Cooperative tracking of vehicle platoons subjected to uncertain dynamics, external disturbance and bounded leader input

This thesis addresses three main issues in the design of distributed controllers for vehicle platoons, namely (i) suppressing the effects of uncertainty, (ii) guaranteeing stability, string stability, and consensus when the platoon is subjected to unknown external disturbances and time-varying bounded leader input and (iii) being applicable to both directed and undirected topologies. Two control approaches based on model reference control scheme are proposed, namely (i) distributed model reference adaptive control (DMRAC) and (ii) distributed model reference control (DMRC). The approaches are designed based on which issues are given more importance in the controller design. DMRAC consists of a reference model and a main control system. Cooperative state variable feedback control (CSVFB) is used as a reference control in the reference model. The main control system utilized CSVFB as a nominal control to track the leader state and a proposed adaptive term to suppress the effect of uncertainty in the inertial time lag and control effectiveness of the followers. Stability analysis shows that global tracking errors of followers with respect to the reference model and with respect to the leader are asymptotically stable. The states of all followers synchronize to both the reference model and leader states. Moreover, with the existence of unknown external disturbances, the global tracking errors remain uniformly ultimately bounded. DMRC consists of a cooperative tracking error reference as a virtual reference and a main control system. The virtual reference is designed based on CSVFB, while the main control system is designed using cooperative tracking error and cooperative disagreement error to attenuate the effects of unknown external disturbance and nonzero bounded leader input. The global disagreement error is shown to be uniformly ultimately bounded and input-to-state string stability (ISSS) through detailed stability analysis, such that the states of each follower synchronize to the leader states with bounded residual error. Performance verifications are conducted through simulations and validate the efficacy of DMRAC and DMRC for the vehicle platoons problem. Both controllers can be applied to vehicle platoon with both directed and undirected topologies. The directed topology must contain at least one spanning tree with the leader as a root node, while the undirected topology must be static and connected with at least one follower receiving information from the leader. Simulation results demonstrate that DMRC outperforms DMRAC when the leader moves with time-varying velocity, i.e., the state of each follower synchronizes to the leader and string stable.