Multiple-model algorithms for distributed tracking of a maneuvering target

The paper deals with distributed tracking of a maneuvering target by means of a network of heterogeneous sensors and communication nodes. To effectively cope with target maneuvers, multiple-model filtering is adopted after being extended to a fully distributed processing framework by means of suitable consensus techniques. Novel Distributed first-order Generalized Pseudo Bayesian (DGPB₁) and Distributed Interacting Multiple Model (DIMM) algorithms are presented. Simulation experiments on critical tracking case studies involving a highly maneuvering target and sensor networks characterized by weak connectivity and target observability properties demonstrate the effectiveness of the proposed distributed multiple-model filters.