SALSA-based motion optimization for robotic manipulators with strong nonlinear dynamic coupling

In this paper, the motion optimization for robotic manipulators is investigated by Support Area Level Set Algorithm (SALAS) through optimizing a high-dimensional nonlinear fitness function. The approach based on level set conception and the ability of Support Vector Machine(SVM) in distribution estimation integrates duel stages sampling strategies to avoid be converge in small search field too early and improve the rate of convergence to the potential solution. The simulations using two links robotic manipulators show satisfied results with SALSA.