Deep reinforcement learning of energy management with continuous control strategy and traffic information for a series-parallel plug-in hybrid electric bus

Wu, Yao; Tan, Huachun; Peng, Jiankun; Zhang, Hailong; He, Haibo

doi:10.1016/j.apenergy.2019.04.021

Cited by 241 publications

(76 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The vehicle speed dataset is collected from the real driving conditions. To reduce the training time and improve the control accuracy, the vehicle velocity is divided into three-speed intervals that are [0-12] m/s, [12][13][14][15][16][17][18][19][20][21][22][23][24] m/s, [24][25][26][27][28][29][30][31][32][33][34][35][36] m/s, and they represent low, medium, and high speed, respectively. Then, the classified speed intervals are adopted to train the DDPG algorithm separately until the algorithm converges, the trained neural network is stored, as depicted in Fig.…”

Section: A a Bi-level Frameworkmentioning

confidence: 99%

“…et al employed the max-value-based policy and the random policy to reduce the overestimation for double Q-learning. Since the DQL cannot output continuous actions, Tan H. and Wu Y. et al [31], [32] proposed energy management strategies based on DDPG and simulation showed that DDPG-based strategy performs better than Q-learning and achieves results closed to DP. In [33] and [34], the authors join realtime topographic information, driver habits, and traffic condition in the training process of DDPG to improve the performance of EMS.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Transfer Deep Reinforcement Learning-Enabled Energy Management Strategy for Hybrid Tracked Vehicle

Guo

Tang

et al. 2020

IEEE Access

View full text Add to dashboard Cite

This paper proposes an adaptive energy management strategy for hybrid electric vehicles by combining deep reinforcement learning (DRL) and transfer learning (TL). This work aims to address the defect of DRL in tedious training time. First, an optimization control modeling of a hybrid tracked vehicle is built, wherein the elaborate powertrain components are introduced. Then, a bi-level control framework is constructed to derive the energy management strategies (EMSs). The upper-level is applying the particular deep deterministic policy gradient (DDPG) algorithms for EMS training at different speed intervals. The lower-level is employing the TL method to transform the pre-trained neural networks for a novel driving cycle. Finally, a series of experiments are executed to prove the effectiveness of the presented control framework. The optimality and adaptability of the formulated EMS are illuminated. The founded DRL and TL-enabled control policy is capable of enhancing energy efficiency and improving system performance.

show abstract

Section: A a Bi-level Frameworkmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Transfer Deep Reinforcement Learning-Enabled Energy Management Strategy for Hybrid Tracked Vehicle

Guo

Tang

et al. 2020

IEEE Access

View full text Add to dashboard Cite

show abstract

“…By equation 5, (6), (12), (13) and neglect all the inertia, the speed and torque relation among power sources and the torque output end is derived as follows…”

Section: Powertrain System Modelingmentioning

confidence: 99%

“…11 The optimization-based control strategies mainly include dynamic programming (DP), equivalent fuel consumption minimization strategy (ECMS), model predictive control (MPC), and deep reinforcement learning algorithm that have emerged with artificial intelligence in recent years. [12][13][14][15] Compared with the rule-based control strategies, the optimization-based control strategies don't need to divide the operation modes for the vehicle, the optimal or sub-optimal solution of the control can be obtained. Therefore, the fuel-saving effect is more obvious.…”

Section: Introductionmentioning

confidence: 99%

Modeling and energy management strategy research of a power-split hybrid electric vehicle

Zou

Huang

et al. 2020

Advances in Mechanical Engineering

View full text Add to dashboard Cite

A power-split hybrid electric vehicle with a dual-planetary gearset is researched in this paper. Based on the lever analogy method of planetary gearsets, the power-split device is theoretically modeled, and the driveline simulation model is built by using vehicle modeling and simulation toolboxes in MATLAB. Six operation modes of the vehicle are discussed in detail, and the kinematic constraint behavior of power sources are analyzed. To verify the rationality of the modeling, a rule-based control strategy (RB) and an adaptive equivalent consumption minimization strategy (A-ECMS) are designed based on the finite state machine and MATLAB language respectively. In order to demonstrate the superiority of A-ECMS in fuel-saving and to explore the impact of different energy management strategies on emission, fuel economy and emission performance of the vehicle are simulated and analyzed under UDDS driving cycle. The simulation results of the two strategies are compared in the end, shows that the modeling is rational, and compared with RB strategy, A-ECMS ensures charge sustaining better, enables power sources to work in more efficient areas, and improves fuel economy by 8.65%, but significantly increases NOx emissions, which will be the focus of the next research work.

show abstract

“…It does not rely on any prediction or predefined rules. Different learning based energy management strategies that are existing are discussed in this paper [21]- [26].…”

Section: Learning-based Emsmentioning

confidence: 99%

Examination on Classification of EVs and Energy Management Strategies of HEV

2019

IJITEE

View full text Add to dashboard Cite

In recent days, the demand for petroleum and emission of pollutant gases continuously increase. This necessitates the electrification power train which replaces Internal Combustion Engine (ICE). Despite pure electric vehicles or Battery Electric Vehicle (EV) reduce the greenhouse gas emissions, there are some major hurdles for EVs to overcome before they totally relieve ICE vehicles form transport sector such as range anxiety, battery storage, economic fall down due to automobile industries, etc. This necessitates Hybrid Electric vehicle (HEV) which combines two different power sources to propel the vehicle. One of the challenges in HEV is how to control the power coming from the two different sources such as battery and ICE. The prime goal of an Energy Management Strategy (EMS) is to manage energy flow such that fuel consumption and emissions are minimized without affecting the vehicle’s performance. In this paper, the different structures of power train and energy management strategies are analysed.

show abstract

Deep reinforcement learning of energy management with continuous control strategy and traffic information for a series-parallel plug-in hybrid electric bus

Cited by 241 publications

References 35 publications

Transfer Deep Reinforcement Learning-Enabled Energy Management Strategy for Hybrid Tracked Vehicle

Transfer Deep Reinforcement Learning-Enabled Energy Management Strategy for Hybrid Tracked Vehicle

Modeling and energy management strategy research of a power-split hybrid electric vehicle

Examination on Classification of EVs and Energy Management Strategies of HEV

Contact Info

Product

Resources

About