Abstract: In view of the insufficient accuracy and reliability of lane-change warning model based on single sensor, and taking no account the real scene factors in determining of ego vehicle acceleration in lane-change process, a lane change warning model based on multi-sensor fusion is proposed, in which the ego vehicle speed, the relative distance and speed between ego vehicle and other vehicles are taken into consideration in determining ego vehicle acceleration. The original lane-change warning model is optimized, and the lane-change decision rules are formulated with time to collision and minimum safety spacing as evaluation indicators. The results of simulation and real vehicle test show that the lane-change warning model based on multi-sensor fusion is more sensitive and efficient, meeting the requirements of road utilization and driver operation, and the effectiveness of lane change warning model is verified, with the advantages of multi-sensors in environmental perception demonstrated

Key words: multi-sensor fusion; safety distance; lane change warning model; lane change decision-making