Forecasting, simulation and PIP control of interurban traffic networks

This paper discusses a number of research areas concerned with dynamic traffic management, including forecasting, simulation and, ultimately, control of motorway traffic flow. A detailed non-linear simulation of the Amsterdam A10 west ring road, utilising a dynamic state space representation called the Statistical Traffic Model (STM) is described. The simulation is employed in the development of a robust, co-ordinated ramp metering system, based on the Non-Minimal State Space (NMSS) approach to control design, which yields adaptive Proportional-Integral-Plus (PIP) controllers. Finally, when employed in predictive mode, the STM requires forecasts of the on-ramp inputs to the motorway. One approach to obtaining such univariate forecasts, namely Dynamic Harmonic Regression (DHR) is also discussed.