Abstract
To accommodate the heavy travel demand in high-density areas, Taipei Bus Station (TBS) is developed as the first multi-level bus terminal in Taipei City. TBS also plays important roles in congestion mitigation, energy conservation and pollutant reduction. Unlike conventional single-level terminals, bus flow interruption while circulating in TBS could significantly impact the service quality and deteriorate environmental condition. Considering time-varying demand and existing Radio Frequency Identification (RFID) monitoring systems, this study constructed an adaptive signal control model combining an artificial neural network (ANN) demand forecasting model to manage bus traffic in TBS. In the case study, the self-retraining demand forecasting algorism is programmed in existing controller/computer to facilitate demand changes. The proposed model has demonstrated itself very efficient in reducing congestion within the terminal.
