In the status quo, traffic control systems operate on predetermined patterns and instructions devised from past data. While this method functions effectively for traffic under normal conditions, it becomes heavily congested and inefficient during rush hour. Furthermore, the constant presence of unexpected emergencies renders pre-determined systems ineffectual. By combining traditional traffic controllers with modern technologies like Internet of Things devices and computer vision, traffic control systems can be greatly improved. Yet there are currently no systems that can affordably fulfill this task. By optimizing traffic signal duration, this allows for both a reduction in delay time for vehicles and a reduction of greenhouse gases emitted. This is especially pertinent considering the recent Intergovernmental Panel on Climate Change report regarding the dramatically worsening state of the climate crisis and the recent Harvard T.H. Chan School of Public Health report regarding the economic and human costs of vehicle emissions. Evidently, there is a compelling need for such a traffic control system to optimize throughput and thus greatly reduce vehicle greenhouse gas emissions. This research presents a novel deep learning traffic control system, called JARVITS (Just A Rather Very Intelligent Traffic System) that can be used for accurate real-time vehicle detection and signal control.
JARVITS: A Novel Deep Learning IoT Traffic Control System for Real-time Detection and Signal Optimization
2022-09-30
1790988 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
- ON-DEVICE REAL-TIME TRAFFIC SIGNAL CONTROL SYSTEM BASED ON DEEP LEARNING
Europäisches Patentamt | 2023
- ON-DEVICE REAL-TIME TRAFFIC SIGNAL CONTROL SYSTEM BASED ON DEEP LEARNING
Europäisches Patentamt | 2022
|