Partially-observed decentralized optimal control for large population two-wheeled vehicles: A differential game approach

Abstract

In this paper, we consider the decentralized noncooperative optimal control problem for a large population of two-wheeled unmanned vehicles with partial observations. Specifically, an individual vehicle is controlled only by its local noisy information measured from a local sensor in order to follow the average behavior (mean field) of the entire population while achieving the overall optimal control performance. We solve the partially-observed linear-quadratic mean field game to obtain decentralized optimal controls for two-wheeled vehicles. These controls are decentralized since they are functions of the local estimated state from the local Kalman filter. We show that the set of the decentralized optimal controls constitutes an E-Nash equilibrium, where E converges to zero as the number of vehicles becomes large. Finally, the theoretical results are validated through simulations and experiments with various operation scenarios for a large population of two-wheeled vehicles. (C) 2020 The Franklin Institute. Published by Elsevier Ltd. All rights reserved.