Developing a testbed for Intelligent Transportation Systems
Abstract
It is estimated that the average UK commuter spent on average 32 hours stuck in traffic jams in 2017 and that our car park is growing from 1.2 billion cars to 2 billion cars within 2035. In recent years there has been growing research interest in the area of traffic optimization to tackle current and future problems concerning congestion. Intelligent Transportation Systems (ITS) aims to provide innovative services to different modes of transport and traffic management. With the partial introduction of self-driving cars and technology advancements within communication technologies, we can create smarter solutions for traffic optimization and increase traffic flow.
While the field of ITS applications is extensively tested, regarding technical and business perspectives, there are few testbed solutions available to test such applications at a general level. The work performed in this thesis focuses on creating a low-cost testbed by using DiddyBorg robots equipped with sensors to emulate a real-world scenario. The proposed testbed is a hybrid between computer simulations and field tests, enabling researchers and developers to test new ITS applications faster with real-world capabilities. The system is a highly modular framework where new applications can be implemented and tested quickly. To evaluate the testbed, two different applications are developed: regular traffic light and virtual traffic light. Virtual traffic light aims to remove the physical traffic lights by enabling cars to communicate and resolve possible conflicts at intersections. Results show that regular traffic lights, in low-density scenarios, reduce the average waiting time in an intersection by around 90%. Observations and results collected during experiments show that the proposed testbed can give valuable insights to researchers and developers within the field of ITS.