A Low Computational Complexity JPDA Filter With Superposition
Robert Blair Angle, Roy L. Streit, Murat Efe
Abstract
Object superposition is a way to derive Bayesian estimators for multiple object tracking using point processes. A low computational complexity Bayesian multiple target tracking filter, based on target superposition, is presented. The concept of superposition is introduced and applied to the well-known Joint Probabilistic Data Association (JPDA) filter to derive the JPDA with superposition (JPDAS) filter. The JPDAS intensity function is evaluated to machine precision “for free” by computing the generating functional of the posterior process using complex arithmetic. A simulated example with eight targets is presented.
Topics & Concepts
Superposition principleComputational complexity theoryFilter (signal processing)AlgorithmProbabilistic logicComputer scienceEstimatorBayesian probabilityArtificial intelligenceMathematicsComputer visionStatisticsMathematical analysisTarget Tracking and Data Fusion in Sensor NetworksInfrared Target Detection MethodologiesGaussian Processes and Bayesian Inference