Robot arm planning and control by τ-Jerk theory and vision-based recurrent ANN observer

Abstract

This work describes a planning path-tracking control for a 6-axis robot manipulator in palettes assembly. Two biologically inspired approaches motivated this work: the general τ - Jerk theory for trajectory tracking and a recurrent bi-layer Hopfield artificial neural network. Equidistant Cartesian points generate free-collision paths between the robot and the palette. Nonlinear regression-based 3rd grade polynomials represents polynomial assembling trajectories. A variational method optimizes paths length. The method is validated through numeric simulations, showing feasibility and effectiveness.