Research On Target Recognition Method Of Automobile Active Cruise System

With the development of vehicle technology,the security of vehicle is valued seriously.As one of the active safety systems,adaptive cruise system can reduce fatigue of drivers and improve the safety and comfort of vehicle.Accurate target recognition method is the key and core technology of adaptive cruise system.Hence,it is very important to study on relevant target recognition method.In thesis,we study on relevant target selection method of adaptive cruise system based on millimeter-wave radar and Kalman filter theory.Firstly,it introduces the fundamental functions and main drawback and significance of adaptive cruise system in the thesis.Secondly,the recognition method is put forward by analyzing the current researching situations at home and abroad and the drawbacks of current target selection algorithm.Specific research contents and results are described as follows:(1)Estimate the vehicle state based on extended Kalman filter theory.Firstly,the three degrees of freedom vehicle model and the relevant extended Kalman filter are built respectively.Secondly,the filter is verified by co-simulation with Matlab/simulink and Carsim.Selecting steering wheel angle in double lane change condition as input,it is estimated on the yaw velocity and sideslip angle of vehicle.The results show that this method is accurate and effective and can be applied to relevant target recognition.(2)Study on relevant target recognition method of adaptive cruise system.A Kalman filter of target vehicle state is built based on distance and azimuth observed by radar,and then the longitudinal and lateral velocity and the yaw rate on target vehicle are estimated respectively.The road curvature is calculated according to the yaw rate and longitudinal velocity of target vehicle.Then,a expected driving trajectory on the host vehicle is defined based on setting road model in the thesis by compensated the host vehicle's direction angle.Finally,a lateral migration of the target vehicle relative to expected trajectory of the host vehicle is calculated.If the lateral migration is less than half road width and the distance is closest to host vehicle,it is just the relevant target of adaptive cruise system.(3)Verify the method by co-simulation with Matlab/simulink and Carsim.Firstly,the accuracy and effectiveness of estimating road curvature,which is obtained by dividing longitudinal velocity with yaw rate,are verified respectively.Then,the accuracy and rationality of estimating longitudinal and lateral velocity of target vehicle are validated.Finally,the accuracy and effectiveness of relevant target recognition method are verified in different cases,such as straight road,curve road,straight road into curve road,curve road into straight road and road-changing.The results show that the proposed target recognition method in the thesis can identify multi-target vehicles accurately.Moreover,this method can identify target vehicle effectively and accurately in complex conditions,like straight road into curve road,curve road into straight road and road-changing.The relevant target recognition method proposed in the thesis has certain reference value and significance for improving the traditional method of target selection on the fixed curvature and perfecting the stability and performance of adaptive cruise system.