A Miss-Detection Probability Based Thresholding Algorithm for an IR-UWB Radar Sensor

Abstract

In this paper, we propose a miss-detection probability based thresholding algorithm for an impulse radio ultra-wideband (IR-UWB) radar sensor. The conventional thresholding algorithm like constant false alarm rate (CFAR) is noise oriented algorithm which is focused on false alarm. Unlike this algorithm, the proposed algorithm is focused on miss-detection. For some applications miss-detection is more critical than false alarm, especially for short range application like IR-UWB radar sensor, such as anticollision system, intrusion detection. Moreover, the target signal oriented thresholding algorithm may better aware of the environment. As we know the amplitude of the signal varies depending on the installation height and angle of the radar. Conventional noise oriented thresholding algorithm is lack of considering these parameters. However, since the proposed algorithm is a target signal oriented thresholding algorithm, it is automatically optimized for these parameters. We analyzed the thresholds generated by the CFAR based thresholding algorithm and miss-detection probability based thresholding algorithm in the experiment result.