In recent works, model-based nonlinear control techniques, such as sliding mode control, model predictive control, model reference adaptive control, model-based impedance control, geometric Jacobian method, an adaptive pressure-based PD controller, finite-element method based control, and a quasi-static model-based controller, have been used to create more effective, robust, and adaptive control algorithms. Although these model-free based techniques eliminate the need for any knowledge about the dynamic model of these highly compliant systems, their performance strongly depends on the learning process (i.e., the amount of data and training for specific conditions and cases) which makes their successful implementation mainly limited to periodic/repetitive tasks with fixed patterns and environmental conditions (e.g., mimicking locomotion gaits on a specific surface).
0 kommentar(er)
