Traffic Incident Generation And Supervised Learning Based Detection Via A Microscopic Simulation Platform
- Resource Type
- Conference
- Authors
- Yang, Huan; Zhao, Han; Liu, Guoqiang; Wang, Yu; Wang, Danwei
- Source
- 2023 IEEE International Conference on Cybernetics and Intelligent Systems (CIS) and IEEE Conference on Robotics, Automation and Mechatronics (RAM) Cybernetics and Intelligent Systems (CIS) and IEEE Conference on Robotics, Automation and Mechatronics (RAM), 2023 IEEE International Conference on. :146-151 Jun, 2023
- Subject
- Robotics and Control Systems
Location awareness
Mechatronics
Microscopy
Supervised learning
Random access memory
Convolutional neural networks
Robots
- Language
- ISSN
- 2326-8239
An effective automatic incident detection (AID) method is vital in the intelligent transportation system. This paper presents an online Convolutional Neural Network (CNN) based traffic incident detection method to detect the incidents in urban networks by three types of traffic flow data (traffic volume, speed and acceleration of vehicles). Regarding the traffic flow data during a time period as an image with RGB channels, this method employs the idea of CNN for dealing with image classification. The online incident detection was realized by using the time sliding window technique. For testing the performance, this paper develops an incident generation simulation model via a microscopic simulation platform VISSIM. The effectiveness and efficiency are demonstrated by several experiments through the simulation and based on the traffic demand estimated by true traffic flow.