Implementation and Results of a Multi-Stage Model Predictive Traffic Light Control System

This paper deals with the implementation and results of the application of a multi-stage traffic light control system which includes a simulation-based traffic estimation and model predictive control. The traffic light control system incorporates a fuzzy system for traffic light phase preselection, followed by a model predictive control to optimise phase combinations and switching times. Predefined phases are selected without restrictions in the order according to a multi-objective optimisation to adapt to the traffic as freely as possible. Initially, the system is tested in simulations and compared with existing methods and analysed afterwards for its effectiveness in a prototype commissioning in field tests. Results indicate high potentials for reducing emissions and waiting times, highlighting the system's value. However, further refinement is necessary for standard implementation. This comprehensive approach demonstrates advancements in traffic management technology, showcasing the potential for enhancing urban mobility and reducing environmental impact.