Engine-in-the-loop study of a hierarchical predictive online controller for connected and automated heavy-duty vehicles 2020-01-0592
This paper presents a cohesive set of engine-in-the-loop (EIL) studies examining the use of hierarchical model-predictive control for fuel consumption minimization in a class-8 heavy-duty truck intended to be equipped with Level-1 connectivity/automation. This work is motivated by the potential of connected/automated vehicle technologies to reduce fuel consumption in both urban/suburban and highway scenarios. We begin by presenting a hierarchical model-predictive control scheme that optimizes multiple chassis and powertrain functionalities for fuel consumption. These functionalities include: vehicle routing, arrival/departure at signalized intersections, speed trajectory optimization, platooning, predictive optimal gear shifting, and engine demand torque shaping. The primary optimization goal is to minimize fuel consumption, but the hierarchical controller explicitly accounts for other key objectives/constraints, including operator comfort and safe inter-vehicle spacing. The main focus of this work is on a sequence of EIL studies intended for evaluating the computational costs and fuel savings associated with these algorithms. These EIL studies involve both the open-loop playback of simulation-based evaluation studies as well as the closed-loop validation of the proposed control strategies, both individually and combined. These studies show that this hierarchy of algorithms is capable of running in real time, with the round-trip communication delay inherent in EIL simulation being one of the key factors affecting the overall fidelity of the EIL results. Moreover, the EIL studies are encouraging, both in terms of the successful hierarchical integration of the underlying algorithms and the preliminary fuel savings seen in the EIL tests. In particular, the EIL results suggest that an aggressive overall goal of reducing vehicle fuel consumption by 20% or more is potentially achievable, especially in urban/suburban scenarios.
Chu Xu, Ben Groelke, Miguel Alvarez Tiburcio, Christian Earnhardt, John Borek, Evan Pelletier, Stephen Boyle, Brian Huynh, Mohamed Wahba, Stephen Geyer, Christopher Graham, Mark Magee, Kyle Palmeter, Mohammad Naghnaeian, Sean Brennan, Stephanie Stockar, Christopher Vermillion, Hosam Fathy
The University of Maryland, North Carolina State University, The University of North Carolina, Penn State University, Ohio State University, Volvo Group North America, Clemson University