2019 · 12 citations · 15 references
EngineeringAutomotive EngineeringMachine LearningPowertrain EngineeringVehicle TestingAdaptive ControlFuel Economy OptimizationIntelligent Transportation SystemsKalman FilteringPowertrain Control SystemsVehicle Control SystemsAerodynamic DragDrag PredictionFuel EconomyNeural Network ControlNeural Network-based EstimationPowertrain OptimizationRolling Resistance AnalysisVehicle State Estimation
The new stringent standards laid down for reducing pollution and greenhouse gas emissions in transportation sector compel the vehicle manufacturers to use different technologies to decarbonize the vehicles and reduce their energy consumption. Powertrain control optimization, especially based on instantaneous accurate information, has an important role in improving the fuel economy of vehicle. However, its successful implementation requires an accurate parameter estimation method. The main purpose of this article is to propose an accurate adaptive method for estimating the rolling resistance and aerodynamic drag, which are two important factors in powertrain optimization of a vehicle. In this regard, two online adaptive methods, namely recursive least squares (RLS) and Kalman filter (KF), are formulated for the parameters estimation of the powertrain by using the experimental data from a real driving cycle and their performances are compared with an offline trained artificial neural networks (ANN). The experimental and simulation results show that KF is more accurate than RLS in terms of predicting the road rolling resistance and aerodynamic drag and its performance is very near to the developed ANN.
15