Powertrain Control for Hybrid-Electric Vehicles Using Supervised Machine Learning

Abstract

This paper presents a novel framework to enable automatic re-training of the supervisory powertrain control strategy for hybrid electric vehicles using supervised machine learning. The aim of re-training is to customize the control strategy to a user-specific driving behavior without human intervention. The framework is designed to update the control strategy at the end of a driving task. A combination of dynamic programming and supervised machine learning is used to train the control strategy. The trained control strategy denoted as SML is compared to an online-implementable strategy based on the combination of the optimal operation line and Pontryagin’s minimum principle denoted as OOL-PMP, on the basis of fuel consumption. SML consistently performed better than OOL-PMP, evaluated over five standard drive cycles. The EUDC performance was almost identical while on FTP75 the OOL-PMP consumed 14.7% more fuel than SML. Moreover, the deviation from the global benchmark obtained from dynamic programming was between 1.8% and 5.4% for SML and between 5.8% and 16.8% for OOL-PMP. Furthermore, a test-case was conducted to emulate a real-world driving scenario wherein a trained controller is exposed to a new drive cycle. It is found that the performance on the new drive cycle deviates significantly from the optimal policy; however, this performance gap is bridged with a single re-training episode for the respective test-case.

On Machine Learning-Based Control for Energy Management in Construction Machines

“…Other similar approaches combining NN learning from optimisation results from DP or other control optimisation techniques were used by other authors as well for the control of on-road vehicles [83][84][85][86][87][88][89][90][91][92]. What is observed from this collection of studies on the topic is the ability of the networks to learn complex decision-making processes from system states and other measured variables and implement them in both simulation domains and in real systems.…”

Project and Development of a Reinforcement Learning Based Control Algorithm for Hybrid Electric Vehicles

“…In [8], an exhaustive chronological review of the most famous real-time capable control methods for (P)HEVs is reported. Starting from the 1990s, more and more sophisticated techniques have been developed, from very simple rule-based EMSs [8], passing through equivalent consumption minimization strategies (ECMS) [9,10] and dynamic programming (DP) [11] related optimizers, until the latest findings of model predictive control (MPC) [12] and learning algorithms [13,14]. Among the latter, Reinforcement Learning (RL) has started to be effectively exploited for control problems from 2015 onwards [8,15]; hence, it can be considered as an up-to-date approach.…”

Development Trends in Vehicle Propulsion Sources—A Short Review

“…The target fuel of future vehicles is hydrogen, which would either be used as the sole fuel [49,50] or as an additive to other fuels [51,52]. Energy management now plays an important role in hybrid or battery systems in the same way as with ICE [53][54][55]. Determining the characteristic ratios of different propulsion sources and whether their shape/path matches with the approximating functions is important for computing and modelling the performance of different vehicle powertrains [56,57].…”