Comparing different schemes for random arrays

In this work, four types of random arrays are compared. In particular, the mean and variance of the array factor are derived. This provides a partial statistical characterisation that allows pointing out some important aspects of random arrays and link them to the number of elements and array aperture. In the absence of a simple and effective analytical apparatus, here great importance is also given to the experimental aspect, especially as far as the side-lobe level is concerned. To this end, Monte Carlo simulations are run to experimentally build the side-lobe distribution as a function of the number of radiators and the average spacing between two adjacent radiators. The obtained results show that random arrays where one is free to impose constraints on the minimum spacing between adjacent elements can obtain performance analogous to those achievable by other schemes which do not put such constraints. However, the former are preferable because they are able to zero the probability that adjacent radiators are separated with less than a certain minimum distance, which allows the mitigation of mutual coupling effects.