Tazaki, Yuichi (Tokyo Inst. of Tech.), Imura, Jun-ichi (Tokyo Inst. of Tech.)

Graph-Based Model Predictive Control of a Planar Bipedal Robot

Scheduled for presentation during the Regular Session "Model Predictive Control I" (MoA10), Monday, July 24, 2006, 12:05−12:30, Room 103

17th International Symposium on Mathematical Theory of Networks and Systems, July 24-28, 2006, Kyoto, Japan

This information is tentative and subject to change. Compiled on April 26, 2024

Abstract

In this paper, we propose a new offline model predictive control for a planar bipedal robot, which we call here Graph-based Model Predictive Control. This method consists of two phases : the graph construction phase and the realtime control phase. The directed graph is constructed off line by i)placing a certain number of nodes on the state space of the robot, ii)computing the optimal path starting from each of the graph nodes with respect to a given cost function, and iii)creating a set of directed edges by taking the first edge of each of the optimal paths. This means that one can achieve receding horizon control in some sense by simply tracing the edges of the directed graph and therefore the real-time computational cost is dramatically reduced compared with the ordinary MPC. In addition, by constructing multiple directed graphs based on different cost functions, one can design multiple motions and switching trajectories among them in a uniform way. The proposed method is applied to the speed changing control problem of a bipedal walker on a two-dimensional plane and its effectiveness is verified by numerical simulation.