Stochastic Decentralized Control of a Platoon of Vehicles Based on the Inclusion Principle

In this paper the Stochastic Inclusion Principle is applied to decentralized Linear Quadratic Gaussian (LQG) suboptimal longitudinal control design of a platoon of automotive vehicles. Starting from a stochastic linearized platoon state model, input/state overlapping subsystems are identified and extracted after an adequate expansion. An algorithm for approximate LQG optimization of these subsystems is developed in accordance with their hierarchical lower-block-triangular (LBT) structure. Vehicle controllers obtained after contraction, which leaves the local Kalman filters uncontracted, provide high performance tracking and noise immunity. Simulation results illustrate characteristic properties of the proposed algorithm.