Adaptive Traffic Light Control Based on Fuzzy Logic with Consideration of Public Transport Priority

Abstract

Adaptive traffic light control taking into account the priority of public transport requires a combination of technological flexibility and social sensitivity. The article analyzes the possibilities of using fuzzy logic as a basis for co9nstructing regulators capable of responding to traffic changes without rigid boundaries and fixed rules. Control using fuzzy rules enables to take into account transport delays, flow density, passenger capacity, and also minimize delays without disturbing the overall balance. Special attention is paid to the multi-agent approach, in which each intersection acts as an independent control unit capable of making decisions, based on local data and interaction with neighboring nodes. Such an architecture allows creating the adaptive control network where the priority for public transport is provided not episodically, but systematically — along the entire route. The article considers hybrid models where fuzzy logic is combined with elements of reinforcement learning and infrastructure communication (V2I). The effectiveness of the approach in reducing the delay time for transport by 12…18 % compared to non-adaptive methods is demonstrated. The formula for the priority index is described, enabling to integrate the factors of delay and vehicle weight. The comparison is made with other models, in particular of type 2 and learning based on GPS data. The conclusion is made about the suitability of the proposed approach for the implementation both at individual intersections and city-wide. In future, simulation testing, scaling to a multi-agent architecture and integration with barrier-free requirements and neural networks are considered. The proposed model not only improves transport efficiency, but also contributes to the formation of a socially- oriented infrastructure sensitive, to the needs of passengers.