An Iterated Greedy Algorithm for Inverse Kinematics of Binary Hyper-Redundant Manipulators

This paper presents a novel solution to the inverse kinematics (IK) problem for hyper-redundant manipulators that use discrete, binary actuators. These robots have a finite workspace, making conventional IK solvers for continuous robots ineffective. We introduce the Multi-Module Search (M-MS) algorithm, an iterative greedy method that efficiently navigates the discrete configuration space. The M-MS method works by iteratively selecting a group of modules and finding their optimal configuration combination to minimize the end-effector error. The algorithm's performance is rigorously tested on both planar and spatial manipulator models. Results demonstrate a key trade-off: searching more modules simultaneously reduces final error but increases computational cost. Furthermore, the proposed M-MS method is shown to outperform several established meta-heuristic algorithms, including Particle Swarm Optimization and Genetic Algorithms, in terms of both accuracy and robustness, establishing it as a superior approach for real-time control of discretely actuated systems.

Keywords:

inverse kinematics , hyper-redundant manipulator , discrete actuation , multi-module search , binary robot , discrete optimization