Optimized inverse kinematics solutions for a 6-DOF robot

Résumé

Inverse kinematics (IK) for surgical robotic manipulators requires a balance between computational speed, pose accuracy, and robustness to poor initialization and near-singular configurations. This paper presents a controlled comparative evaluation of optimization-based IK solvers for a 6-DOF industrial robot (Viper 650s) under a unified constrained optimization framework with identical robot modelling assumptions, constraints, trajectories, and evaluation metrics. Six methods are investigated, spanning standalone and hybrid techniques: Sequential Quadratic Programming (SQP), Branch and Bound (B&B), Ant Colony Optimization (ACO), and three hybrid configurations (ACO-SQP, B&B-SQP, and a three-stage ACO-B&B-SQP pipeline). All methods are assessed using the same evaluation framework to examine the practical effect of combining global exploration, bounded search-space reduction, and local refinement on path-following accuracy. Two representative trajectories (a planar triangle and a spatial helix) are executed with repeated trials, and performance is evaluated in terms of trajectory execution time and geometric path-tracking error, supplemented by repeated-run statistical summaries. The optimization methods are implemented in MATLAB and tested within the Automation Control Environment (ACE). The results show that the ACO-B&B-SQP pipeline achieved the lowest mean path error on both trajectories under the present setup, whereas SQP remained the fastest. The study is positioned as a unified comparative evaluation framework for optimization-based IK on the Viper 650s robot, incorporating the implementation and evaluation of a three-stage ACO-B&B-SQP hybrid pipeline within this common framework.

Citer ce document

Accès au document

Texte intégral en lecture en ligne, réservé aux abonnés SPHAERO et aux membres de l'institution. Se connecter

Auteur(s)

Statistiques

Consultations : 1

Téléchargements : 0