in the vertical direction, and if the offset angle is greater than a first offset threshold, an alarm is issued and an alarm message is uploaded; A tipping assessment module, used to obtain the vibration amplitude and vibration frequency of the forklift during operation through a vibration sensor under the fork arm, and calculate a tipping risk index according to the vibration amplitude, the vibration frequency and the relative height of the centroid of the object; An individual forklift optimization module is used to establish a constraint function in combination with the dumping risk index and the forklift operating state, and optimize the constraint function through a genetic algorithm to obtain the optimal state of the individual forklift; the forklift operating state includes movement speed, acceleration, turning speed and turning radius; The individual forklift path selection module combines the SLAM map and the optimal state of the individual forklift, performs path planning through the A* algorithm, and obtains the optimal path of the individual forklift from the starting point to the target point; The global forklift path selection module is used for individual forklifts to broadcast their own individual forklift optimal path and status information to other individual forklifts, and all individual forklifts distribute and coordinate to adjust their own individual forklift optimal paths to obtain the global optimal path; the status information includes path information, task information and the forklift operating status.

Images