A novel iterative evolving fuzzy logic controller called ant-genetic fuzzy logic controller (AGFLC) was developed for transit signal preemption (TSP) wherein variation of traffic conditions and ambiguity of expert judgement are accounted for. The core logics of this iterative evolving AGFLC algorithm include learning the combination of rules by ant colony optimization and tuning the shapes of membership functions by genetic algorithm. An AGFLC-based TSP model is proposed that provides conditional signal priority to the actuated transit vehicles to minimize the total person delays of the intersections studied. To realize the control performance, both exemplified and field cases were tested at an isolated intersection and consecutive intersections along an arterial. Compared with other models, includinggenetic fuzzy logic controller (GFLC)-based TSP model, net-benefit conditional TSP model, unconditional TSP model, and pre-timed signal without TSP,the results show that the proposed AGFLC-based TSP model has outperformedunder different circumstances. For the covering abstract see ITRD E144727. Reprinted with permission of Elsevier.
Samenvatting