SUMO BASED HARDWARE/SOFTWARE CO-SIMULATION FOR TWO-INTERSECTION ADAPTIVE AND COLLABORATIVE TRAFFIC SIGNAL CONTROLLER

• Main Author: Emkel Ginting, Kendrik
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/73864
• Institution: Institut Teknologi Bandung

Congestion is a problem faced by all big cities in the world. One of the causes of congestion is the ineffective regulation of traffic lights at intersections, especially during rush hour. Traffic management at existing intersections is currently carried out using a fixed-time Traffic Signaling Controller. This causes the Traffic Signaling Controller not to be adaptive to fluctuations in traffic flow at intersections. In addition, the existing coordination between controllers is only limited to time offsets. In this final project, an adaptive and collaborative Traffic Signaling Controller for two adjacent intersections is developed using Reinforcement Learning. The development of this adaptive and collaborative Traffic Signaling Controller needs to be carried out by taking into account various aspects, one of the main ones is safety. By paying attention to safety aspects, the development of adaptive and collaborative Traffic Signaling Controller cannot be implemented directly in the field. Therefore, a system that can simulate actual environmental conditions as a training platform for the controllers that is being developed and to ensure that the controllers work properly before being implemented directly in the field is needed. This final project book specifically discusses the design, implementation, and testing of the traffic environment simulator. This simulator is used as an environment that trains and tests Reinforcement Learning algorithms previously implemented in the field. The simulation is carried out using Simulator for Urban Mobility (SUMO) and taking into account the regulations that apply in Indonesia. In addition, interconnection between subsystems is also designed, especially with miniatures that is controlled by Arduino Mega. In general, the design, implementation and testing of this traffic environment simulator went well. The simulator and miniature subsystems are connected to each other. However, it is necessary to improve the connectivity between systems, namely reducing the delay between the simulator and the miniature.