Order Statistics Based Multidimensional Clustering Scheme for FMCW Radar Targets

Target detection using an frequency-modulated continuous-wave radar system often leads to an unsorted list of n-tuples for found frequencies, magnitudes and -sometimes - phases of sinusoidal components within the input signal. The task covered in this paper is to sort and analyze these unintuitive lists of n-tuples and obtain the parameters of the underlying physical targets. This is achieved by exploiting properties of the distribution of the n-tuples and their order statistics. We show that the assessment of range, velocity and direction of arrival in radar systems can be handled by our clustering algorithm.