A tutorial on particle filters for online nonlinear/non-GaussianBayesian tracking
IEEE Transactions on Signal Processing
Hi-index | 0.00 |
Because the state model is too simple to capture the complicated trajectory, the traditional tracking algorithm can not track the moving people precisely. Simon J. Godsil and others improve particle filter to present variable rate particle filter, which employs more complicated model to track moving target, and samples state variables by changing sampling period. In this paper, we propose an adaptive variable rate particle filter algorithm for large-scale building fire rescue system to track trapped people. The algorithm can adaptively adjust the sampling period by comparing the relationship between particle value and measurement. According to our computer simulation, the method can improve the tracking accuracy, especially as the target trajectory has some drastic changes. we consider that the tracking algorithm is effective and practical.