Near-optimal dynamic trajectory generation and control of an omnidirectional vehicle

This paper describes a computationally inexpensive, yet high performance trajectory generation algorithm for omnidirectional vehicles. It is shown that the associated non-linear control problem can be made tractable by restricting the set of admissible control functions. The resulting problem is linear with coupled control efforts and a near-optimal control strategy is shown to be piecewise constant (bang-bang type). A very favorable trade-off between optimality and computational efficiency is achieved. The proposed algorithm is based on a small number of evaluations of simple closed-form expressions and is thus extremely efficient. The low computational cost makes this method ideal for path planning in dynamic environments. (C) 2003 Published by Elsevier B.V.