Research Of Target Recognition In Uwb Communication Using A Support Vector Machine

Impulse-radio(IR)-based Ultra-wide band(UWB) technology is a recent break-through in broad-band wireless technology. It offers evident advantage such as high throughout, high resolution, low power consumption, and shows great potential when applied in positioning, data transfer, target detection and tracking. Due to its excellent ability in penetrating, target detection using UWB is widely used in military suspicious target alarm, signs of life detection while rescuing, oil detection during resource exploration.According to the given references, in most cases, people use radar system and analyze the echo signal to extract typical characters of target. In this paper, the signal sent by emitter is received and extract target feature from the received signal, then recognize the target. And also, Support Vector Machine (SVM) is an important method of pattern recognition which is widely used in target detection,classification and forecasting. It has distinct advantage in analyzing and classify small scale samples. So in this paper, SVM is used to classify and recognize target after feature extraction.What’s more, we will prove this method in both experimental and practical ways.Firstly, UWB indoor propagation channel models are established using Finite-Difference Time-Domain (FDTD) method to reflect different obstacles. The received signal under different conditions and features are extracted for SVM. Different SVM kernel is used to detect the obstacle and parameters of kernel is picked to optimize the accuracy. Also we modify the noise level to investigate the trends of detection accuracy. This effort offered us a preliminary proof of our method. At last, communication signal in real scenarios is collected, the same steps are taken to investigate the ability of target recognition during IR-UWB communication. Grid research and5-fold cross validation are used to optimize parameters of SVM, and the accuracy is improved. This effort does confirm the feasibility of detecting different targets during IR-UWB communication.