DeeP-LCC: Data-EnablEd Predictive Leading Cruise Control in Mixed Traffic Flow

Abstract

For the control of connected and autonomous vehicles (CAVs), most existing methods focus on model-based strategies. They require explicit knowledge of car-following dynamics of human-driven vehicles that are non-trivial to identify accurately. In this paper, instead of relying on a parametric car-following model, we introduce a data-driven non-parametric strategy, called DeeP-LCC (Data-EnablEd Predictive Leading Cruise Control), to achieve safe and optimal control of CAVs in mixed traffic. We first utilize Willems' fundamental lemma to obtain a data-centric representation of mixed traffic behavior. This is justified by rigorous analysis on controllability and observability properties of mixed traffic. We then employ a receding horizon strategy to solve a finite-horizon optimal control problem at each time step, in which input/output constraints are incorporated for collision-free guarantees. Numerical experiments validate the performance of DeeP-LCC compared to a standard predictive controller that requires an accurate model. Extensive nonlinear traffic simulations further confirm its great potential on improving traffic efficiency, driving safety, and fuel economy.

Experimental validation of DeeP-LCC for dissipating stop-and-go waves in mixed traffic

“…In this section, we review the basics of the DeeP-LCC strategy [18], [19] for CAV control in mixed traffic flow.…”

“…After specifying the inputs and outputs, a discrete-time state-space model for the mixed traffic after linearization can be derived as follows [18], [19] x…”

“…as the past control input sequence of length T ini and the future control input sequence of length N , respectively (similarly for ini , and y ini , y). Under a very mild condition, we have the following result [19,Proposition 2]. Proposition 1: Define the pre-collected combined input sequence as ûd = col u d (1), d (1) .…”

“…Combining the recent Data-EnablEd Predictive Control (DeePC) [17] with the LCC framework [10], we have proposed a Data-EnablEd Predictive Leading Cruise Control (DeeP-LCC) strategy in mixed traffic [18], [19]. DeeP-LCC enables the CAVs to directly utilize measurable traffic data to design safe and optimal control for CAVs.…”

“…Input/output constraints are incorporated in DeeP-LCC for collision avoidance and actuation saturation. Extensive simulations have confirmed the potential of DeeP-LCC in improving traffic efficiency and fuel economy with safety guarantees [19]. To the best of our knowledge, however, data-driven techniques for stabilizing traffic flow have only been verified via traffic simulations and virtual data; see, e.g., [15], [16], [18], [20].…”