Abstract: Vehicle state estimation problem is divided into four aspects: sensor configuration, physical model, estimation algorithm and parameter adaption. The key technologies are later in details discussed. Through comparison of typical sensor configuration for vehicle longitudinal speed, yaw velocity and side-slip angle estimation, an ideal sensor configuration is given. Furthermore, the advantage and disadvantage of kinematic model and dynamic model are analyzed, then the viewpoint is put forward that kinematic model is more suitable for longitudinal speed and yaw velocity estimation, in contrast dynamic model is better for vehicle side-slip angle estimation .The typical estimation algorithms are listed and important issues to be regarded in practical utilization are discussed. It is pointed out that the realization of parameter adaption is an effective mothod to enhance the estimation precision under different driving situation. Recursive least square, combined kalman filter and other parameter observation methods are given.

Key words: Longitudinal speed estimation, Observer, Parameter adaptation, Side-slip angle estimation, Vehicle state estimation, Yaw velocity estimation