Experimental study on vehicle speed estimation using accelerometer and wheel speed measurements

Accurate estimation of longitudinal vehicle speed using general in-vehicle sensors like wheel speed sensors and accelerometers is crucial for good performance of vehicle dynamic control systems. Most of passenger vehicles are two-wheel drive, which provides speed measurement by wheel speed sensors of non-driven wheels. However, when wheels are under braking, this method is not reliable enough since wheels are at high level of slip. An alternative method is to use the direct integration of the longitudinal acceleration, but offset in the acceleration signal may cause excessive errors in estimation results. This work develops a speed estimation method using Kalman filtering algorithm by establishing an updating method of acceleration offset.