Effect of Communications Link Impairments on the Sparsity-based Distributed DoA Estimation

In this paper, distributed direction of arrival (DoA) estimation is proposed for a network of monostatic radar nodes, in which the radar nodes gather target reflected signals during the radar sensing phase, and then efficiently share this data with the central coordinator, during the communications phase. This is achieved by only using one-bit quantized versions of the node signals. We consider a two-node network, each having a uniform linear array (ULA). When viewed from a common reference node, the positions of node antennas form a co-prime array. The central coordinator builds sample covariance matrix using estimated signals and the knowledge of antenna element locations of the two nodes. We assume that the communications channel estimation errors can be modeled as additive Gaussian noise. It is shown that the distributed DoA estimation can be formulated as a sparse-recovery problem and solved within a compressed-sensing framework. Computer simulations show that the proposed distributed DoA estimation can effectively estimate more targets than the total number of available antenna elements at all nodes.