A Hierarchical Economic Model Predictive Controller That Exploits Look-Ahead Information of Roads to Boost Engine Performance

Abstract

Sensors and communication capabilities of connected vehicles provide look-ahead information that can be exploited by vehicle controllers. This work demonstrates the benefits of look-ahead information combined with hierarchical economic model predictive control for the airpath management of compression ignition engines. This work exploits road information predicted with a 0.1-and 2-s horizon to simultaneously control fast and slow engine dynamics, respectively. It controls the variable nozzle turbocharger and dual-loop exhaust gas recirculation, at a 0.01-s rate, to simultaneously optimize NOx, soot, and fuel economy. Simulation studies and hardware-in-loop implementation on an ARM Cortex-A15 processor demonstrate improved NOx, soot, and torque tracking without compromising fuel economy, and a worst case computation time of 8.92 ms.