Motorway incident detection using principal component analysis

Traffic incidents on motorways have a significant impact on traffic operations and consequently cause severe traffic congestion, mobility loss, environmental pollution and energy waste. Conventional incident detection algorithms focus on identification and characterisation of a wide range of different incidents from historic data. In this paper, we introduce a new approach which estimates the probability distribution of incident-free data by means of principal component analysis (PCA). A novel input vector is discriminated by measuring its distance to the centroid of normal data. Both simulated and field data were used in the work to evaluate the reliability and transferability of the methodology. The results show that the technique is promising for incident detection in dynamic traffic monitoring systems