Robot pose optimization based on spectral clustering guided MOSSA considering the flexibility of robot structure and mounting base

Robot pose optimization is an effective method to improve industrial robot performance for machining tasks. However, when the robotic machining system is mounted on a flexible base, the influence of the flexibility deviation of the robot structure and the mounting base caused by external forces and dynamic characteristics on the robot accuracy is extremely complex, which increases the difficulty of obtaining the optimal robot pose. Therefore, this paper proposes a robot pose optimization method based on Spectral Clustering guided Multi-Objective Salp Swarm Algorithm (SC-MOSSA) and Technique for Order Preference by Similarity to Ideal Solution (TOPSIS), to obtain the optimal redundancy angle of the all robot poses for the machining path. Firstly, the application background of this paper is introduced, the influencing factors of robot machining accuracy in large spherical shell machining system are analyzed, and the influencing factors are characterized by the indexes composed of mathematical models. Then, a multi-objective optimization model with nonlinear constraints is established by combining the robot stiffness and base reaction performance index of two machining accuracy influencing factors. SC-MOSSA is proposed to solve the optimization model to calculate the Pareto front that meets the requirements, to reduce the computing time cost and ensure the diversity of solutions. The redundancy angle parameter for machining path is determined by TOPSIS. Finally, a robot machining platform for large spherical shell is established. Based on the helical milling enlarge hole machining path, the effectiveness of the proposed method is verified by simulation and experiment.