Automatically fine-tuned speed control system for fuel and travel-time efficiency: A microscopic simulation case study

摘要

Within the current document a model independent, cognitive and adaptive optimization mechanism, namely CAO, is adopted for providing efficient speed/torque control actions. However since real-life tests were not feasible, a simulation model of a 1.4lt displacement gasoline car, playing the role of the actual car, was adopted while the potential maximum cruising speed levels are chosen so as to emulate a usual suburban route and the vehicle speed control is replicated by a common PID scheme. The control decisions are applied directly to the car throttle/torque pedal itself. The goal of the optimization application was to minimize the vehicle velocity error, while minimizing the total fuel consumption for a certain 10km trip with varying road angles/slopes. It should be noted that CAO module could be applied directly to a car system in a straightforward manner without any preparatory investigations. Initially PID gain values were selected arbitrary while a PID gain - tuning module from Matlab/Simulink was used in order to fine-tune them under certain road conditions. Finally the tuned values were used as initial ones for all CAO's application cases (A, B, C and D). CAO presented substantial improvements in the specified performance index, with respect to the base case speed strategy, as well as to the PID tuning in all simulation scenarios considered.