Genetic algorithm-based multi-dispensing-head cooperative working path planning method and dispensing equipment

Abstract

The invention discloses a multi-dispensing-head cooperative work path planning method based on a genetic algorithm and dispensing equipment. The method comprises the steps that S1, a task point set and a dispensing head parameter set are generated; s2, initializing an initial population of a genetic algorithm; s3, generating a new path planning individual; s4, performing local optimization on the new path planning individual based on a black-wing optimization algorithm, and further optimizing the path planning scheme; s5, generating a path planning scheme with no path conflict; s6, judging whether the path planning population with no path conflict meets a convergence condition or not, outputting an optimal path planning scheme, and if the path planning population does not meet the convergence condition, returning to execute the steps S3 to S5; and S7, cooperative work of the dispensing task is completed. According to the method, a new optimized path can be generated in a shorter time when a sudden task, path change or equipment failure is handled, so that the flexible production capacity of cooperative work of multiple rubber heads is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of glue dispensing machines, and in particular to a method for planning a collaborative working path of multiple glue dispensing heads based on a genetic algorithm and a glue dispensing device. Background Art

[0002] With the development of industrial automation technology, dispensing equipment has been widely used in the fields of electronic manufacturing and automotive parts assembly. As one of the core technologies of dispensing equipment, the collaborative path planning of multiple dispensing heads has a direct impact on production efficiency and product quality. In the dispensing operation, multiple dispensing heads need to process multiple task points at the same time. The efficiency and accuracy of path planning are the key to achieving efficient collaborative operations.

[0003] At present, multi-point glue head path planning mainly relies on static planning and traditional optimization algorithms. Traditional methods have certain feasibility when dealing with fixed scenes and simple task points. For example, the static planning method assigns fixed task points to each glue head by pre-designing a fixed path. However, when the number of task points is large and the distribution is complex, static planning cannot dynamically respond to changes in actual operations, which easily leads to problems such as long paths and frequent path conflicts between glue heads. In addition, traditional optimization algorithms, such as those based on exhaustive search or gradient optimization, are usually difficult to handle multi-objective optimization problems. They are easy to fall into local optimal solutions in path planning and lack global search capabilities.

[0004] In recent years, some intelligent optimization algorithms have begun to be applied to the field of path planning, such as ant colony algorithms and genetic algorithms. Ant colony algorithms and genetic algorithms have improved the global optimization capabilities of path planning to a certain extent. However, when applied to multi-point glue head collaborative work scenarios, the following problems are still exposed:

[0005] The path conflict problem is prominent: the existing methods usually ignore the spatial constraints of the collaborative work of multiple rubber heads during path planning, resulting in cross-conflicts in the paths of different rubber heads, affecting work efficiency and equipment safety.

[0006] Uneven task distribution problem: The existing technology cannot effectively balance the workload of the rubber heads. It is easy for some rubber heads to be overloaded while other rubber heads are idle, resulting in low resource utilization.

[0007] Lack of dynamic adaptability: When the distribution of task points or the status of equipment changes, existing methods cannot quickly adjust path planning, resulting in insufficient flexibility in the production process and difficulty in coping with complex and changing industrial environments.

[0008] In summary, the existing technology has significant deficiencies in path planning efficiency, path conflict avoidance and task allocation balance, and cannot meet the needs of modern industrial dispensing equipment for efficient and dynamic collaborative work. A new method combined with intelligent optimization algorithm is urgently needed to solve the above problems and improve the overall performance and adaptability of multi-dispensing head path planning. Summary of the invention

[0009] One purpose of the present invention is to propose a path planning method and dispensing equipment for collaborative work of multiple dispensing heads based on genetic algorithms. The present invention can generate a new optimized path in a shorter time when dealing with sudden tasks, path changes or equipment failures, thereby ensuring the flexible production capacity of collaborative work of multiple dispensing heads.

[0010] According to an embodiment of the present invention, a method for planning a collaborative working path of multiple glue heads based on a genetic algorithm comprises the following steps:

[0011] S1. Obtain the spatial coordinates of the task points to be processed of the multi-dispensing glue head dispensing task and the parameter information of the dispensing task, and generate a task point set and a dispensing head parameter set;

[0012] S2. Initialize the initial population of the genetic algorithm, the initial population is composed of a plurality of path planning individuals, each path planning individual includes a task point allocation scheme and a dispensing path planning scheme, and in the initialization process, the task point set and the dispensing head parameter set are used to preferentially allocate task points with higher attractiveness to the appropriate dispensing head by combining the attractiveness calculation method of the black-winged kite optimization algorithm;

[0013] S3. Perform genetic algorithm operations on the initial population in S2 to generate new path planning individuals;

[0014] S4. Based on the black kite optimization algorithm, the new path planning individual is locally optimized, the black kite optimization algorithm is used to simulate the prey tracking behavior of the black kite, the task point allocation is optimized by dynamically adjusting the attractiveness of the task points in the task point set, and the path planning scheme is further optimized;

[0015] S5. Perform path conflict detection on the optimized path planning individuals, identify path intersection conflict points based on the task point set and the dispensing head parameter set, and adjust the conflicting individuals by path redistribution to generate a path planning solution without path conflict;

[0016] S6. Determine whether the path planning population without conflict meets the convergence condition, wherein the convergence condition includes that the maximum number of iterations reaches a preset threshold or the fitness value of the population reaches a target threshold. When the path planning population meets the convergence condition, the optimal path planning scheme is output. If the path planning population does not meet the convergence condition, the execution of steps S3 to S5 is returned.

[0017] S7. Input the final optimized path planning scheme into the control module of the dispensing equipment, and use the dispensing task parameter information to drive multiple dispensing heads to perform the dispensing task according to the optimized path to complete the collaborative work of the dispensing task.

[0018] Optionally, the S1 specifically includes the following steps:

[0019] S11. Obtain the spatial coordinates of the task point to be processed, where the spatial coordinates are the absolute positions of the task point in three-dimensional space:

[0020] P={P 1 ,P 2 ,…,P n};

[0021] Among them, P i =(x i ,y i ,z i ), x i ,y i ,z i Represent the task point P i x, y, z coordinates in three-dimensional space;

[0022] S12. Obtain parameter information of the dispensing head, the parameter information including the number of dispensing heads m, the dispensing head movement speed set V, the dispensing head maximum load capacity set Q and the dispensing task execution priority set U;

[0023] Dispensing head movement speed set V in v j Indicates the movement speed of the dispensing head j;

[0024] Maximum load capacity of dispensing head set Q in q j Indicates the maximum load capacity of the dispensing head j;

[0025] The execution priority set of the dispensing task is u in i Represents the task point P i The higher the priority of the task point P i Will be given priority in path planning;

[0026] S13. Generate a task point set based on the spatial coordinate set P of the task point and the parameter information of the dispensing task. Each element of the task point set represents a combination of the coordinates of the task point and its dispensing task execution priority:

[0027] P task ={(P i ,u i )|i=1,2,…,n};

[0028] S14. Construct a dispensing head parameter set, where each element of the dispensing head parameter set represents a combination of the movement speed and the maximum load capacity of the dispensing head:

[0029] H={(v j ,q j )|j=1,2,…,m}.

[0030] Optionally, S2 includes the following steps:

[0031] S21. Calculate the attractiveness of the task points. Based on the attractiveness model of the black kite optimization algorithm, the task point set P task The task point P in i Calculate attraction α i , the attraction α i By the task point priority u i , the distance between the task point and the dispensing head d ij and dispensing head load capacity q j Sure:

[0032]

[0033] Among them, t represents the current iteration number, u i Point P for the task i The priority of d ij (t) is the current task point P i To the position x of the dispensing head j at iteration t j (t),y j (t),z j (t), q j is the maximum load capacity of the dispensing head j, v j (t) is the speed of the dispensing head j at iteration t, which indicates the influence of the dispensing head movement on the attraction, γ is the load weight adjustment factor, which controls the influence of the dispensing head load on the attraction, β is the speed weight adjustment factor, which controls the influence of the dispensing head movement speed on the attraction, ∈ is a very small positive number, which is used to avoid the denominator being zero;

[0034] S22. The attraction strategy of the black kite optimization algorithm is used to simulate the black kite's preferential allocation behavior to high-attractive targets during prey tracking. i ,u i ) and dispensing head (v j ,q j )'s attractiveness value α i (t) Calculate and sort by attraction value α i(t) Prioritize the allocation of task points with an attraction higher than a preset value to the dispensing head with the largest attraction, and dynamically adjust the remaining load capacity of the dispensing head during the allocation process so that the allocation of each dispensing head does not exceed its maximum load capacity q j ;

[0035] S23. Based on the relationship between the assigned task points and the glue dispensing heads, an initial path planning individual is constructed. The path planning individual includes the task point sequence and glue dispensing path of each glue dispensing head. The attack strategy of the black kite optimization algorithm is applied to the initial path planning individual to locally optimize the sequence and allocation plan of the task points.

[0036] Optionally, the S23 specifically includes the following steps:

[0037] S231. Based on the distribution relationship between the task points and the dispensing heads, generate the initial task point sequence S for each dispensing head j , the task point sequence represents the set of task points that the dispensing head j needs to visit in sequence;

[0038] S232. Calculate the total length L of each dispensing head path in the initial path planning individual j , the total path length L j Indicates that the dispensing head j follows the task point sequence S j The total path length L when executing the task j ;

[0039] S233. Apply the attack strategy of the black kite optimization algorithm to locally optimize the initial path planning individual. The optimization process includes:

[0040] Path exchange optimization: task point sequence S at dispensing head j j In the example above, the order of the two task points is exchanged and the total length of the path after the exchange is calculated. ′ j , if L ′ j <L j , then update the task point sequence to the optimized order.

[0041] Task point redistribution: based on dispensing head load capacity q j And the task point attraction α i (t), for the task point sequence S j Reallocate the task points with low attraction;

[0042] S234. Iterate and perform local optimization, setting the number of optimization iterations T opt and convergence condition, when the total path length of all dispensing heads is When it stops decreasing or reaches the maximum number of iterations, the optimized path planning individual is output;

[0043] S235. Outputting the optimized path planning individual, the path planning individual includes the task point sequence S of each dispensing head j and dispensing path, the optimized path planning individual meets the requirements of minimizing the path length and balancing the load of the dispensing head.

[0044] Optionally, S3 specifically includes the following steps:

[0045] S31. Select path planning individuals based on fitness values, select path planning individuals with fitness values ​​higher than a preset threshold from the path planning population to enter the next generation, and the fitness value F is calculated based on the path length L of the path planning individual j , path conflict cost C conflict Load balance of dispensing head imbalance Defined as:

[0046]

[0047] Among them, w 1 ,w 2 ,w 3 is the weight coefficient, Indicates the total length of all dispensing head paths, C conflict represents the path conflict cost, U imbalance Indicates the dispensing head load imbalance;

[0048] S32. Randomly select the task point sequences of two path planning individuals for partial mapping and crossover, retain the relative order of the task point sequences to generate a new task point allocation scheme, map the crossover task point allocation scheme to the path planning population to generate a new path planning individual, and retain the elite individual with the highest fitness value from participating in the crossover operation to maintain the stability of the optimal solution of the population;

[0049] S33. Randomly select a sequence from the task point sequence of the new path planning individual and arrange it in reverse order, and calculate the path length L after mutation ′ j , if L ′ j <L j , then update the path planning individual, randomly exchange the positions of two task points in the task point sequence of the new path planning individual, recalculate the fitness value F', if F'>F, retain the mutation result;

[0050] S34. Re-evaluate the fitness value of the mutated path planning population and remove new path planning individuals whose fitness value is lower than the threshold.

[0051] Optionally, the S4 specifically includes the following steps:

[0052] S41. Based on the task point set Ptask And the dispensing head parameter set H for the new path planning individual task point P k ∈P task Calculate the multi-target attraction value α k (t), the attractiveness value combines priority, distance, load balance and path conflict risk and is calculated as:

[0053]

[0054] Among them, u k Point P for the task k The priority of d kj (t) is the task point P k The Euclidean distance to the dispensing head j at iteration t, q j is the maximum load capacity of the dispensing head j, R k,j (t) is the path conflict risk factor, indicating that the task point P k The probability of a possible conflict with the current path of dispensing head j is calculated by the frequency of historical path intersections, γ 1 is the load adjustment factor, which controls the impact of load capacity, δ is the path conflict risk adjustment factor, which reduces the impact of path conflict on attractiveness;

[0055] S42. Simulate the collaborative prey capture behavior of black-winged kites, intelligently focus on the task points with attraction higher than the threshold, and dynamically adjust the weight of the dispensing head to the task points with high attraction:

[0056]

[0057] Among them, ω k,j (t) is the distance between dispensing head j and task point P k The allocation weight, η is the dynamic adjustment factor, The average attraction value of the task point set is used to increase the priority of the dispensing head for high attraction task points;

[0058] S43. Update the current position X of the path planning individual according to the intelligent focusing strategy combined with the attack and gliding behavior of the black kite optimization algorithm i (t+1) and speed V i (t+1):

[0059] V i (t+1)=λ·V i (t)+c 1 ·r 1 ·(X best -X i (t))+c 2 ·r 2 ·(X gbest -X i(t))+φ·ΔX cluster ;

[0060] X i (t+1)=X i (t)+V i (t+1);

[0061] Where λ is the momentum coefficient, c 1 ,c 2 is the learning factor, r 1 ,r 2 is a random number, X best is the optimal position of the individual in the current path planning, X gbest is the global optimal path planning solution, ΔX cluster is the aggregation center offset of the high-attractive task points within the population, which is used to simulate the flocking behavior of black-winged kites; φ is the focusing behavior adjustment factor, which controls the degree of concentration of individuals in the population on the high-attractive task points;

[0062] S44. According to the change of the attractiveness of the task points in the path planning scheme, the following two aspects are collaboratively optimized:

[0063] Load balancing optimization is to dynamically adjust the task point allocation to make the workload of the dispensing head close to balanced;

[0064] Path reordering optimization is to dynamically reorder non-critical task points in the task point sequence to reduce path length and path conflicts.

[0065] S45. Calculate the global attraction variance and the total length of path planning after each iteration. If the attraction variance is less than the set threshold or reaches the maximum number of iterations, output the optimized path planning solution; otherwise, return to execute steps S42 to S44.

[0066] Optionally, the S5 specifically includes the following steps:

[0067] S51. Based on the individuals in the optimized path planning scheme, time and space conflict detection is performed on the task points in the dispensing head path:

[0068]

[0069] Among them, C conflict (j,k) indicates the dispensing heads j and k at the task point P i The path conflict cost, T j (i) and T k (i) Dispensing heads j and k at task point P i The occupied time interval, is the indicator function, if two time intervals overlap, the conflict cost increases;

[0070] A path conflict matrix C is generated by detecting the conflicts of task points in all dispensing head paths, and the matrix elements of the path conflict matrix C represent the conflict intensity between the dispensing heads;

[0071] S52. Identify the conflict points corresponding to the non-zero elements in the path conflict matrix C, and select the task point P with the largest conflict intensity. c , based on the task point priority u c Adjust the priority to ensure the path execution order of high priority task points, and record the relevant glue head information of the conflict point. c ,k c ;

[0072] S53. Conflict point P c The task allocation is re-optimized, and the execution dispensing head j' of the conflict point is reselected based on the load capacity and attraction value of the dispensing head:

[0073]

[0074] Among them, α c (j) is the distance between dispensing head j and task point P c The attractiveness, q j is the maximum load capacity of the dispensing head j, w j is the current load of dispensing head j, indicating that the dispensing head with the largest remaining load capacity after allocation will be executed first;

[0075] S54. Update the path sequence of the dispensing head related to the conflict point based on the reallocated task point, adjust the path execution order, so that the execution time of the conflict point does not overlap with other task points, and the total path length meets the optimization goal:

[0076]

[0077] in, Click i for the task x and i x+1 The distance between Click i for the task k-1 The distance from the conflict point c, For conflict point c and task point i k+1 distance.

[0078] S55. Repeat the path conflict detection, conflict point identification and path reallocation until all elements of the path conflict matrix C are zero or the maximum number of iterations is reached, and output a multi-point glue head collaborative working path planning solution with no path conflicts.

[0079] A glue dispensing device comprises an internally arranged processor and memory, wherein the memory stores computer instructions, and when the processor runs the computer instructions, a method for planning a collaborative working path of multiple glue dispensing heads based on a genetic algorithm is executed.

[0080] The beneficial effects of the present invention are:

[0081] (1) The present invention combines the dynamic hunting behavior of the black kite optimization algorithm with the genetic algorithm to perform multi-objective optimization on the attractiveness of the path planning individuals. Taking into account the priority of the task points, the probability of path intersection conflicts between rubber heads, and the load capacity of the rubber heads, the multi-objective attractiveness function and intelligent focusing strategy are proposed. The path conflict points can be detected and dynamically adjusted in real time, and the path conflict problem can be effectively avoided by path redistribution and task point sequence optimization, thereby ensuring the safety of collaborative operations between rubber heads and the global optimization of path planning.

[0082] (2) The present invention is based on the collaborative work of genetic algorithm and black kite optimization algorithm. It realizes load balancing of multi-glue head task allocation by dynamically adjusting the task allocation weights of the glue dispensing heads. In the task allocation process, an intelligent focusing strategy and a multi-dimensional collaborative optimization method are adopted to give priority to allocating task points with higher attractiveness, and dynamically adjust the remaining load capacity of the glue heads, so that the workload of each glue head tends to be balanced, avoiding overload operation or idleness.

[0083] (3) The present invention introduces a dynamic parameter adjustment mechanism and a path conflict matrix model to enable the path planning scheme to quickly respond to changes in the distribution of task points or abnormal conditions in the rubber head state. The fluttering and gliding behaviors of the black kite optimization algorithm are combined with a dynamic attraction model, so that path planning can achieve real-time adjustment and adaptive optimization in complex industrial environments. When responding to sudden tasks, path changes or equipment failures, new optimized paths can be generated in a shorter time, thereby ensuring the flexible production capacity of multiple rubber heads working in collaboration. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0085] Figure 1 The present invention provides a method for planning a collaborative working path for multiple glue dispensing heads based on a genetic algorithm and a flow chart of a glue dispensing device. DETAILED DESCRIPTION

[0086] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0087] refer to Figure 1 , a multi-point glue head collaborative working path planning method based on genetic algorithm, comprising the following steps:

[0088] S1. Obtain the spatial coordinates of the task points to be processed of the multi-dispensing glue head dispensing task and the parameter information of the dispensing task, and generate a task point set and a dispensing head parameter set;

[0089] S2. Initialize the initial population of the genetic algorithm, which is composed of multiple path planning individuals. Each path planning individual includes a task point allocation scheme and a dispensing path planning scheme. During the initialization process, the task point set and the dispensing head parameter set are used to preferentially allocate task points with higher attractiveness to the appropriate dispensing head by combining the attractiveness calculation method of the black-winged kite optimization algorithm;

[0090] S3. Perform genetic algorithm operations on the initial population in S2 to generate new path planning individuals;

[0091] S4. Based on the black kite optimization algorithm, the new path planning individual is locally optimized, the black kite optimization algorithm is used to simulate the prey tracking behavior of the black kite, the task point allocation is optimized by dynamically adjusting the attractiveness of the task points in the task point set, and the path planning scheme is further optimized;

[0092] S5. Perform path conflict detection on the optimized path planning individuals, identify path intersection conflict points based on the task point set and the dispensing head parameter set, and adjust the conflicting individuals by path redistribution to generate a path planning solution without path conflict;

[0093] S6. Determine whether the path planning population without conflict meets the convergence condition, the convergence condition includes that the maximum number of iterations reaches a preset threshold or the population fitness value reaches a target threshold. When the path planning population meets the convergence condition, the optimal path planning solution is output. If the path planning population does not meet the convergence condition, return to execute steps S3 to S5;

[0094] S7. Input the final optimized path planning scheme into the control module of the dispensing equipment, and use the dispensing task parameter information to drive multiple dispensing heads to perform the dispensing task according to the optimized path to complete the collaborative work of the dispensing task.

[0095] In this implementation, S1 specifically includes the following steps:

[0096] S11. Obtain the spatial coordinates of the task point to be processed, which are the absolute positions of the task point in three-dimensional space:

[0097] P={P 1 ,P 2 ,…,P n};

[0098] Among them, Pi =(x i ,y i ,z i ), x i ,y i ,z i Represent the task point P i x, y, z coordinates in three-dimensional space;

[0099] S12. Obtain parameter information of the dispensing head, the parameter information includes the number of dispensing heads m, the dispensing head movement speed set V, the dispensing head maximum load capacity set Q and the dispensing task execution priority set U;

[0100] Dispensing head movement speed set V in v j Indicates the movement speed of the dispensing head j;

[0101] Maximum load capacity of dispensing head set Q in q j Indicates the maximum load capacity of the dispensing head j;

[0102] The execution priority set of the dispensing task is u in i Represents the task point P i The higher the priority of the task point P i Will be given priority in path planning;

[0103] S13. Generate a task point set based on the spatial coordinate set P of the task point and the parameter information of the dispensing task. Each element of the task point set represents a combination of the coordinates of the task point and its dispensing task execution priority:

[0104] P task ={(P i ,u i )|i=1,2,…,n};

[0105] S14. Construct a dispensing head parameter set, where each element of the dispensing head parameter set represents a combination of the movement speed and the maximum load capacity of the dispensing head:

[0106] H={(v j ,q j )|j=1,2,…,m}.

[0107] In this implementation, S2 includes the following steps:

[0108] S21. Calculate the attractiveness of the task points. Based on the attractiveness model of the black kite optimization algorithm, the task point set P task The task point P in i Calculate attraction α i , attraction α i By the task point priority u i, the distance between the task point and the dispensing head d ij and dispensing head load capacity q j Sure:

[0109]

[0110] Among them, t represents the current iteration number, u i Point P for the task i The priority of d ij (t) is the current task point P i To the position x of the dispensing head j at iteration t j (t),y j (t),z j (t), q j is the maximum load capacity of the dispensing head j, v j (t) is the speed of the dispensing head j at iteration t, which indicates the influence of the dispensing head movement on the attraction, γ is the load weight adjustment factor, which controls the influence of the dispensing head load on the attraction, β is the speed weight adjustment factor, which controls the influence of the dispensing head movement speed on the attraction, ∈ is a very small positive number, which is used to avoid the denominator being zero;

[0111] S22. The attraction strategy of the black kite optimization algorithm is used to simulate the black kite's preferential allocation behavior to high-attractive targets during prey tracking. i ,u i ) and dispensing head (v j ,q j )'s attractiveness value α i (t) Calculate and sort by attraction value α i (t) Prioritize the allocation of task points with an attraction higher than a preset value to the dispensing head with the largest attraction, and dynamically adjust the remaining load capacity of the dispensing head during the allocation process so that the allocation of each dispensing head does not exceed its maximum load capacity q j ;

[0112] S23. Based on the relationship between the assigned task points and the glue dispensing heads, an initial path planning individual is constructed. The path planning individual includes the task point sequence and glue dispensing path of each glue dispensing head. The attack strategy of the black kite optimization algorithm is applied to the initial path planning individual to locally optimize the sequence and allocation plan of the task points.

[0113] In this implementation manner, S23 specifically includes the following steps:

[0114] S231. Based on the distribution relationship between the task points and the dispensing heads, generate the initial task point sequence S for each dispensing head j , the task point sequence represents the set of task points that the dispensing head j needs to visit in sequence;

[0115] S232. Calculate the total length L of each dispensing head path in the initial path planning individual j , the total path length L j Indicates that the dispensing head j follows the task point sequence S j The total path length L when executing the task j ;

[0116] S233. Apply the attack strategy of the black kite optimization algorithm to locally optimize the initial path planning individual. The optimization process includes:

[0117] Path exchange optimization: task point sequence S at dispensing head j j In the example above, the order of the two task points is exchanged and the total length of the path after the exchange is calculated. ′ j , if L ′ j <L j , then update the task point sequence to the optimized order.

[0118] Task point redistribution: based on dispensing head load capacity q j And the task point attraction α i (t), for the task point sequence S j Reallocate the task points with low attraction;

[0119] S234. Iterate and perform local optimization, setting the number of optimization iterations T opt and convergence condition, when the total path length of all dispensing heads is When it stops decreasing or reaches the maximum number of iterations, the optimized path planning individual is output;

[0120] S235. Output the optimized path planning individual, which includes the task point sequence S of each dispensing head. j and dispensing path, the optimized path planning individual meets the requirements of minimizing the path length and balancing the load of the dispensing head.

[0121] In this implementation, S3 specifically includes the following steps:

[0122] S31. Select path planning individuals based on fitness values, select path planning individuals with fitness values ​​higher than a preset threshold from the path planning population to enter the next generation, and the fitness value F is calculated based on the path length L of the path planning individual j , path conflict cost C conflict Load balance of dispensing head imbalance Defined as:

[0123]

[0124] Among them, w 1 ,w2 ,w 3 is the weight coefficient, Indicates the total length of all dispensing head paths, C conflict represents the path conflict cost, U imbalance Indicates the dispensing head load imbalance;

[0125]

[0126] Among them, w j is the current workload of glue head j, is the average load of all pads.

[0127] S32. Randomly select the task point sequences of two path planning individuals for partial mapping and crossover, retain the relative order of the task point sequences to generate a new task point allocation scheme, map the crossover task point allocation scheme to the path planning population to generate a new path planning individual, and retain the elite individual with the highest fitness value from participating in the crossover operation to maintain the stability of the optimal solution of the population;

[0128] S33. Randomly select a sequence from the task point sequence of the new path planning individual and arrange it in reverse order, and calculate the path length L after mutation ′ j , if L ′ j <L j , then update the path planning individual, randomly exchange the positions of two task points in the task point sequence of the new path planning individual, recalculate the fitness value F', if F'>F, retain the mutation result;

[0129] S34. Re-evaluate the fitness value of the mutated path planning population and remove new path planning individuals whose fitness value is lower than the threshold.

[0130] In this implementation, S4 specifically includes the following steps:

[0131] S41. Based on the task point set P task And the dispensing head parameter set H for the new path planning individual task point P k ∈P task Calculate the multi-target attraction value α k (t), the attractiveness value combines priority, distance, load balance and path conflict risk and is calculated as:

[0132]

[0133] Among them, u k Point P for the task k The priority of d kj (t) is the task point P kThe Euclidean distance to the dispensing head j at iteration t, q j is the maximum load capacity of the dispensing head j, R k,j (t) is the path conflict risk factor, indicating that the task point P k The probability of a possible conflict with the current path of dispensing head j is calculated by the frequency of historical path intersections, γ 1 is the load adjustment factor, which controls the impact of load capacity, δ is the path conflict risk adjustment factor, which reduces the impact of path conflict on attractiveness;

[0134] S42. Simulate the collaborative prey capture behavior of black-winged kites, intelligently focus on the task points with attraction higher than the threshold, and dynamically adjust the weight of the dispensing head to the task points with high attraction:

[0135]

[0136] Among them, ω k,j (t) is the distance between dispensing head j and task point P k The allocation weight, η is the dynamic adjustment factor, The average attraction value of the task point set is used to increase the priority of the dispensing head for high attraction task points;

[0137] S43. Update the current position X of the path planning individual according to the intelligent focusing strategy combined with the attack and gliding behavior of the black kite optimization algorithm i (t+1) and speed V i (t+1):

[0138] V i (t+1)=λ·V i (t)+c 1 ·r 1 ·(X best -X i (t))+c 2 ·r 2 ·(X gbest -X i (t))+φ·ΔX cluster ;

[0139] X i (t+1)=X i (t)+V i (t+1);

[0140] Where λ is the momentum coefficient, c 1 ,c 2 is the learning factor, r 1 ,r 2 is a random number, X best is the optimal position of the individual in the current path planning, X gbestis the global optimal path planning solution, ΔX cluster is the aggregation center offset of the high-attractive task points within the population, which is used to simulate the flocking behavior of black-winged kites; φ is the focusing behavior adjustment factor, which controls the degree of concentration of individuals in the population on the high-attractive task points;

[0141] S44. According to the change of the attractiveness of the task points in the path planning scheme, the following two aspects are collaboratively optimized:

[0142] Load balancing optimization is to dynamically adjust the task point allocation to make the workload of the dispensing head close to balanced;

[0143] Path reordering optimization is to dynamically reorder non-critical task points in the task point sequence to reduce path length and path conflicts.

[0144] S45. Calculate the global attraction variance and the total length of path planning after each iteration. If the attraction variance is less than the set threshold or reaches the maximum number of iterations, output the optimized path planning solution; otherwise, return to execute steps S42 to S44.

[0145] In this implementation, S5 specifically includes the following steps:

[0146] S51. Based on the individuals in the optimized path planning scheme, time and space conflict detection is performed on the task points in the dispensing head path:

[0147]

[0148] Among them, C conflict (j,k) indicates the dispensing heads j and k at the task point P i The path conflict cost, T j (i) and T k (i) Dispensing heads j and k at task point P i The occupied time interval, is the indicator function, if two time intervals overlap, the conflict cost increases;

[0149] A path conflict matrix C is generated by detecting the conflicts of task points in all dispensing head paths, and the matrix elements of the path conflict matrix C represent the conflict intensity between the dispensing heads;

[0150] S52. Identify the conflict points corresponding to the non-zero elements in the path conflict matrix C, and select the task point P with the largest conflict intensity. c , based on the task point priority u c Adjust the priority to ensure the path execution order of high priority task points, and record the relevant glue head information of the conflict point. c ,k c ;

[0151] S53. Conflict point P c The task allocation is re-optimized, and the execution dispensing head j' of the conflict point is reselected based on the load capacity and attraction value of the dispensing head:

[0152]

[0153] Among them, α c (j) is the distance between dispensing head j and task point P c The attractiveness, q j is the maximum load capacity of the dispensing head j, w j is the current load of dispensing head j, indicating that the dispensing head with the largest remaining load capacity after allocation will be executed first;

[0154] S54. Update the path sequence of the dispensing head related to the conflict point based on the reallocated task point, adjust the path execution order, so that the execution time of the conflict point does not overlap with other task points, and the total path length meets the optimization goal:

[0155]

[0156] in, Click i for the task x and i x+1 The distance between Click i for the task k-1 The distance from the conflict point c, For conflict point c and task point i k+1 distance.

[0157] S55. Repeat the path conflict detection, conflict point identification and path reallocation until all elements of the path conflict matrix C are zero or the maximum number of iterations is reached, and output a multi-point glue head collaborative working path planning solution with no path conflicts.

[0158] A glue dispensing device comprises an internally arranged processor and memory, wherein computer instructions are stored in the memory, and when the processor runs the computer instructions, a collaborative working path planning method for multiple glue dispensing heads based on a genetic algorithm is executed.

[0159] Embodiment 1:

[0160] Example In the daily operation of the dispensing production line, a certain task requires dispensing of 200 electronic circuit boards. There are 150 task points distributed in different areas. The spatial distribution of the task points is complex, including high-density areas and low-density areas, with priorities ranging from high to low. There are 6 dispensing heads working together on the production line. The maximum load capacity of each dispensing head is 1000 units of glue, and the movement speed is 500 mm / s. The task requires that all dispensing operations be completed within 90 minutes and the dispensing path must be as short as possible to avoid path conflicts between dispensing heads.

[0161] In the traditional path planning method, the system assigns task points to the rubber head through a preset path. In the embodiment, the task point P 1 , P 2 and P 3 Assigned to glue head 1, and task point P 4 , P 5 and P 6 Assigned to glue head 2, however, in the actual operation process, it is found that glue head 1 and glue head 2 are at task point P 4 The nearby paths intersected, causing the two devices to wait for each other, delaying the task completion time. In addition, the task points were unevenly distributed, with the load of glue head 1 reaching 95%, while the load of glue head 3 was only 60%, resulting in low resource utilization.

[0162] In order to solve these problems, the method of the present invention is applied to this task. First, the system inputs the spatial coordinates and priority information of the task points to generate a task point set P. task = {P 1 ,P 2 ,…,P 150} and the dispensing head parameter set H = {(v 1 ,q 1 ),(v 2 ,q 2 ),…,(v 6 ,q 6 )}, when generating the initial population, the system combines the multi-objective attraction model to calculate the attraction value of the task point. In the embodiment, the task point P 10 The priority is "high", the distance from the rubber head 1 is 100 mm, and the attraction is calculated as follows:

[0163]

[0164] By sorting by attraction, the system preferentially assigns task points with higher attraction to suitable rubber heads.

[0165] During the task execution, the system monitors the path conflicts and task point priority changes in real time. In the embodiment, at the 30th minute, the system detects that the task point P 20 and P 21 The paths intersect, and glue head 2 and glue head 4 conflict because they arrive at the path intersection at the same time. The conflict point is identified and redistributed through the path, and the task point P is 20 The execution rubber head of the rubber head is adjusted from rubber head 4 to rubber head 3, and the task point P is adjusted at the same time. 20 The path order is optimized to avoid path crossing and reduce the conflict cost to 0.

[0166] The dynamic adjustment capability is particularly prominent in this task. In the embodiment, at the 45th minute, the system receives the newly added task point P 151The point is the highest priority emergency task point with coordinates (50, 60, 20). The traditional method cannot dynamically adjust the path and needs to wait for the current task to be completed before replanning. The method of the present invention completes the path adjustment within 3 seconds through real-time attraction calculation and path optimization. 151 Assign it to glue head 1 and reorder the path of glue head 1 to give priority to P 151 .

[0167] The following is the performance comparison data of the method of the present invention, the traditional static path planning method, and the conventional genetic algorithm method:

[0168] index Method of the present invention Traditional static methods Conventional genetic algorithm Completion time (minutes) 85 110 95 Total path length (mm) 11000 14500 12300 Path conflict count 0 12 5 Rubber head load balance coefficient 0.91 0.68 0.81 Dynamic task response time (seconds) 3 300 45

[0169] By comparing the data, it can be seen that the method of the present invention is superior to the traditional method in terms of path conflict avoidance, path length optimization and dynamic adaptability of tasks. In terms of dynamic task response, the traditional static method requires 300 seconds to replan the path, while the method of the present invention only takes 3 seconds to complete the path adjustment, and the dynamic adaptability is significantly improved. In addition, through load balancing optimization, the load balancing coefficient of the rubber head of the present invention is increased to 0.91, the resource utilization rate is significantly improved, and the path conflict is reduced from 12 times in the traditional method to 0 times, ensuring the smoothness of the collaborative work of multiple rubber heads.

[0170] This embodiment shows that the method of the present invention can effectively solve the problems of path conflict, uneven task distribution and insufficient dynamic adaptability in the collaborative work of multiple dispensing heads, and provides an efficient and reliable solution for dispensing operations in complex industrial environments.

[0171] The present invention combines the dynamic hunting behavior of the black kite optimization algorithm in the genetic algorithm to perform multi-objective optimization on the attractiveness of the path planning individuals, and comprehensively considers the priority of the task point, the probability of path intersection conflict between rubber heads, and the load capacity of the rubber heads to propose a multi-objective attractiveness function and an intelligent focusing strategy. The path conflict points can be detected and dynamically adjusted in real time, and the path conflict problem can be effectively avoided by path reallocation and task point sequence optimization, thereby ensuring the safety of collaborative operations between rubber heads and the global optimization of path planning.

[0172] The present invention is based on the collaborative work of a genetic algorithm and a black-winged kite optimization algorithm, and realizes load balancing of task allocation of multiple glue heads by dynamically adjusting the task allocation weights of the glue dispensing heads. In the task allocation process, an intelligent focusing strategy and a multi-dimensional collaborative optimization method are adopted to give priority to allocating task points with higher attractiveness, and dynamically adjust the remaining load capacity of the glue heads, so that the workload of each glue head tends to be balanced, avoiding overload operation or idleness.

[0173] The present invention introduces a dynamic parameter adjustment mechanism and a path conflict matrix model, so that the path planning scheme can quickly respond to changes in the distribution of task points or abnormal conditions of the rubber head state. The fluttering and gliding behaviors of the black kite optimization algorithm are combined with the dynamic attraction model, so that path planning can achieve real-time adjustment and adaptive optimization in complex industrial environments. When responding to sudden tasks, path changes or equipment failures, new optimized paths can be generated in a shorter time, thereby ensuring the flexible production capacity of multiple rubber heads working together.

[0174] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for collaborative working path planning of multiple glue heads based on genetic algorithm, characterized in that: The steps include: S1. Obtain the spatial coordinates of the task points to be processed of the multi-dispensing glue head dispensing task and the parameter information of the dispensing task, and generate a task point set and a dispensing head parameter set; S2. Initialize the initial population of the genetic algorithm, the initial population is composed of a plurality of path planning individuals, each path planning individual includes a task point allocation scheme and a dispensing path planning scheme, and in the initialization process, the task point set and the dispensing head parameter set are used to preferentially allocate task points with higher attractiveness to the appropriate dispensing head by combining the attractiveness calculation method of the black-winged kite optimization algorithm; S3. Perform genetic algorithm operations on the initial population in S2 to generate new path planning individuals; S4. Based on the black kite optimization algorithm, the new path planning individual is locally optimized, the black kite optimization algorithm is used to simulate the prey tracking behavior of the black kite, the task point allocation is optimized by dynamically adjusting the attractiveness of the task points in the task point set, and the path planning scheme is further optimized; S5. Perform path conflict detection on the optimized path planning individuals, identify path intersection conflict points based on the task point set and the dispensing head parameter set, and adjust the conflicting individuals by path redistribution to generate a path planning solution without path conflict; S6. Determine whether the path planning population without conflict meets the convergence condition, wherein the convergence condition includes that the maximum number of iterations reaches a preset threshold or the fitness value of the population reaches a target threshold. When the path planning population meets the convergence condition, the optimal path planning scheme is output. If the path planning population does not meet the convergence condition, the execution of steps S3 to S5 is returned. S7. Input the final optimized path planning scheme into the control module of the dispensing equipment, and use the dispensing task parameter information to drive multiple dispensing heads to perform the dispensing task according to the optimized path to complete the collaborative work of the dispensing task.

2. The method for collaborative working path planning of multiple glue heads based on genetic algorithm according to claim 1 is characterized in that: The S1 specifically includes the following steps: S11. Obtain the spatial coordinates of the task point to be processed, where the spatial coordinates are the absolute positions of the task point in three-dimensional space: P={P1,P2,…,P n }; Among them, P i =(x i ,y i ,z i ), x i ,y i ,z i Represent the task point P i x, y, z coordinates in three-dimensional space; S12. Obtain parameter information of the dispensing head, the parameter information including the number of dispensing heads m, the dispensing head movement speed set V, the dispensing head maximum load capacity set Q and the dispensing task execution priority set U; Dispensing head movement speed set V in v j Indicates the movement speed of the dispensing head j; Maximum load capacity of dispensing head set Q in q j Indicates the maximum load capacity of the dispensing head j; The execution priority set of the dispensing task is u in i Represents the task point P i The higher the priority of the task point P i Will be given priority in path planning; S13. Generate a task point set based on the spatial coordinate set P of the task point and the parameter information of the dispensing task. Each element of the task point set represents a combination of the coordinates of the task point and its dispensing task execution priority: P task ={(P i ,u i )∣i=1,2,…,n}; S14. Construct a dispensing head parameter set, where each element of the dispensing head parameter set represents a combination of the movement speed and the maximum load capacity of the dispensing head: H={(v j ,q j )∣j=1,2,…,m}。 3. The method for collaborative working path planning of multiple glue heads based on genetic algorithm according to claim 1 is characterized in that: The S2 comprises the following steps: S21. Calculate the attractiveness of the task points. Based on the attractiveness model of the black kite optimization algorithm, the task point set P task The task point P in i Calculate attraction α i , the attraction α i By the task point priority u i , the distance between the task point and the dispensing head d ij and dispensing head load capacity q j Sure: Among them, t represents the current iteration number, u i Point P for the task i The priority of d ij (t) is the current task point P i To the position x of the dispensing head j at iteration t j (t),y j (t),z j (t), q j is the maximum load capacity of the dispensing head j, v j (t) is the speed of the dispensing head j at iteration t, which indicates the influence of the dispensing head movement on the attraction, γ is the load weight adjustment factor, which controls the influence of the dispensing head load on the attraction, β is the speed weight adjustment factor, which controls the influence of the dispensing head movement speed on the attraction, ∈ is a very small positive number, which is used to avoid the denominator being zero; S22. The attraction strategy of the black kite optimization algorithm is used to simulate the black kite's preferential allocation behavior to high-attractive targets during prey tracking. i ,u i ) and dispensing head (v j ,q j )'s attractiveness value α i (t) Calculate and sort by attraction value α i (t) Prioritize the allocation of task points with an attraction higher than a preset value to the dispensing head with the largest attraction, and dynamically adjust the remaining load capacity of the dispensing head during the allocation process so that the allocation of each dispensing head does not exceed its maximum load capacity q j ; S23. Based on the relationship between the assigned task points and the glue dispensing heads, an initial path planning individual is constructed. The path planning individual includes the task point sequence and glue dispensing path of each glue dispensing head. The attack strategy of the black kite optimization algorithm is applied to the initial path planning individual to locally optimize the sequence and allocation plan of the task points.

4. The method for collaborative working path planning of multiple glue heads based on genetic algorithm according to claim 1 is characterized in that: The S23 specifically includes the following steps: S231. Based on the distribution relationship between the task points and the dispensing heads, generate the initial task point sequence S for each dispensing head j , the task point sequence represents the set of task points that the dispensing head j needs to visit in sequence; S232. Calculate the total length L of each dispensing head path in the initial path planning individual j , the total path length L j Indicates that the dispensing head j follows the task point sequence S j The total path length L when executing the task j ; S233. Apply the attack strategy of the black kite optimization algorithm to locally optimize the initial path planning individual. The optimization process includes: Path exchange optimization: task point sequence S at dispensing head j j In the example above, the order of the two task points is exchanged and the total length of the path after the exchange is calculated. ′ j , if L ′ j <L j , then update the task point sequence to the optimized order. Task point redistribution: based on dispensing head load capacity q j And the task point attraction α i (t), for the task point sequence S j Reallocate the task points with low attraction; S234. Iterate and perform local optimization, setting the number of optimization iterations T opt and convergence condition, when the total path length of all dispensing heads is When it stops decreasing or reaches the maximum number of iterations, the optimized path planning individual is output; S235. Outputting the optimized path planning individual, the path planning individual includes the task point sequence S of each dispensing head j and dispensing path, the optimized path planning individual meets the requirements of minimizing the path length and balancing the load of the dispensing head.

5. The method for collaborative working path planning of multiple glue heads based on genetic algorithm according to claim 1 is characterized in that: The S3 specifically includes the following steps: S31. Select path planning individuals based on fitness values, select path planning individuals with fitness values ​​higher than a preset threshold from the path planning population to enter the next generation, and the fitness value F is calculated based on the path length L of the path planning individual j , path conflict cost C conflict Load balance of dispensing head imbalance Defined as: Among them, w1, w2, w3 are weight coefficients, Indicates the total length of all dispensing head paths, C conflict represents the path conflict cost, U imbalance Indicates the dispensing head load imbalance; S32. Randomly select the task point sequences of two path planning individuals for partial mapping and crossover, retain the relative order of the task point sequences to generate a new task point allocation scheme, map the crossover task point allocation scheme to the path planning population to generate a new path planning individual, and retain the elite individual with the highest fitness value from participating in the crossover operation to maintain the stability of the optimal solution of the population; S33. Randomly select a sequence from the task point sequence of the new path planning individual and arrange it in reverse order, and calculate the path length L after mutation ′ j , if L ′ j <L j , then update the path planning individual, randomly exchange the positions of two task points in the task point sequence of the new path planning individual, recalculate the fitness value F', if F'>F, retain the mutation result; S34. Re-evaluate the fitness value of the mutated path planning population and remove new path planning individuals whose fitness value is lower than the threshold.

6. The method for collaborative working path planning of multiple glue heads based on genetic algorithm according to claim 1 is characterized in that: The S4 specifically comprises the following steps: S41. Based on the task point set P task And the dispensing head parameter set H for the new path planning individual task point P k ∈P task Calculate the multi-target attraction value α k (t), the attractiveness value combines priority, distance, load balance and path conflict risk and is calculated as: Among them, u k Point P for the task k The priority of kj (t) is the task point P k The Euclidean distance to the dispensing head j at iteration t, q j is the maximum load capacity of the dispensing head j, R k,j (t) is the path conflict risk factor, indicating that the task point P k The probability of a possible conflict with the current path of the dispensing head j is calculated by the frequency of historical path intersections. γ1 is the load adjustment factor, which controls the impact of load capacity, and δ is the path conflict risk adjustment factor, which reduces the impact of path conflict on attractiveness; S42. Simulate the collaborative prey capture behavior of black-winged kites, intelligently focus on the task points with attraction higher than the threshold, and dynamically adjust the weight of the dispensing head to the task points with high attraction: Among them, ω k,j (t) is the distance between dispensing head j and task point P k The allocation weight, η is the dynamic adjustment factor, The average attraction value of the task point set is used to increase the priority of the dispensing head for high attraction task points; S43. Update the current position X of the path planning individual according to the intelligent focusing strategy combined with the attack and gliding behavior of the black kite optimization algorithm i (t+1) and speed V i (t+1): V i (t+1)=λ·V i (t)+c1·r1·(X best -X i (t))+c2·r2·(X gbest -X i (t))+φ· ΔX cluster ; X i (t+1)=X i (t)+V i (t+1); Among them, λ is the momentum coefficient, c1, c2 are learning factors, r1, r2 are random numbers, and X best is the optimal position of the individual in the current path planning, X gbest is the global optimal path planning solution, ΔX cluster is the aggregation center offset of the high-attractive task points within the population, which is used to simulate the flocking behavior of black-winged kites; φ is the focusing behavior adjustment factor, which controls the degree of concentration of individuals in the population on the high-attractive task points; S44. According to the change of the attractiveness of the task points in the path planning scheme, the following two aspects are collaboratively optimized: Load balancing optimization is to dynamically adjust the task point allocation to make the workload of the dispensing head close to balanced; Path reordering optimization is to dynamically reorder non-critical task points in the task point sequence to reduce path length and path conflicts. S45. Calculate the global attraction variance and the total length of path planning after each iteration. If the attraction variance is less than the set threshold or reaches the maximum number of iterations, output the optimized path planning solution; otherwise, return to execute steps S42 to S44.

7. The method for collaborative working path planning of multiple glue heads based on genetic algorithm according to claim 1 is characterized in that: The S5 specifically includes the following steps: S51. Based on the individuals in the optimized path planning scheme, time and space conflict detection is performed on the task points in the dispensing head path: Among them, C conflict (j,k) indicates the dispensing heads j and k at the task point P i The path conflict cost, T j (i) and T k (i) Dispensing heads j and k at task point P i The occupied time interval, is the indicator function, if two time intervals overlap, the conflict cost increases; A path conflict matrix C is generated by detecting the conflicts of task points in all dispensing head paths, and the matrix elements of the path conflict matrix C represent the conflict intensity between the dispensing heads; S52. Identify the conflict points corresponding to the non-zero elements in the path conflict matrix C, and select the task point P with the largest conflict intensity. c , based on the task point priority u c Adjust the priority to ensure the path execution order of high priority task points, and record the relevant glue head information of the conflict point. c ,k c ; S53. Conflict point P c The task allocation is re-optimized, and the execution dispensing head j' of the conflict point is reselected based on the load capacity and attraction value of the dispensing head: Among them, α c (j) is the distance between dispensing head j and task point P c The attractiveness, q j is the maximum load capacity of the dispensing head j, w j is the current load of dispensing head j, indicating that the dispensing head with the largest remaining load capacity after allocation will be executed first; S54. Update the path sequence of the dispensing head related to the conflict point based on the reallocated task point, adjust the path execution order, so that the execution time of the conflict point does not overlap with other task points, and the total path length meets the optimization goal: in, Click i for the task x and i x+1 The distance between Click i for the task k-1 The distance from the conflict point c, For conflict point c and task point i k+1 distance. S55. Repeat the path conflict detection, conflict point identification and path reallocation until all elements of the path conflict matrix C are zero or the maximum number of iterations is reached, and output a multi-point glue head collaborative working path planning solution with no path conflicts.

8. A dispensing device, comprising a processor and a memory disposed therein, characterized in that: The memory stores computer instructions, and the processor executes the method for collaborative working path planning of multiple glue heads based on genetic algorithm as described in any one of claims 1 to 7 when running the computer instructions.

Images