Estimation on Location, Velocity, and Acceleration With High Precision for Collision Avoidance

Abstract

An approach is proposed to estimate the location, velocity, and acceleration of a target vehicle to avoid a possible collision. Radial distance, velocity, and acceleration are extracted from the hybrid linear frequency modulation (LFM)/frequency-shift keying (FSK) echoed signals and then processed using the Kalman filter and the trilateration process. This approach proves to converge fast with good accuracy. Two other approaches, i.e., an extended Kalman filter (EKF) and a two-stage Kalman filter (TSKF), are used as benchmarks for comparison. Several scenarios of vehicle movement are also presented to demonstrate the effectiveness of this approach.