Fourier series-based algorithm for control optimization in pendulum capsule drive: an experimental study

This study presents a novel greedy version of a Fourier series-based algorithm for control optimization in a pendulum capsule drive. The method improves computational efficiency by optimizing the control function in smaller segments. A 3D-printed prototype with a PID-controlled actuation system validates the approach experimentally. Results show that the optimized control enhances locomotion, with speed increasing as more harmonics are used. Experimental findings closely match numerical predictions, with minor discrepancies attributed to friction modeling and mechanical backlash. The proposed method proves effective for optimizing control in nonlinear and discontinuous systems.