NLOS identification for UWB based on channel impulse response

The localization accuracy of ultra-wide band (UWB) system could be dramatically degraded, if the signal is propagated under non-line-of-sight (NLOS) condition. The detection of the NLOS propagation is very important to guarantee the accuracy of the UWB system. Based on the channel impulse response (CIR) sample, the NLOS condition could be identified. However, for the decawave chips, each CIR sample contains 1015 points. Thus the real-time realization of the NLOS detection with CIR is very hard, since the import and calculation of such a large amount of data cause to huge delay. In order to reduce the delay, the minimal needed size of the points in CIR for accurate NLOS identification is discussed in this paper. The support vector machine (SVM) is used for the classification based on the original CIR points or the eight different features extracted from each CIR. Furthermore, a new method is proposed for the identification based on the convolution algorithm. Compared to the existing approach with CIR, the needed CIR points for the detection are dramatically reduced, which makes the on-line identification realization possible. The accuracy of the NLOS identification with less CIR points is even better. The new proposed method using convolution algorithm also shows very promising results compared the other approaches.