Extended Kalman Filter with Sliding Window For Vehicle Tracking in Urban Area

For intelligent transportation system (ITS) in 5G, the range between road-side unit (RSU) and vehicle user equipment (VUE), as well as the degree of arrival (DOA), are collected by the RSU as primary measurements. These measurements, including time and azimuth, usually suffer the influences of clock offset and noise in real-world situations where the statistical properties of the signal change over time. This paper proposes the Sliding Window Extended Kalman filter (SWEKF) to enhance the accuracy of vehicle position estimation and tracking. The slide windowing technique breaks measured signals into smaller segments. A window of a fixed size is moved over the signal, and calculations are performed on each segment, making it useful for dealing with non-stationary signals. The algorithm of SWEKF is developed in details, and simulations are performed. Simulation results show that our scheme outperforms the conventional EKF in both position estimation and tracking sensibility.