Multisensor-multitarget tracking based on belief propagation against false data injection attacks and denial of service attacks

This paper is concerned with the multisensor multitarget tracking where the sensor network can potentially be compromised by adversarial attacks, including false data injection (FDI) attacks and denial of service (DoS) attacks. We propose a multisensor multitarget tracking algorithm against FDI and DoS attacks based on belief propagation (BP) message passing method. With the factorization of joint posterior density, the statistical structure of the tracking problem is described by a factor graph. A BP-based algorithm is derived based on the factor graph for an efficient evaluation of the marginal posterior densities of the target states. The marginal posterior densities are then utilized for the detection and estimation of the multitarget states. Then, we develop an efficient Gaussian mixture implementation of the proposed BP-based algorithm for the linear Gaussian measurement and state evolution model. Simulation results illustrate that the proposed multisensor multitarget tracking algorithm can provide reliable tracking performance against FDI and DoS attacks.