Título: Intelligent Urban Traffic Mobility based on Visible Light Communication, Projecto IPL/IDI&CA2024_INUTRAM_ISEL
Orador: Paula Louro (DEETC/ISEL, CTS/Instituto de Desenvolvimento de Novas Tecnologias, LASI/Universidade Nova de Lisboa)
Abstract:
This seminar presents an innovative solution for urban traffic management by integrating Visible Light Communication (VLC) technology with learning-based traffic signal control. The approach aims to enhance safety and reduce congestion through efficient and adaptive traffic control strategies suitable for various urban environments.
The system combines VLC localization with reinforcement learning (RL) to optimize traffic flow and safety. Vehicular-VLC (V-VLC) communication gathers real-time traffic information, which is used to adapt traffic signals dynamically. Additionally, a novel pedestrian mobility model tailored for outdoor scenarios incorporates diverse pedestrian behaviours into traffic signal control. This model, validated in a case study of downtown Lisbon, integrates pedestrian preferences to optimize routing algorithms.
A deep RL-based adaptive traffic control model ensures equitable green light allocation for both vehicles and pedestrians, minimizing total user delays and addressing pedestrian interruptions like jaywalking. Real-time optimization adjusts signal control based on pedestrian and vehicle factors, improving overall traffic flow and reducing waiting times. Simulation experiments using real intersection data demonstrate the system's effectiveness, highlighting significant improvements in traffic management for both vehicles and pedestrians.
