Considerate and Cooperative Model Predictive Control for Energy-Efficient Truck Platooning of Heterogeneous Fleets

Abstract

Connectivity-enabled automation of distributed control systems allow for better anticipation of system disturbances and better prediction of the effects of actuator limitations on individual agents when incorporating a model. Automated convoy of heavy-duty trucks in the form of platooning is one such application designed to maintain close gaps between trucks to exploit drafting benefits and improve fuel economy, and has traditionally been handled with classically-designed connected and adaptive cruise control (CACC). This paper is motivated by demonstrated limitations of such a control strategy, in which a classical CACC was unable to efficiently handle real-world road grade and velocity transient disturbances without the assistance of fleet operator intervention, and is non-adaptive to varied hardware and loading conditions of the operating truck. This automation strategy is addressed by forming a cooperative model predictive control (MPC) for eco-platooning that considers interactions with trailing trucks to incentivize platoon harmonization under road disturbances, velocity transients, and engine limitations, and further improves energy economy by reducing unnecessary engine effort. This is accomplished for each truck by sharing load, maximum engine power, transmission ratios, control states, and intended trajectories with its nearest neighbors. The performance of the considerate and cooperative strategy was demonstrated on a real-world driving scenario against a similar non-considerate control strategy, and overall it was found that the considerate strategy significantly improved harmonization between the platooned trucks in a real-time implementable manner.