Performance evaluation of CFAR detector for delay spread analysis of underwater acoustic channel

In this paper, we present the performance evaluation of constant false alarm rate (CFAR) detector for delay spread analysis of underwater acoustic channel, in shallow water environment. The model is developed with ray tracing of a Towing Tank, and validated by measurement with pseudo-noise (PN) sequence probe signal. Delay spread is obtained by using correlation between original of transmitter part with received signal. Threshold decision is based on CFAR technique, and has been developed by using three methods: Cell Averaging-CFAR (CA-CFAR), least square filter, and Sousa model. Comparison the three models with analytic model based on ray-theory, showed that the combination of CFAR with Sousa model achieved better performance that the others. This combination technique was able to differentiated samples of noise with samples of multipath components more accurately. The output shows the delay spread is similar to the analytical model.