Design of the deep learning based electric vehicle charging forecasting network and navigation system
In the modern city, the number of Electric Vehicle (EV) is increasing rapidly for its low emission and better dynamic performance, leading to an increasing demand of EV charging. However, due to the limited number of EV charging facilities, catering the huge demand of the time consuming EV charging...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis-Master by Coursework |
| Language: | English |
| Published: |
Nanyang Technological University
2022
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/163230 |
| _version_ | 1826125853819404288 |
|---|---|
| author | Song, Yaofeng |
| author2 | Su Rong |
| author_facet | Su Rong Song, Yaofeng |
| author_sort | Song, Yaofeng |
| collection | NTU |
| description | In the modern city, the number of Electric Vehicle (EV) is increasing rapidly for its low emission and better dynamic performance, leading to an increasing demand of EV charging. However, due to the limited number of EV charging facilities, catering the huge demand of the time consuming EV charging becomes an unignorable problem. In this paper, we aim to improve the efficiency of the EV charging station usage and save time for EV users by designing a station availability forecasting network and an EV navigation system. On one hand, there are multiple works focusing on time series forecasting using Deep Learning (DL) methods for traffic speed, traffic flow or public transportation demand. However, EV charging availability forecasting is barely mentioned and a big number of DL model designs in the traffic area ignore the importance of spatial information and external factors such as weather and Places of Interest (POI) information. Our design, Attribute-Augmented Spatiotemporal Graph Informer Network (AST-GIN), fully intakes the spatial information and external factors, and outperforms many other state-of-art time series forecasting methods. On the other hand, we build an EV navigation system on the basis of the
traffic simulator SUMO for Deep Reinforcement Learning (DRL) experiments. |
| first_indexed | 2024-10-01T06:43:10Z |
| format | Thesis-Master by Coursework |
| id | ntu-10356/163230 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T06:43:10Z |
| publishDate | 2022 |
| publisher | Nanyang Technological University |
| record_format | dspace |
| spelling | ntu-10356/1632302022-11-30T02:07:22Z Design of the deep learning based electric vehicle charging forecasting network and navigation system Song, Yaofeng Su Rong School of Electrical and Electronic Engineering RSu@ntu.edu.sg Engineering::Electrical and electronic engineering In the modern city, the number of Electric Vehicle (EV) is increasing rapidly for its low emission and better dynamic performance, leading to an increasing demand of EV charging. However, due to the limited number of EV charging facilities, catering the huge demand of the time consuming EV charging becomes an unignorable problem. In this paper, we aim to improve the efficiency of the EV charging station usage and save time for EV users by designing a station availability forecasting network and an EV navigation system. On one hand, there are multiple works focusing on time series forecasting using Deep Learning (DL) methods for traffic speed, traffic flow or public transportation demand. However, EV charging availability forecasting is barely mentioned and a big number of DL model designs in the traffic area ignore the importance of spatial information and external factors such as weather and Places of Interest (POI) information. Our design, Attribute-Augmented Spatiotemporal Graph Informer Network (AST-GIN), fully intakes the spatial information and external factors, and outperforms many other state-of-art time series forecasting methods. On the other hand, we build an EV navigation system on the basis of the traffic simulator SUMO for Deep Reinforcement Learning (DRL) experiments. Master of Science (Computer Control and Automation) 2022-11-30T02:07:22Z 2022-11-30T02:07:22Z 2022 Thesis-Master by Coursework Song, Y. (2022). Design of the deep learning based electric vehicle charging forecasting network and navigation system. Master's thesis, Nanyang Technological University, Singapore. https://hdl.handle.net/10356/163230 https://hdl.handle.net/10356/163230 en ISM-DISS-02885 application/pdf Nanyang Technological University |
| spellingShingle | Engineering::Electrical and electronic engineering Song, Yaofeng Design of the deep learning based electric vehicle charging forecasting network and navigation system |
| title | Design of the deep learning based electric vehicle charging forecasting network and navigation system |
| title_full | Design of the deep learning based electric vehicle charging forecasting network and navigation system |
| title_fullStr | Design of the deep learning based electric vehicle charging forecasting network and navigation system |
| title_full_unstemmed | Design of the deep learning based electric vehicle charging forecasting network and navigation system |
| title_short | Design of the deep learning based electric vehicle charging forecasting network and navigation system |
| title_sort | design of the deep learning based electric vehicle charging forecasting network and navigation system |
| topic | Engineering::Electrical and electronic engineering |
| url | https://hdl.handle.net/10356/163230 |
| work_keys_str_mv | AT songyaofeng designofthedeeplearningbasedelectricvehiclechargingforecastingnetworkandnavigationsystem |