The classification of the wheeled vehicles and tracked vehicles is one of the difficulties of vehicles recognition. Rotation of the wheels and running of the track are the typical forms of micro-motion

the micro-Doppler signatures of the micro-motion can be the crucial proof for classification of the two kinds of vehicles. Firstly

the micro-Doppler difference of different micro-motion of the two vehicles are analyzed based on the echoes under radar with short dwell time;Secondly

based on the short-time stability and correlation of the echoes

the Auto Regression (AR) model of echoes is established

the forward and backward prediction method is applied in the extraction of AR model coefficients

which are used in the classification of the two kinds of vehicles

and the selection of the order of AR model is operated

the separation abilities between forward and backward prediction coefficients and forward prediction coefficients are compared;At last

after the pre-processing of the measured data

the classification of the two vehicles is accomplished with high accuracy based on the extraction of AR model coefficients. The effectiveness and robustness of the method are proved by the simulation results.