Vehicle lateral and longitudinal velocity estimation based on Unscented Kalman Filter

Vehicle lateral velocity (vy) and longitudinal velocity (vx) play an important role in modern vehicle active safety control systems. However using sensors to measure vy and vx are very expensive, it is necessary to estimate vy and vx from other variables measured easily. In this paper, a novel estimation method for vy and vx is proposed. This method is based on Unscented Kalman Filter (UKF) from steering angle, yaw rate and acceleration, which can be measured easily and cheaply by sensors. This method was evaluated under a variety of maneuvers and road conditions. The simulation results demonstrated that the proposed method was robust and even if condition changed, this method could still estimate vy and vx accurately.