Adaptive Kernel Kalman Filter Based Belief Propagation Algorithm for Maneuvering Multi-Target Tracking
Mengwei Sun, Mike E. Davies, Ian K. Proudler, James R. Hopgood
Abstract
This letter incorporates the adaptive kernel Kalman filter (AKKF) into the belief propagation (BP) algorithm for multi-target tracking (MTT) in single-sensor systems. The algorithm is capable of tracking an unknown and time-varying number of targets, in the presence of false alarms, clutter and measurement-to-target association uncertainty. Experiment results reveal that the proposed method has a favourable tracking performance using the generalized optimal sub-patten assignment (GOSAP) metrics at substantially less computation cost than the particle filter (PF) based multi-target tracking (MTT) BP algorithm.
Topics & Concepts
Tracking (education)ClutterKalman filterComputer scienceKernel (algebra)AlgorithmComputationRadar trackerArtificial intelligenceKernel adaptive filterTracking systemFilter (signal processing)Computer visionMathematicsFilter designRadarPedagogyCombinatoricsPsychologyTelecommunicationsTarget Tracking and Data Fusion in Sensor NetworksIndoor and Outdoor Localization TechnologiesDistributed Sensor Networks and Detection Algorithms