Real time traffic flow forecasting and predictive ramp-metering using spectral analysis

An algorithm for the implementation of short term prediction of traffic with real time updating based on spectral analysis is described. The prediction is based on the characterisation of the flow based on modal functions associated with a covariance matrix constructed from historical flow data. This paper describes the intended online implementation of the method which predicts within-day traffic flow using a forecasting horizon of 1 hour 15 minutes with 15 minute step. Thus, every 15 minutes, the traffic flow for a further one 1 hour 15 minutes is predicted. As well as forecasting to this horizon, a weighted averaging technique is developed which allows the prediction of one 15 minute step ahead by using current and previous predictions of traffic flows at the given time instant while placing more weight on the more recent predictions. The development of a MATLAB program for the implementation is described and results are presented for a number of different schemes. The procedure is then applied in a ramp metering scenario in which the control inputs are based on flow predictions made 15 min previously.