Research On Target Recognition Of Adaptive Cruise System

The application of the adaptive cruise control system(ACC)on the vehicle effectively reduces the driver's operational burden,aleviates driving fatigue,and improves safety while improving driving comfort.The existing ACC target recognition methods are difficult to meet the requirements in the complicated road conditions such as curves,lane changes,etc.When identifying targets,the main vehicle prediction trajectory is used as the reference without considering the actual lane line,and vehicles entering the main lane cannot be identified in advance.Misidentification can greatly compromise comfort and increase the risk of accidents.Therefore,researching a target identification method that accords with the actual driving specification and identifies target accurately,and can identify the vehicle in front of the host vehicle changing into the current lane in time,will further improve the work performance of ACC and extend its use scope.The ACC target recognition method studied in this paper includes two parts: the lane-recognition model and the identification of lane-change in side lane model.In the lane-recognition model,Kalman Filter based on vehicle linear two-degree-offreedom model is used to filter yaw rate obtained by sensor.The road curvature estimation model uses the host vehicle status data to estimate the road curvature in real time.Based on the assumption of a fixed curvature road,combined with lane line identification system data and radar data,a prediction lane is obtained,and the lane relationship can be determined based on the relative position relationship between the target vehicle and the predicted lane line.In the model of identification of lane-change in side lane,primary selection of recognition range to reduce model calculations.The lane change may be prejudged according to whether or not there is a continuous active steering.Based on the two situations of active steering,the problem of lane change recognition is transformed into a problem of the distinction between lane change and entrance/exit bend.The fourth and eighth polynomials are used to match the goodness of the lane change and the entry/exit curve.Paired t-test was introduced to test the significance of the differences to complete the identification of lane change trajectoriesBased on Simulink&Carsim's co-simulation environment,the ACC target recognition method is verified.The results show that the ACC target recognition method designed in this paper can accurately recognize the lane relationship of the front target vehicle in the straight and fixed curvature curves,and then determine the follow-up target.When the vehicle in the side lane cut into the host lane,the lane change behavior can be recongnized when it does not cross the lane line,and the recognition accuracy and real-time performance are better.