2020/09/15 by Christopher Thron, Thron, Christopher, Khoi Q. Tran +3
Computer Science · Engineering · Earth and Planetary Sciences · #Target Tracking and Data Fusion in Sensor Networks #Structural Health Monitoring Techniques #Meteorological Phenomena and Simulations
paper · pdf · doi:10.48550/arxiv.2009.08310
This paper develops an accurate, efficient filter (called the `TT filter')\nfor tracking multiple targets using a spatially-distributed network of\namplitude sensors that estimate distance but not direction. Several innovations\nare included in the algorithm that increase accuracy and reduce complexity. For\ninitial target acquisition once tracking begins, a constrained Hessian search\nis used to find the maximum likelihood (ML) target vector, based on the\nmeasurement model and a Gaussian approximation of the prior. The Hessian at the\nML vector is used to give an initial approximation of the negative log\nlikelihood for the target vector distribution: corrections are applied if the\nHessian is not positive definite due to the near-far problem. Further\ncorrections are made by applying a transformation that matches the known\nnonlinearity introduced by distance-only sensors. A set of integration points\nis constructed using this information, which are used to estimate the mean and\nmoments of the target vector distribution. Results show that the TT filter\ngives superior accuracy and lower complexity than previous alternatives such as\nKalman-based or particle filters.\n