Fused Model Predictive Control Techniques for Strategic Platooning Amongst Heterogeneous Pairs of Heavy-Duty Trucks

Abstract

With pairs or groups of heterogeneous vehicles (with different masses, aerodynamic coefficients, etc.), collaborative platooning can be advantageous in some scenarios due to aerodynamic drag reduction, while being detrimental in other scenarios due to mismatches in vehicle properties. This paper introduces two controllers capable of alternating between independent vehicle velocity trajectory optimization (VTO) and a collaborative platooning/VTO approach based on the aggregate fuel savings of all vehicles within the platoon. The first uses the difference in mass between the vehicles within a platoon and the upcoming road grade to decide whether platooning will be economically advantageous, relying on a support vector classification algorithm to make the switching decision. The second runs both independent VTO and collaborative VTO/platooning in parallel, making a decision based on which method predicts the least amount of fuel consumption over an upcoming stretch of highway. The performance of these techniques was evaluated using a medium-fidelity Simulink model of a heavy-duty truck. Results show a 5.1% to 14.1% decrease in fuel consumption for the following vehicle of a platoon as compared to a baseline controller not platooning, where the exact fuel consumption improvement depends on the desired following distance. These results were also compared to a baseline that platooned over the entire route, providing evidence that there are situations where disengaging from a platoon is beneficial in the presence of heterogeneity.