A comparison of Genetic Algorithms (GA), Gradient Descent (GD) and Simulated Annealing (SA) and how they differ in effectiveness when optimizing linkage lengths of Jansen's Linkage to achieve a desired foot trajectory

Jansen's linkage, a planar thirteen-bar single-degree-of-freedom mechanism designed by Theo Jansen in the 1990s, converts rotary motion into efficient walking gaits for applications in legged robotics and terrain navigation. Determining optimal linkage parameters that produce desired foot trajectories presents a challenging non-linear, multi-parameter optimization problem. This study compares three canonical optimization algorithms: Gradient Descent (GD), Genetic Algorithm (GA), and Simulated Annealing (SA). And evaluates their effectiveness in minimizing trajectory deviation from Jansen's established mechanism design.

GitHub: github.com/XenenDev/Jansen-Linkage-Research