Extended (Conventional) Co-Prime Arrays and Difference Set Analysis: Low Latency Approach

TL;DR

This paper develops a comprehensive analysis of extended co-prime arrays, providing closed-form equations for their difference set properties and demonstrating low latency spectrum estimation through simulations.

Contribution

It introduces fundamental equations for the difference set of extended co-prime arrays and analyzes their bias, variance, and computational complexity, enhancing spectrum estimation methods.

Findings

Closed-form equations for weight function, bias window, and variance.

Optimal parameter choice ($M \\approx N/2$) improves bias characteristics.

Simulation confirms low latency spectrum estimation effectiveness.

Abstract

The co-prime array is a sub-Nyquist acquisition scheme for the estimation of second order statistics. It cannot generate all the difference values in the co-prime range and hence, one of the sub-array is extended to enable the estimation of the second order statistics at each difference value in the co-prime range. Recently, the difference set for the co-prime array was studied and low latency temporal spectrum estimation was demonstrated. In this paper, the fundamentals of the difference set of the extended co-prime array, also known as the conventional co-prime array, is developed. The closed-form equations for the weight function, correlogram bias window, and the variance are described. This is provided for the entire difference set, continuous difference set and for the prototype co-prime period. It is shown that the choice of $M\approx \frac{N}{2}$ generates a bias function with a large relative amplitude between the main-lobe and side-lobe peaks, where $M$ and $N$ are co-prime pairs. The expressions for the number of multiplications and additions required to compute the autocorrelation for the entire, continuous, and prototype range is derived. Simulation results demonstrate low latency spectrum estimation using the correlogram method.