Improving ROCs of Constant False Alarm Coded Anti-collision Radar in Very Noised Cases

Abstract

Abstract The use of Constant False Alarm Coded Anticollision Radar (CFACAR) is very interesting in automotive environment. Due to the orthogonality properties of used codes this system is most robust to multi-user interferences. The actual version of the receiver called in this paper Single Correlation Receiver (SCR), is not able to detect the targets in very low input Signal to Noise Ratio (SNR). To resolve this problem, we present a new receiver called Averaging Correlation Receiver (ACR), that computes the average of the M later correlations. Then, we developed the expression of detection and false alarm probabilities for the new receiver in mono and multi-user scenarios. These probabilities are used to plot the new Receiver Operating Characteristics (ROCs). They are drawn for different values of input SNR and length M of ACR. There is a suitable value of M, according to some equation, that can be taken to have a good detection (ROCs more perfect). Also, we found that for a fixed SNR, we must increase sufficiently the length M but it is possible only for low relative velocity of the target. For a velocity of 5Km/h with M = 1055, we can lessen the value of the SNR until we reach SNR = -45 dB.