Task Planning Research Considering Operational Fatigue On Human-Robot Collaboration

Human-robot collaboration(HRC)is an emerging manufacturing model that can meet the needs of modern manufacturing for personalization,customization and small batch production.Human-robot collaboration combines respective advantages of humans and robots,leveraging the dexterity,perception,and decision-making capabilities of humans and the repetitive,hard-working,and highly accurate characteristics of robots.Leveraging the complementary capabilities and characteristics of workers and robots can make manufacturing systems more flexible,safe,efficient,and cost-effective.Despite the great potential of human-machine collaboration,the very different characteristics of humans and robots increase the complexity of the system and pose a great challenge for collaborative task planning.Ensuring the effective cooperation between humans and robots through collaborative task planning is one of the key issues of HRC.In this paper,we consider the factors of operational fatigue and the productivity of HRC.In this paper,we conduct an in-depth study on HRC task planning from the perspective of task classification and task scheduling.The main research contents of this paper are listed as below.(1)Machine learning-based task classification for HRC is studied.Task classification is the foundation of HRC task scheduling.The goal of task classification is to determine the best executor type for the task.The consequences of task assignment are analyzed from the perspective of task execution,and a cost matrix of misclassification is established.A cost-sensitive HRC task classification method is proposed with the goal of minimizing the total misclassification cost from the perspective of task decision outcomes.The cost-sensitive task classification method is implemented by solving the combined optimization of the ECOC coding matrix and feature selection by the imperial competition algorithm,and the improved errorcorrecting output coding(ECOC)algorithm for decoding output based on feature weights.Numerical experiments demonstrate that this method is outperformed the 6 comparison classification methods.(2)The HRC task scheduling problem based on the shortest job-cycle is studied.A novel task scheduling model for HRC is proposed,which integrates the rest time of workers inside job-cycle to provide workers with appropriate rest-recovery time by taking advantages of HRC.Based on the task scheduling model,a genetic-based method is proposed to solve the task scheduling problem of job-cycle minimization with the constraint of worker’s fatigue.The experimental results demonstrate that the proposed task scheduling model has significant advantages in shortening the job-cycle.The proposed task scheduling model outperformed the comparison model in 91.6% of scenarios.(3)A multi-objective HRC task scheduling problem with minimization of the job-cycle and the worker fatigue is studied.The existing task scheduling model is optimized and the mathematical job-cycle expression is proposed.Furthermore,the multi-objective optimization problem is solved by an improved chemical reaction algorithm(ICRO).The four operators of the CRO are improved with a greedy strategy to enhance the solution of the task scheduling problem.Numerical experiments demonstrate that the improved chemical reaction algorithm based on the greedy strategy significantly outperforms the 3 comparison algorithms.(4)Base on above studies of task scheduling with determined performing time,the HRC task scheduling problem with undetermined performing time is studied.In the task scheduling,the task performing time of human workers is represented by the triangular fuzzy number(TFN).A new TFN stochastic simulation method is proposed to solve the cycle time and worker fatigue with Monte Carlo method.Furthermore,a decomposition framework-based butterfly algorithm(MOEA/D-BOA)is proposed for solving the Pareto set of the multiobjective task scheduling problem.The experimental results demonstrate the feasibility and effectiveness of the proposed algorithm are better than other comparison algorithms.(5)A comprehensive simulation production experiment of HRC cable assembly is built.With the data testing and statistical analysis of the participants’ operation status,the participants’ operation fatigue is monitored.By comparing the fatigue of monitoring results with the calculation fatigue from the task scheduling model,the effectiveness of the proposed HRC task scheduling model of this paper is verified.