Robust Data-EnablEd Predictive Leading Cruise Control via Reachability Analysis

Data-driven predictive control promises modelfree wave-dampening strategies for Connected and Autonomous Vehicles (CAVs) in mixed traffic flow. However, the performance suffers from unknown noise and disturbances, which could occur in offline data collection and online predictive control. In this paper, we propose a Robust Data-EnablEd Predictive Leading Cruise Control (RDeeP-LCC) method based on reachability analysis, aiming to achieve safe and optimal control of CAVs under bounded process noise and external disturbances. Precisely, we decouple the mixed platoon system into an error system and a nominal system, and tighten the constraint via the data-driven reachable set technique. Then, the enhanced safety constraint is integrated with the data-driven predictive control formulation to achieve stronger robust control performance for CAVs. Simulations validate the effectiveness of the proposed method in mitigating traffic waves with better robustness.