Optimisation of Cubic Polynomial Robot Motion Using Flexible Polyhedron Search, Weighted-Sum-GA and Pareto-Based GA: A Comparative Stdy

This project was concerned with the motion optimisation of a SCARA robot subject to kinematics constraints, where two methods are used, namely: genetic-based algorithms and the flexible polyhedron search. Both methods use cubic spline functions to generate motion profiles. In formulating the GA's, a tailor-fit operators and procedures were used to seek an application-dependent structure. In addition, an initial evaluation is reported in the form of comparisons between a Pareto-based and weighted-sum (parametric) approaches to multi-criteria optimisation. Case study results are summarised for the RTX robot with six joints.