Moving Vehicle Tracking Under Measurement Uncertainties, Multi-Step Random Delays, and Packet Dropouts

The unbiased finite impulse response (UFIR) filter, Kalman filter (KF), and game theory $H_{\infty} $ filter are developed for vehicle tracking in wireless sensor networks (WSNs) with multi-step random delays and multiple dropouts. The problem with data delays is solved for given delay probabilities by converting a model with delays to another without delays. The dropouts are detected at the receiver by a data sensor, and lost data are compensated for by prediction. The filters developed are experimentally tested for accuracy and robustness using the Global Positioning System-based vehicle tracking database, where the measured coordinates are transmitted over WSN with delays. It is shown that the UFIR filter is the most robust to uncertain delays and the KF is the most accurate when the delay probabilities are known exactly. The game theory $H_{\infty }$ filter is the most accurate under ideal conditions and prone to divergency otherwise.