A method for human-machine collaborative assembly task allocation considering task matching and human factors

Abstract

The present invention discloses a method for allocating tasks for human-machine collaborative assembly that considers task matching and human factors, including S1. collecting the time required for workers and robots to complete each task and the fatigue coefficient information of workers completing each assembly task; S2. establishing an evaluation system for task matching, and using a fuzzy comprehensive evaluation method to calculate the task matching of workers and robots for each task; S3. constructing a task allocation model that considers both task matching and human factors based on the collected information and the calculated task matching; S4. using a non-dominated sorting genetic algorithm with an elite strategy to solve the constructed task allocation model, and outputting an optimal task allocation plan. The technical solution of the present invention can greatly improve the matching degree between workers and robots and task design requirements in task allocation while ensuring assembly efficiency, thereby greatly reducing the problem of high assembly error rate caused by low task matching.

Description

Technical Field

[0001] The present invention relates to the technical field of human-machine collaborative assembly, and in particular to a human-machine collaborative assembly task allocation method that considers task matching and human factors. Background Art

[0002] In the context of intelligent manufacturing, the flexibility of the production process is increasingly required. Manufacturing operations require automation systems to be highly flexible and adaptable to cope with rapid changes in customer needs, products, and processes. Traditional manual assembly is very costly; and traditional automation systems using industrial robots are inflexible to adapt to dynamic manufacturing environments. In response to this situation, collaborative assembly between humans and robots has become a hot topic in the manufacturing industry. In recent years, especially as product life cycles continue to shorten and manufacturing forms gradually shift from mass production to multi-variety small-batch production, the task allocation of human-robot collaborative assembly has become an important and frequent activity. In human-robot collaborative assembly, unreasonable task allocation is an important reason for low assembly efficiency and great risks to workers' physical and mental health. Therefore, how to reasonably allocate tasks according to the characteristics of workers and robots in human-robot collaborative assembly has become an important decision-making issue.

[0003] Existing methods for human-machine collaborative assembly task allocation focus on pursuing assembly efficiency, ignoring the problem of matching the skills of workers and robots with the design requirements of tasks, which leads to workers and robots often completing tasks they are not good at, resulting in a high assembly error rate; in addition, existing methods do not pay enough attention to the physical and mental health of workers in human-machine collaborative assembly, and workers have great health risks. Therefore, in the allocation of human-machine collaborative assembly tasks, it is of great practical significance to consider the matching of the skills of workers and robots with the design requirements of tasks, and to pay attention to the physical and mental health of workers in human-machine collaborative assembly.

[0004] The existing methods of human-machine collaborative assembly task allocation have the following main shortcomings: First, in task allocation, the main goal is to reduce the unit product assembly time and improve assembly efficiency, and not enough attention is paid to whether the skills of workers and robots match the task design requirements. As a result, in actual production, workers and robots often have a high assembly error rate due to low task matching. Second, as workers are an important part of human-machine collaborative assembly, the existing task allocation methods do not fully consider the physiological and psychological characteristics of workers. In the existing task allocation methods, workers often suffer from excessive fatigue due to completing too many tasks in a row, which poses a great safety hazard to the physical and mental health of workers. Summary of the invention

[0005] In view of the defects in the prior art, the present invention provides a method for allocating human-machine collaborative assembly tasks taking into account task matching and human factors.

[0006] In order to achieve the above purpose, the present invention adopts the following technical solutions:

[0007] A method for allocating tasks of human-machine collaborative assembly taking into account task matching and human factors, characterized by comprising:

[0008] S1. Collect the time required for workers and robots to complete each task and the fatigue coefficient information of workers completing each assembly task;

[0009] S2. Establish a task matching evaluation system and use fuzzy comprehensive evaluation method to calculate the task matching between workers and robots for each task;

[0010] S3. Based on the collected information and the calculated task matching degree, a task allocation model that considers both task matching degree and human factors is constructed;

[0011] S4. Use the non-dominated sorting genetic algorithm with elite strategy to solve the constructed task allocation model and output the optimal task allocation plan.

[0012] Furthermore, the step S2 specifically includes:

[0013] S21. Select relevant indicators of the task matching evaluation system; the indicators include primary indicators and secondary indicators;

[0014] S22. Use network analysis method to calculate the weight of each indicator in the primary and secondary indicators;

[0015] S23. Based on the calculated weights, a fuzzy comprehensive evaluation method is used to calculate the task matching degree of each task for workers and robots.

[0016] Furthermore, the step S22 specifically includes:

[0017] S221. Use the 9-level scaling method to compare the importance of indicators in pairs, and perform consistency test to determine whether the consistency mark CI is less than 0.1. If so, execute step S222; if not, continue to execute step S221;

[0018] S222. Construct a judgment matrix corresponding to the indicator, and operate and normalize all judgment matrices to obtain a supermatrix, and weight the elements in the supermatrix to obtain a weighted supermatrix;

[0019] S223. Perform limit operation on the weighted supermatrix to obtain a limit supermatrix, and obtain the limit relative ranking vector of each indicator element based on the limit supermatrix.

[0020] Furthermore, the step S23 specifically includes:

[0021] S231. Use the fuzzy comprehensive evaluation method to grade the task matching degree between the worker and the robot for each design requirement of each task; wherein the grading levels include high matching degree, relatively high matching degree, medium matching degree, relatively low matching degree, and low matching degree, wherein the corresponding interval numbers of high matching degree, relatively high matching degree, medium matching degree, relatively low matching degree, and low matching degree are {[0.8,1), [0.6,0.8), [0.4,0.6), [0.2,0.4), [0,0.2)};

[0022] S232. Score the matching degree between workers and robots for each indicator;

[0023] S233. Based on the classification results and scoring results, the matching degree between the worker and the robot in each process is calculated, which is expressed as:

[0024] c jk =Z mn ·w mn [1, 0.8, 0.6, 0.4, 0.2] / (1 + 0.8 + 0.6 + 0.4 + 0.2)

[0025]

[0026]

[0027] Among them, c jk represents the task matching degree of the jth task completed by operator k; r mn0 It represents the membership degree of the oth level of the nth secondary indicator under the mth primary indicator; Z mn represents the membership matrix; v represents the total number of experts; u represents the number of experts given level o.

[0028] Further, step S3 specifically includes:

[0029] S31. Establish the assumptions and related parameters of the model;

[0030] S32. Construct the objective function of the task allocation model considering task matching and human factors;

[0031] S33. Determine constraint conditions for the constructed objective function based on the parameters.

[0032] Further, the objective function constructed in step S32 is expressed as:

[0033] f 1 =MinT

[0034]

[0035] Among them, f 1represents the minimization of assembly time; f 2 represents the maximum total task matching degree; T represents the time point within a given range; c jk represents the task matching degree of operator k completing task j; x jkt represents the decision variable.

[0036] Further, step S33 specifically includes:

[0037] S331. The minimum assembly time per unit product is the latest time when all operators complete the assembly task, expressed as:

[0038]

[0039] Where T represents the time point within a given range; K represents the number of executable operators; t jk represents the time required for operator k to complete task j; x jkt represents the decision variable; j represents the different process numbers in the task; k represents the operator; t represents time;

[0040] S332. At any time, each process j is completed by only one operator, expressed as:

[0041]

[0042] S333. At any time, each process j is completed by only one operator, which means that each operator can only handle one task at a time and cannot handle multiple tasks at the same time, expressed as:

[0043]

[0044] Among them, J represents the number of processes;

[0045] S334. Any subsequent task j in sequential assembly is not started until its previous task i has been completed, expressed as:

[0046]

[0047] where p j Indicates the index number of the task immediately preceding task j;

[0048] S335. Only when all parallel processes of parallel assembly are completed can the next sequential process be started, which is expressed as:

[0049]

[0050]

[0051] Among them, t ik represents the time required for operator k to complete task i; xikt ,y ji represents a decision variable; M represents an infinite number;

[0052] S336. The fatigue level of workers at any time should be within the safety threshold, expressed as:

[0053]

[0054]

[0055]

[0056] F oi ≤F max

[0057] Among them, τ i represents the rest time of workers before completing task i; F(τ i ) indicates that the worker passes through τ i Fatigue before rest; R(τ i ) indicates that the worker has passed τ i Fatigue after rest; F oi represents the worker's fatigue level after completing the i-th process of the o-th product; μ represents the worker's fatigue recovery index; F o(i-1) represents the worker's fatigue after completing the i-1th process of the oth product; i represents the fatigue coefficient of the i-th task; t i F represents the time required for the operator to complete task i; max Indicates the safety threshold of worker fatigue.

[0058] Furthermore, the step S4 specifically includes:

[0059] S41. Generate an initial chromosome population;

[0060] S42. Calculate the chromosome fitness to obtain a population group; wherein the fitness includes unit product assembly time and total task matching degree;

[0061] S43. Perform non-dominated sorting on all populations, select half of the individuals in the population to enter the next generation according to the Parato level, and determine whether the Parato levels are the same. If so, select the individual with greater crowding degree;

[0062] S44. Two-point crossover method was used to perform chromosome crossover processing;

[0063] S45. Use single-point mutation to mutate chromosomes;

[0064] S46. Use a repair strategy to check whether the chromosomes after crossover mutation meet the fatigue constraint. If not, randomly assign a task completed by a worker to the robot until the fatigue constraint is met, thereby generating a new population.

[0065] S47. The generated new population and the population in step S42 are combined into a new population, and the non-dominated sorting is performed again. According to the Parato level, 2 / 3 of the individuals in the population are selected to enter the next generation. It is determined whether the parato levels are the same. If so, the individual with a greater crowding degree is selected, and the number of iterations is increased by 1;

[0066] S48. Determine whether the number of iterations reaches the set value. If so, output the final population; if not, return to step S43;

[0067] S49. Normalize the indicator dimensions of the decision-making process to obtain the optimal allocation plan.

[0068] Furthermore, the total task matching degree in step S42 is expressed as:

[0069]

[0070] Where C represents the total task matching degree; N represents the number of assembly tasks; c jk It represents the task matching degree of operator k completing task j.

[0071] Furthermore, in step S49, normalization processing is performed to obtain the optimal allocation solution, which is expressed as:

[0072]

[0073]

[0074] B i =y′ i1 × 1 +y′ i2 × 2

[0075] in, represents the worst value of objective function 1 among all solutions, represents the worst value of objective function 2 among all solutions, represents the optimal value of objective function 1 among all solutions, represents the optimal value of objective function 2 among all solutions; y′ i1 , y′ i2 are the dimensionless values ​​after being processed by objective function 1 and objective function 2; B i represents the final comparison value of the i-th solution, where all B i Medium B iThe i-th solution corresponding to the minimum value is the optimal allocation solution.

[0076] Compared with the prior art, the present invention takes into account the matching of workers' and robots' own skills with the task design requirements while considering assembly efficiency, and also considers the physical and mental health of workers in task allocation. The technical solution of the present invention can greatly improve the matching degree between workers and robots and task design requirements in task allocation while ensuring assembly efficiency, thereby greatly reducing the problem of high assembly error rate caused by low task matching; at the same time, the fatigue resistance of robots is utilized in the task allocation process, and reasonable shift scheduling is performed, which greatly reduces the fatigue of workers in the assembly process, effectively ensures the physical and mental health of workers, and improves the happiness of workers. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Figure 1 This is a flow chart of a method for allocating tasks of human-machine collaborative assembly taking into account task matching and human factors, provided in Embodiment 1;

[0078] Figure 2 is a task priority relationship diagram of an assembly sequence provided in Example 2;

[0079] Figure 3 is a schematic diagram of the task allocation scheme of the method provided in Example 2;

[0080] Figure 4 This is a schematic diagram of a task allocation scheme of a traditional method provided in Example 2;

[0081] Figure 5 is a graph showing the variation of worker fatigue levels in the present method and the traditional method provided in Example 2;

[0082] Figure 6 is a schematic diagram of the task allocation scheme of the method provided in Example 3;

[0083] Figure 7 This is a schematic diagram of a task allocation scheme of a traditional method provided in Example 3;

[0084] Figure 8 It is a curve diagram of the change of worker fatigue degree of the present method and the traditional method provided in Example 3. DETAILED DESCRIPTION

[0085] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0086] The purpose of the present invention is to provide a method for allocating human-machine collaborative assembly tasks taking into account task matching and human factors in order to address the defects of the prior art.

[0087] Embodiment 1

[0088] This embodiment provides a method for allocating tasks of human-machine collaborative assembly taking into account task matching and human factors. Figure 1 As shown, it is characterized by comprising:

[0089] S1. Collect the time required for workers and robots to complete each task and the fatigue coefficient information of workers completing each assembly task;

[0090] S2. Establish a task matching evaluation system and use fuzzy comprehensive evaluation method to calculate the task matching between workers and robots for each task;

[0091] S3. Based on the collected information and the calculated task matching degree, a task allocation model that considers both task matching degree and human factors is constructed;

[0092] S4. Use the non-dominated sorting genetic algorithm with elite strategy to solve the constructed task allocation model and output the optimal task allocation plan.

[0093] In step S2, a task matching evaluation system is established, and the task matching degree between the worker and the robot for each task is calculated using a fuzzy comprehensive evaluation method. Specifically, it includes:

[0094] S21. Select relevant indicators of the task matching evaluation system; the indicators include primary indicators and secondary indicators;

[0095] The first-level indicators are the complexity of the parts themselves, the complexity of the parts operation process, and the human risk of the task. The second-level indicators of the complexity of the parts themselves include the weight of the parts, the size of the parts, the shape of the parts, the sensitivity of the parts, and the stability of the parts. The complexity of the operation process includes the assembly direction of the parts, whether the parts need to be pressed, the resistance of the parts insertion, and the fastening method of the parts. The human risk of the task includes the MSD risk of the task, the physical load of the task, and the mental load of the task. The human risk of the task is affected by the complexity of the parts themselves and the complexity of the operation process.

[0096] S22. Use network analysis method to calculate the weight of each indicator in the primary and secondary indicators;

[0097] S221. Use the 9-level scaling method to compare the importance of indicators in pairs, and perform consistency test to determine whether the consistency mark CI is less than 0.1. If so, execute step S222; if not, continue to execute step S221;

[0098] Invite several technical and management personnel of the company to use the 9-level scaling method to conduct pairwise comparison of significance and conduct consistency test. If the CI is less than 0.1, proceed to the next step, otherwise it is necessary to consult the management and technical personnel again.

[0099] S222. Construct a judgment matrix corresponding to the indicator, and operate and normalize all judgment matrices to obtain a supermatrix, and weight the elements in the supermatrix to obtain a weighted supermatrix;

[0100] S223. Perform limit operation on the weighted supermatrix to obtain a limit supermatrix, and obtain the limit relative ranking vector of each indicator element based on the limit supermatrix.

[0101] S23. Based on the calculated weights, a fuzzy comprehensive evaluation method is used to calculate the task matching degree of each task for workers and robots.

[0102] S231. The fuzzy comprehensive evaluation method is used to classify the task matching degree between workers and robots for each design requirement of each task; there are 5 levels of classification, including high matching degree, relatively high matching degree, medium matching degree, relatively low matching degree, and low matching degree, among which the corresponding interval numbers of high matching degree, relatively high matching degree, medium matching degree, relatively low matching degree, and low matching degree are {[0.8,1), [0.6,0.8), [0.4,0.6), [0.2,0.4), [0,0.2)};

[0103] S232. Ask several relevant industry personnel to score the matching degree between workers and robots for each indicator;

[0104] S233. Based on the classification results and scoring results, the matching degree between the worker and the robot in each process is calculated, which is expressed as:

[0105] c jk =Z mn ·w mn [1, 0.8, 0.6, 0.4, 0.2] / (1 + 0.8 + 0.6 + 0.4 + 0.2)

[0106]

[0107]

[0108] Among them, c jk represents the task matching degree of the jth task completed by operator k; r mn0 It represents the membership degree of the oth level of the nth secondary indicator under the mth primary indicator; Z mn represents the membership matrix; v represents the total number of experts; u represents the number of experts given level o.

[0109] In step S3, based on the collected information and the calculated task matching degree, a task allocation model that considers both task matching degree and human factors is constructed. Specifically, it includes:

[0110] S31. Establish the assumptions and related parameters of the model;

[0111] Assumptions of the model: (1) Tasks are independent, and the assembly process and the assembly relationship between each process are known and strictly follow the predetermined assembly sequence. (2) Each assembly task can only be performed by one operator. (3) Workers and robots share the workplace and time. (4) The time for workers and robots to complete each process is fixed, regardless of the dynamic changes in completion time. (5) When producing the same product, the task allocation plan should be the same.

[0112] Symbols of known parameters in the model and explanation of decision variables: T represents the time point within a given range, J represents the number of processes, i and j represent different process numbers in the task, K represents the number of operators that can be executed, m represents the number of workers in the assembly unit, n represents the number of robots in the assembly unit, k i represents the i-th operator, P j represents the number of tasks immediately preceding j, p j The index number of the task j's immediate predecessor, U j represents the number of tasks following task j, u j represents the index number of the task j's immediate predecessor, c jk represents the task matching degree of operator k completing task j, t jkrepresents the time required for operator k to complete task j, F io represents the fatigue degree of the i-th process to complete the o-th product, F max represents the safety threshold of worker fatigue, x jkt ,y ji is the decision variable, u represents the fatigue recovery index, τ i represents the rest time of the worker before completing task i, and C represents the total task matching degree.

[0113] in:

[0114]

[0115]

[0116] S32. Construct the objective function of the task allocation model considering task matching and human factors;

[0117] Determine the goal to be pursued, that is, the objective function. There are two goals to be pursued. One is to minimize the assembly time f 1 ; The second is to maximize the total task matching degree f 2 , expressed as:

[0118] f 1 =MinT

[0119]

[0120] Among them, f 1 represents the minimization of assembly time; f 2 represents the maximum total task matching degree; T represents the time point within a given range; c jk represents the task matching degree of operator k completing task j; x jkt represents the decision variable.

[0121] S33. Determine constraint conditions for the constructed objective function based on the parameters.

[0122] Based on the parameters obtained and established above, determine the constraints of the model, including:

[0123] S331. The minimum assembly time per unit product is the latest time when all operators complete the assembly task, expressed as:

[0124]

[0125] Where T represents the time point within a given range; K represents the number of executable operators; t jk represents the time required for operator k to complete task j; x jkt represents the decision variable; j represents the different process numbers in the task; k represents the operator; t represents time;

[0126] S332. At any time, each process j is completed by only one operator, expressed as:

[0127]

[0128] S333. At any time, each process j is completed by only one operator, which means that each operator can only handle one task at a time and cannot handle multiple tasks at the same time, expressed as:

[0129]

[0130] Among them, J represents the number of processes;

[0131] S334. Any subsequent task j in sequential assembly is not started until its previous task i has been completed, expressed as:

[0132]

[0133] where p j Indicates the index number of the task immediately preceding task j;

[0134] S335. Only when all parallel processes of parallel assembly are completed can the next sequential process be started, which is expressed as:

[0135]

[0136]

[0137] Among them, t ik represents the time required for operator k to complete task i; x ikt ,y ji represents a decision variable; M represents an infinite number;

[0138] S336. The fatigue level of workers at any time should be within the safety threshold, expressed as:

[0139]

[0140]

[0141]

[0142] F oi ≤F max

[0143] Among them, τ i represents the rest time of workers before completing task i; F(τ i ) indicates that the worker passes through τ i Fatigue before rest; R(τi ) indicates that the worker has passed τ i Fatigue after rest; F oi represents the worker's fatigue level after completing the i-th process of the o-th product; μ represents the worker's fatigue recovery index; F o(i-1) represents the worker's fatigue after completing the i-1th process of the oth product; i represents the fatigue coefficient of the i-th task; t i F represents the time required for the operator to complete task i; max Indicates the safety threshold of worker fatigue.

[0144] In step S4, a non-dominated sorting genetic algorithm with an elite strategy is used to solve the constructed task allocation model and output the optimal task allocation solution.

[0145] S41. Generate an initial chromosome population;

[0146] The length of the chromosome is twice the number of tasks; the first half of the chromosome corresponds to the order in which the tasks are completed, and the second half of the chromosome corresponds to the operator of the first half of the tasks in sequence. The chromosome numbers of the first half are randomly generated according to the immediate relationship of the process. The chromosome numbers of the second half are randomly generated from 1 to (m+n). Check whether the fatigue level of the generated task allocation plan meets the safety threshold in the process of producing all the required products. If not, regenerate it randomly.

[0147] S42. Calculate the chromosome fitness to obtain a population group; wherein the fitness includes unit product assembly time and total task matching degree;

[0148] The assembly time of the unit product can be obtained through step S331.

[0149] The total task matching degree is expressed as:

[0150]

[0151] Where C represents the total task matching degree; N represents the number of assembly tasks; c jk It represents the task matching degree of operator k completing task j.

[0152] S43. Perform non-dominated sorting on all populations, select half of the individuals in the population to enter the next generation according to the Parato level, and determine whether the Parato levels are the same. If so, select the individual with greater crowding degree;

[0153] S44. Use the two-point crossover method to perform chromosome crossover processing; specifically:

[0154] S441. Take the front halves of the two chromosomes. The front halves of the two parent chromosomes are represented by P1 and P2, and the chromosomes generated after the crossover are represented by C1 and C2. For the process part, this paper adopts the two-point crossover method. In order to ensure that the solution generated after the crossover is still a solution that satisfies the process priority relationship, the crossover steps are as follows: take any two crossover points, and the crossover points divide the two chromosomes into three parts. Delete the head and tail of P1, delete the gene numbers in P2 that are the same as the head and tail of P1, and turn the remaining part into the middle part of the daughter chromosome 1, which together with the head and tail of the original parent chromosome constitute the daughter chromosome 1. Similarly, the daughter chromosome 2 is obtained;

[0155] S442. Take the second half of the two chromosomes and generate new chromosomes using the two-point crossover method;

[0156] S443. Resynthesize the front and back halves into new chromosomes.

[0157] S45. Use single-point mutation to mutate chromosomes; specifically:

[0158] S451. Take the first half of the chromosome, randomly select a mutation point, and regenerate the chromosome after the mutation point according to the priority relationship of the tasks

[0159] S452. Take the second half of the chromosome, randomly select a mutation point, and turn it into a number of other operators other than this number.

[0160] S46. It is impossible to ensure that the new chromosomes generated after crossover mutation meet the fatigue constraint, so a repair strategy is used to check whether the chromosomes after crossover mutation meet the fatigue constraint. If not, a task completed by a worker is randomly assigned to the robot until the fatigue constraint is met to generate a new population;

[0161] S47. The generated new population and the population in step S42 are combined into a new population, and the non-dominated sorting is performed again. According to the Parato level, 2 / 3 of the individuals in the population are selected to enter the next generation. It is determined whether the parato levels are the same. If so, the individual with a greater crowding degree is selected, and the number of iterations is increased by 1;

[0162] S48. Determine whether the number of iterations reaches the set value. If so, output the final population; if not, return to step S43 until the number of iterations reaches the set threshold;

[0163] S49. Since the indicator dimensions of the decision-making process are different, the indicator dimensions of the decision-making process are normalized to obtain the optimal allocation plan, which is expressed as:

[0164]

[0165]

[0166] B i =y′ i1 × 1 +y′ i2 × 2

[0167] in, represents the worst value of objective function 1 among all solutions, represents the worst value of objective function 2 among all solutions, represents the optimal value of objective function 1 among all solutions, represents the optimal value of objective function 2 among all solutions; y′ i1 , y′ i2 are the dimensionless values ​​after being processed by objective function 1 and objective function 2; B i represents the final comparison value of the i-th solution, where all B i Medium B i The i-th solution corresponding to the minimum value is the optimal allocation solution.

[0168] This embodiment takes into account the matching of workers' and robots' own skills with the task design requirements while considering assembly efficiency, and also considers the physical and mental health of workers in task allocation. The technical solution of this embodiment can greatly improve the matching degree between workers and robots and task design requirements in task allocation while ensuring assembly efficiency, thereby greatly reducing the problem of high assembly error rate caused by low task matching; at the same time, the fatigue resistance of robots is utilized in the task allocation process, and reasonable shift scheduling is performed, which greatly reduces the fatigue of workers in the assembly process, effectively ensures the physical and mental health of workers, and improves the happiness of workers.

[0169] Embodiment 2

[0170] The method for allocating human-machine collaborative assembly tasks that considers task matching and human factors provided in this embodiment is different from the first embodiment in that:

[0171] This embodiment takes the example of a customized automobile gearbox produced by Y Company. The assembly of the gearbox consists of 34 processes. Figure 2 shown.

[0172] First, the time required for workers and robots to complete each task and the fatigue coefficient information of workers completing each assembly task are collected, as shown in Table 1 below to obtain relevant process information;

[0173] Table 1 Process information table of the process

[0174]

[0175]

[0176] After establishing an evaluation system for task matching, the fuzzy comprehensive evaluation method is used to calculate the task matching between workers and robots for each task; the task matching between workers and robots for 34 tasks can be obtained, as shown in Table 2.

[0177] Table 2 Task matching between workers and robots for different tasks

[0178]

[0179]

[0180] Based on the collected information and the calculated task matching degree, a task allocation model is constructed which takes both task matching degree and human factors into consideration.

[0181] The above model is solved by using non-dominated sorting genetic algorithm with elite strategy, with a population size of 200, a crossover rate of 0.95, a mutation rate of 0.1, 400 iterations, and w 1 =0.5,w 2 =0.5.

[0182] In the ρ(1,1) collaborative production mode, the allocation scheme of this method is used, such as Figure 3 As shown in the figure, the unit product assembly time is T = 310, C = 23.9263, the fatigue peak value of the entire assembly cycle is 0.582, and the fatigue range after stabilization is 0.04832-0.3786. Figure 5 The curve marked in .

[0183] In the ρ(1,1) collaborative production mode, the proposed method is not used, and the traditional method is used. The task allocation scheme is as follows Figure 4 At this time, the unit product assembly time is T = 268, the task matching degree is C = 21.7863, the fatigue peak value of the entire assembly cycle is 1, and the fatigue change range after stabilization is 0.09555-1. The fatigue change is as follows Figure 5 The curve in .

[0184] After comparison, it can be concluded that although the solution proposed by this method increases the unit product assembly time by 15.6%, it increases the task matching by 9.8%, reduces the fatigue peak by 41.8%, and makes the overall fatigue range of workers at an extremely low level after stabilization. The improvement of task matching means that workers and robots can complete more tasks they are good at, reducing the assembly error rate; and the decrease in fatigue peak means that the occupational health of workers is greatly guaranteed. Therefore, we can see the superiority of the method proposed in this article.

[0185] Embodiment 3

[0186] The method for allocating human-machine collaborative assembly tasks that considers task matching and human factors provided in this embodiment is different from the second embodiment in that:

[0187] The example used in this implementation example is a collaborative production mode different from the example in the second embodiment. The second embodiment is a collaborative production mode of 1 worker and 1 robot, i.e., ρ(1,1), while the present embodiment is a collaborative production mode of 1 worker and 2 robots, ρ(1,2). Other relevant information is the same as that in the second embodiment.

[0188] In the ρ(1,2) collaborative production mode, using this method, the solution is as follows Figure 6 At this time, the unit product assembly time T = 264, the total task matching degree C = 23.4568, the peak value of the fatigue during the entire assembly cycle is 0.5915, and the fatigue range after stabilization is 0.06887-0.5915. The fatigue changes of workers are as follows Figure 8 Annotated curve.

[0189] In the ρ(1,2) collaborative production mode, without using the strategy proposed in this paper, the traditional method is used, and the solution is as follows Figure 7 At this time, T = 245, C = 21.9376, the peak value of fatigue during the entire assembly cycle is 0.9575, and the fatigue range after stabilization is 0.1663-0.9575. The fatigue changes of workers are as follows Figure 8 Annotated curve.

[0190] After comparison, it can be concluded that under the ρ(1,2) collaborative production mode, although the solution proposed by this method increases the unit product assembly time by 7.76%, it improves the task matching degree by 6.93%, reduces the fatigue peak by 38.22%, and makes the overall fatigue range of workers at an extremely low level after stabilization. The improvement of task matching degree means that workers and robots can complete more tasks they are good at, reducing the assembly error rate; and the decrease in fatigue peak means that the occupational health of workers is greatly guaranteed. Therefore, we can see that our method can achieve the same superiority in Example 2ρ(1,2) collaborative production mode.

[0191] The specific embodiments described herein are merely examples of the spirit of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in similar ways, but they will not deviate from the spirit of the present invention or exceed the scope defined by the appended claims.

Claims

1. A method for allocating human-machine collaborative assembly tasks considering task matching and human factors, characterized in that: include: S1. Collect the time required for workers and robots to complete each task and the fatigue coefficient information of workers completing each assembly task; S2. Establish a task matching evaluation system and use fuzzy comprehensive evaluation method to calculate the task matching between workers and robots for each task; S3. Based on the collected information and the calculated task matching degree, a task allocation model that considers both task matching degree and human factors is constructed; S4. Use the non-dominated sorting genetic algorithm with elite strategy to solve the constructed task allocation model and output the optimal task allocation solution; Step S3 specifically includes: S31. Establish the assumptions and related parameters of the model; S32. Construct the objective function of the task allocation model considering task matching and human factors; S33. Determine the constraints of the objective function constructed according to the parameters; The objective function constructed in step S32 is expressed as: f1=MinT Where f1 represents minimizing the assembly time; f2 represents maximizing the total task matching degree; T represents the time point within a given range; c jk represents the task matching degree of operator k completing task j; x jlt represents the decision variable; The step S4 specifically includes: S41. Generate an initial chromosome population; S42. Calculate the chromosome fitness to obtain a population group; wherein the fitness includes unit product assembly time and total task matching degree; S43. Perform non-dominated sorting on all populations, select half of the individuals in the population to enter the next generation according to the Parato level, and determine whether the Parato levels are the same. If so, select the individual with greater crowding degree; S44. Two-point crossover method was used to perform chromosome crossover processing; S45. Use single-point mutation to mutate chromosomes; S46. Use a repair strategy to check whether the chromosomes after crossover mutation meet the fatigue constraint. If not, randomly assign a task completed by a worker to the robot until the fatigue constraint is met, thereby generating a new population. S47. The generated new population and the population in step S42 are combined into a new population, and the non-dominated sorting is performed again. According to the Parato level, 2 / 3 of the individuals in the population are selected to enter the next generation. It is determined whether the parato levels are the same. If so, the individual with a greater crowding degree is selected, and the number of iterations is increased by 1; S48. Determine whether the number of iterations reaches the set value. If so, output the final population; if not, return to step S43; S49. Normalize the indicator dimensions of the decision-making process to obtain the optimal allocation plan.

2. According to claim 1, a method for allocating tasks of human-machine collaborative assembly taking into account task matching and human factors is characterized in that: The step S2 specifically includes: S21. Select relevant indicators of the task matching evaluation system; the indicators include primary indicators and secondary indicators; S22. Use network analysis method to calculate the weight of each indicator in the primary and secondary indicators; S23. Based on the calculated weights, a fuzzy comprehensive evaluation method is used to calculate the task matching degree of each task for workers and robots.

3. The method for allocating human-machine collaborative assembly tasks considering task matching and human factors according to claim 2 is characterized in that: The step S22 specifically includes: S221. Use the 9-level scaling method to compare the importance of indicators in pairs, and perform consistency test to determine whether the consistency mark CI is less than 0.

1. If so, execute step S222; if not, continue to execute step S221; S222. Construct a judgment matrix corresponding to the indicator, and operate and normalize all judgment matrices to obtain a supermatrix, and weight the elements in the supermatrix to obtain a weighted supermatrix; S223. Perform limit operation on the weighted supermatrix to obtain a limit supermatrix, and obtain the limit relative ranking vector of each indicator element based on the limit supermatrix.

4. The method for allocating human-machine collaborative assembly tasks considering task matching and human factors according to claim 2 is characterized in that: The step S23 specifically includes: S231. Use the fuzzy comprehensive evaluation method to grade the task matching degree between the worker and the robot for each design requirement of each task; wherein the grading levels include high matching degree, relatively high matching degree, medium matching degree, relatively low matching degree, and low matching degree, wherein the corresponding interval numbers of high matching degree, relatively high matching degree, medium matching degree, relatively low matching degree, and low matching degree are {[0.8,1), [0.6,0.8), [0.4,0.6), [0.2,0.4), [0,0.2)}; S232. Score the matching degree between workers and robots for each indicator; S233. Based on the classification results and scoring results, the matching degree between the worker and the robot in each process is calculated, which is expressed as: c jk =Z mn ·w mn ·[1,0.8,0.6,0.4,0.2] / (1+0.8+0.6+0.4+0.2) Among them, c jk represents the task matching degree of the jth task completed by operator k; r mn0 It represents the membership degree of the oth level of the nth secondary indicator under the mth primary indicator; Z mn represents the membership matrix; v represents the total number of experts; u represents the number of experts given level o.

5. The method for allocating human-machine collaborative assembly tasks considering task matching and human factors according to claim 4 is characterized in that: Step S33 specifically includes: S331. The minimum assembly time per unit product is the latest time when all operators complete the assembly task, expressed as: Where T represents the time point within a given range; K represents the number of executable operators; t jk represents the time required for operator k to complete task j; x jkt represents the decision variable; j represents the process number in the task; k represents the operator; t represents time; S332. At any time, each process j is completed by only one operator, expressed as: S333. At any time, each process j is completed by only one operator, expressed as: Among them, J represents the number of processes; S334. Any subsequent task j in sequential assembly is not started until its previous task i has been completed, expressed as: where p j Indicates the index number of the task immediately preceding task j; S335. Only when all parallel processes of parallel assembly are completed can the next sequential process be started, which is expressed as: Among them, t ik represents the time required for operator k to complete task i; x ikt ,y ji represents a decision variable; M represents an infinite number; S336. The fatigue level of workers at any time should be within the safety threshold, expressed as: F oi ≤F max Among them, τ i represents the rest time of workers before completing task i; F(τ i ) indicates that the worker passes through τ i Fatigue before rest; R(τ i ) indicates that the worker has passed τ i Fatigue after rest; F oi represents the worker's fatigue level after completing the i-th process of the o-th product; μ represents the worker's fatigue recovery index; F o(i-1) represents the worker's fatigue after completing the i-1th process of the oth product; i represents the fatigue coefficient of the i-th task; t i F represents the time required for the operator to complete task i; max Indicates the safety threshold of worker fatigue.

6. The method for allocating human-machine collaborative assembly tasks considering task matching and human factors according to claim 5, characterized in that: The total task matching degree in step S42 is expressed as: Where C represents the total task matching degree; N represents the number of assembly tasks; c jk It represents the task matching degree of operator k completing task j.

7. The method for allocating human-machine collaborative assembly tasks considering task matching and human factors according to claim 6, characterized in that: In step S49, normalization processing is performed to obtain the optimal allocation solution, which is expressed as: B i =y′ i1 ×w1+y′ i2 ×w2 in, represents the worst value of objective function 1 among all solutions, represents the worst value of objective function 2 among all solutions, represents the optimal value of objective function 1 among all solutions, represents the optimal value of objective function 2 among all solutions; y′ i1 , y′ i2 are the dimensionless values ​​after being processed by objective function 1 and objective function 2; B i represents the final comparison value of the i-th solution, where all B i Medium B i The i-th solution corresponding to the minimum value is the optimal allocation solution.

Images