Research on the shortest path planning for non-standard workpiece Arc welding based on genetic algorithm

Efficient path planning for robotic arc welding of non-standard workpieces with spatially distributed welds is challenging due to the trajectory optimization complexity. To address this, a path optimization method for multi-plane structural components is proposed. By adapting the traveling salesman problem framework, a mathematical model for multi-linear weld path planning is established. The approach defines key process constraints including primary welding path, transition path, and inter-node spacing, and employs a customized genetic algorithm with adaptive operators and fitness functions to obtain near‑optimal solutions. An integrated technical framework is further developed, enabling workpiece modelling and positioning via VTK and point cloud processing, as well as welding data extraction and motion commands conversion through VTK Widget interaction and Karel-Socket communication. Experiments on self-developed multi-plane components show stable convergence under varying population sizes and elite ratios. This confirms the robustness and practical value of the proposed method for improving welding efficiency.