Research On Classification And Recognition Technology Of Vehicle Millimeter Wave Radar

Automated or assisted driving is an important technical approach to increase the level of intelligence in increasingly complex and congested road traffic,and promote the transformation and upgrading of the automotive industry.With the rapid development of integrated circuit and radar system technologies,the vehicle-mounted millimeter wave radar has broken through the applicaton limitations of volume,power consumption,cost and other factors.With high reliability and high precision detection capabilityes all-day and all-weather,millimeter wave radar has become the core sensor of the vehicle.The research on automatic target classification and recognition technology based on the vehicle-mounted millimeter wave radar can provide technical support for the accurate forming automatic driving strategies.Focusing on the technical challenges of classification and recognition of road traffic subjects such as pedestrians,motorized and non-motorized vehicles,large and small motorized vehicles by vehicle-mounted millimeter wave radar,this thesis carried out feature extraction besed on target motion and electromagnetic scattering features,studied classification and recognition technology on multi-radar-features fusion,and conducted experimental research based on the measured data.The results of road target classification and recognition based on the above method have obtained high accuracy and high promoted application value.The main work and contributions of this paper are as follows:1.The working mode and the applicable requirements of millimeter wave radar system are analysed.The measurement principles of target location parameters and feature information based on the LFMCW radar signal.2.The key features required for the classification and recognition of different types of road targets are analyzed.The extraction and processing technologies based on the target motion features and electromagnetic scattering features are studied.Complete the feature extraction and transformation various types of targets based on the measured data,and provide the basis and data foundation for the classification and recognition system.3.Contructed a classification machanism based on road targets and studied basic classification and recognition methods and algorithm characteristics.According to actual application scenario requirements,constructed a decision tree and support vector machine combined classifier based on multi-feature fusion,and carried out the training and design of the classifier based on the measured data.4.Built an experimental platform based on AWR1243 millimeter wave radar,and conducted the experimental test and regognition performance evaluation of fusion classification algorithm.The correct rate of classification and recognition was better than 90%,which proves the feasibility of target classification and recognition scheme of millimeter wave radar.