Abstract：ALVINN(Autonomous Land Vehicle in a Neural Network) is the best intelligent vehicle road follower based on neural network in the world. Road following camera is fixed on ALVINN,so information about road may be missed during steering and following performance may be decreased,even the task of road following cannot be completed. In this paper,an improved based on road follower always is proposed,in which the azimuth of road following camera is efficiently controlled so that the road is always in the center of image. This opens possibilities for improving road following performance. Performance improvement is verified by computer simulation results.
 Pomerleau D A.Neural Network Perception for Mobile Robot Guidance.Kluwer Academic Publishing,Boston,1993;  Zhu Z,Yang S,Xu G,Lin X,Shi D.Fast Road Classification and Orientation Estimation Using Omni-view Images and Neural Networks.IEEE Trans.IP,1998,7(8):1182-1197;  Sunkthankar R,Pomerleau D A,Thorpe C E.Panacea:An Active Sensor Controller for the ALVINN Autonomous Driving System.Report CMU-RI-TR-93-09,Robotics Institute,Carnegie Mellon University,1993