Freeway work zone traffic state estimation with fault diagnosis

Abstract

Freeway work zone can cause disruption to local traffic and have adverse impacts on mobility and safety of road users and those who work in the work zone. To ensure an effective traffic management strategy, it is essential to accurately and instantaneously estimate the traffic states at the work zone area. While many traffic state estimation methods are proposed by previous studies, few of them consider the occurrence of freeway sensor faults, which may result in a large deviation in state estimation and potentially lead to an inappropriate traffic management strategy. To overcome the impacts of sensor faults and provide accurate traffic state estimation, this study presents an approach using sensor fault diagnosis for traffic state estimation at freeway work zone area. Considering the capacity drop, the switching mode model with Kalman filter was used to estimate the traffic states. With the analysis of the density residuals generated by traffic sensors and probes, the fault diagnosis can detect the type of sensor faults and reconfigure the estimation model. The proposed system is implemented and evaluated in traffic simulator SUMO under a realistic freeway work zone environment. The results show that the developed system can accurately identify the type of fault in short time. An accurate traffic state estimation is provided and fairly maintained under fault-free and sensor-fault scenarios respectively.