Resource Scheduling for Multi-Target Tracking in Multi-Radar Systems With Imperfect Detection

Abstract

In this paper, an effective joint radar assignment and power scheduling (JRAPS) scheme is proposed for the multi-target tracking (MTT) task in multi-radar systems (MRSs) under imperfect detection, namely, target related measurements are collected with the probability of detection (<inline-formula><tex-math notation="LaTeX">$\bm {P}_{\bm {D}}$</tex-math></inline-formula>) less than 1. The centralized fusion architecture is adopted by the MRS. Specifically, during each sampling time interval, each radar of the MRS is assigned to track certain targets with controllable transmitting power. The measurements collected by all selected local radars, with <inline-formula><tex-math notation="LaTeX">$\bm {P}_{\bm {D}}\leq 1$</tex-math></inline-formula>, are sent to a central radar to obtain the global MTT results. To maximize the global MTT task performance, the proposed JRAPS scheme implements the online target-to-radar assignment and transmitting power allocation scheme based on the feedback of the MTT results. The posterior Cram&#x00E9;r-Rao lower bound (PCRLB) with <inline-formula><tex-math notation="LaTeX">$\bm {P}_{\bm {D}}\leq 1$</tex-math></inline-formula> is derived and utilized as the tracking performance metric since it provides a more accurate lower bound on the target state estimates under imperfect detection. Then, an overall cost function (OCF) is formulated based on the derived PCRLB to quantify the global MTT performance. Combined with the practical resource constraints of the MRS, the formulated JRAPS problem is shown to be non-convex. Therefore, we further propose a fast three-stage iterative method to solve this problem efficiently. Simulation results verify the superiority and effectiveness of the proposed JRAPS strategy in terms of both tracking accuracy and target detection performance.