Resource allocation for target tracking in multiple radar architectures over lossy networks

Abstract

Communication packet losses are unavoidable in two-orientation communication between a fusion center (FC) and radar nodes, called a lossy network. In this paper, we propose a joint node selection and power allocation (JSPA) strategy of multiple radar architecture over a lossy network. First, a centralized multitarget tracking architecture considering bilateral packet loss is developed. Based on the centralized fusion architecture, the deterministic lower bound of the expected error covariance (DLBEEC) considering bilateral packet loss is derived. Then, minimizing the sum of weighted DLBEECs and the number of selected nodes simultaneously, we formulate the JSPA strategy as a bi-objective optimization problem. This optimization is an NP-hard problem involving binary and continuous variables, which is solved by a sparsity-enhancing sequential convex programming (SE-SCP) approach. Finally, numerical results show that the proposed JSPA strategy outperforms the traditional allocation strategies ignoring packet losses.