Combined Cooperative Adaptive Cruise Control using Collective Initial Excitation based Distributed Parameter Estimator

Abstract

In cooperative adaptive cruise control (CACC), autonomous vehicles are grouped into a string of platoon and, the main objective is to automatically adapt their speed using on-board sensors and communication with the preceding vehicle to maintain a desired inter-vehicle distance. Cruise control is achieved in the presence of parametric uncertainty in the vehicle dynamics using principles of adaptive control. This work proposes a novel combined CACC strategy for an uncertain homogeneous platoon with guaranteed parameter convergence and asymptotic string stability. A novel distributed consensus-based parameter estimator is proposed in conjunction with a model reference adaptive control (MRAC) algorithm using a direct control-gain update law. The algorithm ensures exponential parameter estimation error convergence to zero as well as asymptotic convergence of tracking-error to zero. Conventional CACC protocols require a condition of persistence of excitation (PE) for parameter convergence, which is required for better transient performance in converging to a string stable configuration. The PE condition is highly restrictive in the context of cruise control since velocity profiles which are demanded in the platoon model do not typically satisfy the PE condition. In contrast, the proposed scheme can ensure parameter convergence under a significantly milder condition, coined as collective initial excitation (C-IE). The C-IE condition is an extension of the concept of initial excitation (IE), which is recently proposed in the context of adaptive control of single agent system. Unlike IE, the C-IE condition caters to distributed estimation in the context of multi-agent systems. As far as the authors are aware, this is the first work on CACC framework, which ensures exponential convergence of parameter estimation error of each vehicle under the mild condition of C-IE, which further leads to asymptotic convergence of the entire vehicle platoon to a string stable configuration. Simulation study dictates that the proposed CACC architecture outperforms the existing CACC algorithms in terms of tracking and estimation performance.