In order to solve the problems of poor target optimization and insufficient planning accuracy of live-working robots in complex environments, this paper proposes a multi-objective optimization method based on genetic algorithm, combined with dynamic weight adjustment and real-time perception technology, aiming to achieve high efficiency, safety and energy balance in path planning. Firstly, a path planning model is constructed based on typical scenarios of live working (such as lead installation and wire breaking operations). The shortest path length, highest obstacle avoidance efficiency and lowest energy consumption are taken as optimization goals, and the robot’s motion range and working environment constraints are set. Then, a genetic algorithm is used to design a multi-objective optimization scheme and transform the path planning problem into a fitness function solution. The priority of each optimization objective is adjusted in real time through a dynamic weight mechanism, while gene retention and adaptive mutation operations are used to improve the algorithm’s search capability and convergence efficiency. Finally, by combining lidar and visual sensors, dynamic obstacle information in the working environment is obtained, and obstacles are avoided through local path replanning strategies to ensure the stability of robot operations and optimize subsequent task paths. The experimental results show that the average length of the unoptimized path is 21.56 meters, and the average length of the optimized path is 18.22 meters. The obstacle avoidance success rate can reach 97.7%, effectively solving the problem of path planning in complex environments and significantly improving work efficiency and safety. The multi-objective optimization method proposed in this paper achieves an effective balance between path length, obstacle avoidance efficiency and energy consumption, and improves the planning performance of the live-line working robot in complex environments. It helps to optimize the real-time decision-making ability in dynamic environments and promote the intelligent development of power operations.
Xu et al. (Thu,) studied this question.