Performance prediction of tracking sensors for surface vehicle collision avoidance

Multiple complementary sensors can be valuable when designing a robust target tracking system for collision avoidance on a moving platform. In this paper, the achievable performance of several sensor configurations are assessed using the Cramer-Rao Lower Bound (CRLB) and an alternative performance measure closely related to the Extended Kalman Filter (EKF). Targets are modeled using constant velocity (CV) and constant turn rate (CT) models. Simulations indicate that RADAR performs well overall, but that stereo bearing sensors can improve accuracy at relatively short ranges and that AIS velocity measurements can improve the ability to estimate target maneuvers.