Driving an emergency vehicle through a red light poses a safety risk and increases travel time, especially during peak traffic hours. Existing solutions for traffic signal preemption have limitations in range or require prior knowledge of the vehicle's destination.
In this thesis, I propose a centralized GPS-based detection system that does not require knowledge of the destination. The developed prototype consists of an in-vehicle module and a central server, which communicate over a mobile network. The module periodically transmits the vehicle's location. The server uses an algorithm to determine the direction of approach and the appropriate traffic light phase. The system is tested across seven traffic scenarios; it responds correctly in four of them. In the cases of roadside parking, a U-turn, and a turn before the intersection, the algorithm does not perform correctly because it remains in emergency mode longer than necessary.
|
