Study On Target Recognition For Vehicle Active Anti-Collision Warning System Based On MWR

The frequent traffic accidents cause huge economic loss to the society, and bring greatcost of life to humans, which has been taking people’s great attention on the safety of the car.Vehicle active anti-collision warning system is becoming the focus in the automobileadvanced technology. The using of active anti-collision warning system based on MWR(millimeter wave radar) has small influence by the road environment and weatherenvironment, plays an important role to reduce the incidence of traffic accidents.Researching the target recognition technology basing on the millimeter wave radar, which isfrom the major scientific and technological innovation project of Jilin province departmentof "vehicle active anti-collision warning system research and development", and the maincontents are as follows:Firstly this paper analyzes the principle of radar, and MWR is chosen as the collisionavoidance radar by combining with the performance characteristics of vehicle activeanti-collision radar. When the complex target is recognized, the target is consisted ofmultiple scattering centers. According to the meaning of radar target recognition technology,the target positional information and feature information is adopted as the discriminantinformation of target recognition. By analyzing intermediate frequency signal, targetidentification information can be gotten directly, and the related theory of intermediatefrequency signal is discussed.Then the basic theory of target recognition is studied. By the comparative study oftarget feature information extraction method, the extracted RCS (radar cross section)sequence is selected as benchmark parameter of target character information. That is to say,the differences between the target RCS sequences are utilized to realize the identification ofthe target type.Next the type and characteristics of the goal in front of car are analyzed and the nearestdangerous target information as well as pedestrian information within the detection range are the outputs of the target recognition system. Then the positioning algorithm and targetfeature information extraction algorithm are designed, namely the positioning information istarget orientation, speed and distance information, and target feature information extractionalgorithm is to extract sequence characteristic parameters of radar cross-section, and thefuzzy controller is designed, and the velocity information, RCS sequence information are asinput characteristic parameters to realize the target type recognition. According to thelocation information, the multi-objective algorithm is designed.Lastly the simulation experiments analysis is carried out. The people,vehicle and fenceRCS features library are built, the transmitting and receiving signal model, the targetcharacteristic and positional parameters model as well as the fuzzy discrimination model ofthe target are built individually. The target recognition model simulation is analyzed withthe help of the constructed models and the effectiveness of target recognition algorithms isalso verified.Through this study, the target recognition algorithms are determined in the field oftarget recognition technology, which establishes a theoretical framework for the targetrecognition technology of active anti-collision warning system and lays out a theoreticalfoundation for the future study.