Battery-Involved Energy Management for Hybrid Electric Bus Based on Expert-Assistance Deep Deterministic Policy Gradient Algorithm

Wu, Jingda; Wei, Zhongbao; Liu, Kailong; Quan, Zhongyi; Li, Yunwei

doi:10.1109/tvt.2020.3025627

Cited by 158 publications

(75 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, the energy management for a multi-agent system governed microgrid in the Internet of Energy was not investigated. In Reference [35], the authors proposed a new energy management-based machine learning method for an electric hybrid bus with an emphasis on thermal security and the degradation of the lithium-ion batteries structure. In Reference [36], researchers proposed a novel energy management for a grid connected mode and an island mode using an adaptive neuro fuzzy interface system.…”

Section: Literature Review Of Theoretical Backgroundmentioning

confidence: 99%

A Novel Real-Time Electricity Scheduling for Home Energy Management System Using the Internet of Energy

et al. 2021

View full text Add to dashboard Cite

This paper presents a novel scheduling scheme for the real-time home energy management systems based on Internet of Energy (IoE). The scheme is a multi-agent method that considers two chief purposes including user satisfaction and energy consumption cost. The scheme is designed under environment of microgrid. The user impact in terms of energy cost savings is generally significant in terms of system efficiency. That is why domestic users are involved in the management of domestic appliances. The optimization algorithms are based on an improved version of the rainfall algorithm and the salp swarm algorithm. In this paper, the Time of Use (ToU) model is proposed to define the rates for shoulder-peak and on-peak hours. A two-level communication system connects the microgrid system, implemented in MATLAB, to the cloud server. The local communication level utilizes IP/TCP and MQTT and is used as a protocol for the global communication level. The scheduling controller proposed in this study succeeded the energy saving of 25.3% by using the salp swarm algorithm and saving of 31.335% by using the rainfall algorithm.

show abstract

Section: Literature Review Of Theoretical Backgroundmentioning

confidence: 99%

A Novel Real-Time Electricity Scheduling for Home Energy Management System Using the Internet of Energy

et al. 2021

View full text Add to dashboard Cite

show abstract

“…[ 22 ] addresses the particular case of an autonomous underwater vehicle, where it is compared a DQN architecture with a Deep Deterministic Policy Gradient (DDPG) approach for continuous action space. Another interesting application of such continuous optimization algorithms can be found in [ 23 , 24 ] where it is used a Soft Actor-Critic architecture to improve the thermal and energetic efficiency of a vehicle under power constraints. For some applications like the one presented in this paper, there is no need of a continuous domain and a discretized space is enough, but it is remarkable how the DRL can adapt to these problems as well.…”

Section: Related Workmentioning

confidence: 99%

A Dimensional Comparison between Evolutionary Algorithm and Deep Reinforcement Learning Methodologies for Autonomous Surface Vehicles with Water Quality Sensors

Luis

Reina

Toral

2021

Sensors

View full text Add to dashboard Cite

The monitoring of water resources using Autonomous Surface Vehicles with water-quality sensors has been a recent approach due to the advances in unmanned transportation technology. The Ypacaraí Lake, the biggest water resource in Paraguay, suffers from a major contamination problem because of cyanobacteria blooms. In order to supervise the blooms using these on-board sensor modules, a Non-Homogeneous Patrolling Problem (a NP-hard problem) must be solved in a feasible amount of time. A dimensionality study is addressed to compare the most common methodologies, Evolutionary Algorithm and Deep Reinforcement Learning, in different map scales and fleet sizes with changes in the environmental conditions. The results determined that Deep Q-Learning overcomes the evolutionary method in terms of sample-efficiency by 50–70% in higher resolutions. Furthermore, it reacts better than the Evolutionary Algorithm in high space-state actions. In contrast, the evolutionary approach shows a better efficiency in lower resolutions and needs fewer parameters to synthesize robust solutions. This study reveals that Deep Q-learning approaches exceed in efficiency for the Non-Homogeneous Patrolling Problem but with many hyper-parameters involved in the stability and convergence.

show abstract

“…The authors of [54] propose an energy management strategy based on machine learning for HEB. The Deep Deterministic Policy Gradient (DDPG) algorithm was used to increase the cold start efficiency and optimize HEB power allocation.…”

Section: Hydrogen As a Fuel Of The Futurementioning

confidence: 99%

Fuel Cells Based on Natural Polysaccharides for Rail Vehicle Application

et al. 2021

View full text Add to dashboard Cite

This manuscript shows the use of natural polysaccharides such as starch and cellulose as a carbon source for fuel cells. To achieve this, two innovative methods of obtaining hydrogen have been shown: by adsorption and by enzyme. The carbonization path of the material results in excellent sorption properties and allows gas with high efficiency to be obtained. The enzymatic method for the degradation of the compound is more expensive because specific enzymes (such as laccase, tyrosinase) must be used, but it allows greater control of the properties of the obtained material. A scientific novelty is the use of natural raw materials, the use of which increases the biodegradability of the electrochemical system and also reduces the cost of raw materials and increases the range of their acquisition. Energy should be generated where it is used. Another goal is decentralization, and thanks to the proposed solutions, hydrogen cells represent an innovative alternative to today’s energy giants—also for independent power supply to households. The proposed harvesting paths are intended to drive rail vehicles in order to reduce emissions and secondary pollution of the environment. The goals of both methods were easy recycling, high efficiency, increased environmental friendliness, low cost and a short hydrogen production path.

show abstract

Battery-Involved Energy Management for Hybrid Electric Bus Based on Expert-Assistance Deep Deterministic Policy Gradient Algorithm

Cited by 158 publications

References 43 publications

A Novel Real-Time Electricity Scheduling for Home Energy Management System Using the Internet of Energy

A Novel Real-Time Electricity Scheduling for Home Energy Management System Using the Internet of Energy

A Dimensional Comparison between Evolutionary Algorithm and Deep Reinforcement Learning Methodologies for Autonomous Surface Vehicles with Water Quality Sensors

Fuel Cells Based on Natural Polysaccharides for Rail Vehicle Application

Contact Info

Product

Resources

About