Optimal-Horizon Model-Predictive Control with Differential Dynamic Programming
This work addresses the problem of adaptive horizon selection in model-predictive control for robotics, offering a novel method that is incremental in improving control flexibility.
The paper tackles trajectory optimization by developing an algorithm that determines the horizon online within the Differential Dynamic Programming framework, achieving exact one-step convergence for linear quadratic time-invariant problems and enabling real-time nonlinear model-predictive control, with demonstrated efficacy in obstacle-avoidance for planar robots.
We present an algorithm, based on the Differential Dynamic Programming framework, to handle trajectory optimization problems in which the horizon is determined online rather than fixed a priori. This algorithm exhibits exact one-step convergence for linear, quadratic, time-invariant problems and is fast enough for real-time nonlinear model-predictive control. We show derivations for the nonlinear algorithm in the discrete-time case, and apply this algorithm to a variety of nonlinear problems. Finally, we show the efficacy of the optimal-horizon model-predictive control scheme compared to a standard MPC controller, on an obstacle-avoidance problem with planar robots.