Exploiting sensor mobility and covariance sparsity for distributed tracking of multiple sparse targets

Abstract

The problem of distributed tracking of multiple targets is tackled by exploiting sensor mobility and the presence of sparsity in the sensor data covariance matrix. Sparse matrix decomposition relying on norm-one/two regularization is integrated with a kinematic framework to identify informative sensors, associate them with the targets, and enable them to follow closely the moving targets. Coordinate descent techniques are employed to determine in a distributed way the target-informative sensors, while the modified barrier method is employed to minimize proper error covariance matrices acquired by extended Kalman filtering. Different from existing approaches which force all sensors to move, here, local updating recursive rules are obtained only for the target-informative sensors that can update their location and follow closely the corresponding targets while staying connected. Simulations advocate that the proposed scheme outperforms alternative tracking schemes while accurately tracks multiple targets by imposing mobility only on the target-informative sensors.

Survey: mobile sensor networks for target searching and tracking

“…The extended Kalman filter is mainly used by mobile sensor networks to handle non-linear measurement models, (3b), while the target model, (3a), is usually linear. This is the case for both Ren et al [29] (distance-to-target measurement model) and Wu et al [30] (range-bearing sensor model). In Ren et al [29] the authors also exploit the covariance matrix, P k , in their control law.…”

“…This is the case for both Ren et al [29] (distance-to-target measurement model) and Wu et al [30] (range-bearing sensor model). In Ren et al [29] the authors also exploit the covariance matrix, P k , in their control law. Martínez and Bullo [31] use a general target model in their derivation, and in simulation, use an 8-shaped movement for the target.…”

“…First, it is depended on the initial state estimate of the target for the estimate error to converge to zero. In Ren et al [29], the authors use a sparsity decomposition scheme to initialize the targets to ensure the initial estimate error. InWu et al [30] and Martínez and Bullo [31] there is less discussion on how to initialize the target positions.…”

“…Both [83] Searching [140,144,118] Tracking [28,134,101,27,116,46,117] Single Target Both [] Searching [30,133] Tracking [24,26,42,47,48,70,125,111,110,112,113,124,115,135,107] Fig. 5: Classification of centralized approaches presented in this survey Distributed Multiple Targets Both [72,34,142,127,126,128,84] Searching [130,121,138] Tracking [53,119,129,82,122,94,95,29,81,106,…”