Ali Joodi Aalhasan, Bao Song, and Yongxing Liu
Industrial robotics, collision-free workspace, optimizing manipulatorsmotion, non-dominated sorting genetic algorithm II (NSGA-II),Gilbert–Johnson–Keerthi (GJK) algorithm
Ensuring efficient and collision-free motion in industrial robotics is essential for productivity and safety, especially in complex workspace environments. This paper presents an innovative approach by integrating the non-dominated sorting genetic algorithm II with the Gilbert–Johnson–Keerthi algorithm for optimised collision avoidance and motion planning. Unlike prior methods, this integration allows for precise collision detection and seamless optimisation of multiple objectives, such as jerk minimisation, motion smoothness, and computational efficiency. Our approach utilises hybrid trajectory modelling with B-Splines and B´ezier curves to achieve fluid transitions between waypoints, combined with Spherical Linear Rotation Interpolation to ensure accurate orientation control. This method has been validated in simulations using CoppeliaSim and practical experiments with TA6-R5, a 6-DoF robotic arm, demonstrating notable improvements in reducing joint jerks and maintaining motion stability. The experimental results demonstrate notable improvements in robot motion efficiency and joint dynamics, with a notable reduction in jerk relative to previous benchmarks. This approach using a 6-DoF robot allowed us to establish a practical baseline for our method, demonstrating its applicability across simpler and more complex robotic systems.
Important Links:
Go Back
