A Simulation Parameter Identification Method for Single Peg and Hole Assembly Based on Virtual-Real Comparison

In recent years, intelligent algorithms represented by reinforcement learning have been increasingly applied to solve single peg and hole assembly tasks. These intelligent algorithms put forward higher requirements on the accuracy of assembly contact force simulation. This paper proposes a simulation parameter identification method based on minimizing the virtual and real contact force error to identify difficult-to-determine parameters like friction coefficient, contact stiffness, and contact damping. This ensures that the simulation results are close to the actual single peg and hole assembly contact force. A single peg and hole assembly experimental verification shows that the proposed method can identify those simulation parameters, and the simulation parameters obtained have a high degree of accuracy.