Toshiyuki Ohtsuka

Kyoto University

Tuesday, December 03, 2019, 11:00

Room 01-012, Georges-Köhler-Allee 102, Freiburg 79110, Germany

In this talk, a real-time algorithm of nonlinear model predictive control (NMPC) is introduced for switched systems with state-dependent switches. The algorithm consists of two parts: 1) the local optimization of the control input and switching instants by solving the two-point boundary-value problem (TPBVP) for the whole horizon under a given switching sequence, and 2) the detection of an additional switch and the reconstruction of the solution taking into account the additional switch. We demonstrate the effectiveness of the proposed method through numerical simulations of a compass-like biped walking robot, which contains state-dependent switches and state jumps.