An obstacle-avoidance inverse kinematics method for robotic manipulator in overhead multi-line environment

The inverse kinematics problem plays a crucial role in robotic manipulator planning, autonomous control, and object grasping. This problem can be solved in simple environments based on existing studies. However, it is still challenging to quickly find a feasible inverse kinematic solution when obstacle avoidance is required. In this paper, we present a nonconvex composite programming method to solve the inverse kinematics problem with overhead obstacle-avoidance requirements. Our method enables efficient obstacle avoidance by directly calculating the minimum distance between the manipulator and the overhead environment. We construct end-effector error functions based on the Product of Exponentials model and explicitly provide their gradient formula. We derive the minimum distance based on the geometry parametric equation and directly utilize it to construct the obstacle avoidance function. We propose an enhanced version of adaptive moment estimation based on short-time gradient information to improve optimization performance. Finally, we conduct simulations and experiments in overhead line environments. Comparative results with other optimization methods demonstrate that our proposed method achieves a high success rate with a low solution time.