Robotic Simulator with High-Precision Perception of Contact Dynamics

Yongcan Zhou,Yang Pan,Tianjian Lei,Junpeng Chen

DOI: https://doi.org/10.1109/icdl55364.2023.10364550

2023-01-01

Abstract:A robotic simulator constitutes a pivotal tool in achieving accurate modeling and simulation within a virtual environment, especially concerning comprehensive robot training. This paper introduces a novel robotic simulator that employs a continuous compliant contact model, ensuring continuous and stable calculations during the contact process, with a specific focus on precise contact force computation during object collisions. The proposed physics engine captures energy transformations for diverse contact surface shapes and material properties, employing spring-dampers in collision detection. Dynamic equations are derived based on the spring stiffness system, damping coefficient, and external forces applied during the contact phase, enabling effective determination of dynamic parameters throughout the contact process, including position, velocity, and contact force during penetration. Experimental validation of the physics engine's contact solver involves investigating bouncing balls under varying coefficients of restitution. The conducted experiments successfully capture continuous variations in parameters such as contact force during the penetration process. The primary contribution of this paper lies in advancing physics engines to achieve accuracy in simulating and modeling virtual environments, particularly when applied to large-scale robot training. Beyond this context, the high-precision perception of contact dynamics within our robotic simulator bears potential implications in areas such as industrial automation, robotic manipulation, and virtual prototyping.