Travel time estimation in urban road networks

The development of two methods for the estimation of link travel times at signalised junctions under traffic responsive control are described. In the first method, neural networks were trained with data from a microscopic traffic simulator of an isolated junction operating under traffic responsive control. In the second method, a simple queueing model was evaluated with data from a registration plate survey in a SCOOT region in Nottingham. The first method, unlike the second, did not require knowledge of the signal timings or parameters, such as saturation flow, but did require lots of training data.