Methods for integrating control algorithms of light signal systems for various simulation environments with special consideration of public transport

Kurzfassung

Simulation is a crucial part of today’s research and development due to its high efficiency. It is mostly effiective in environments where otherwise high costs or even lives were to be put on stake for testing new approaches. This is both the matter of fact when facing the domain of traffic control. If a traffic light system is malfunctioning the result can vary from a traffic jam, which would create an economic loss on a regional level. This might increase over time if this malfunction is persistent and even lead to accidents that cost lives. For example when all lights at an intersection show green the same time. To avoid all that it is very important to test everything that goes live on real hardware beforehand and ensure its reliability and fail-safety. Furthermore, the more complex the mechanics to control traffic lights and thus traffic flow in general become too. The existing hardware, especially the controllers, should be integrated into this testing pipeline to estimate the effects of hardware lifespan on the deployed algorithms and to proof that it can actually run the algorithms. This thesis provides a prototype and a design for implementing a scalable and flexible architecture for traffic light signal systems in the DLR’s simulation framework DOS, a possibility to import traffic light programs created with the simulation framework SUMO and examining how hardware testing could be a future feature of DLR’s test environment. Furthermore it’s discussing how the presented architecture is not only living up to the promise of scalability and flexibility, but also how control schemes with public transportation consideration can be straight forward deployed on it. Lastly, it presents a preparation for a future feature of the simulation environment, the live-stream of traffc light data from the city of Braunschweig to its virtual counterpart.