Sequential measurement-driven multi-target Bayesian filter for nonlinear multi-target models
- Resource Type
- Conference
- Authors
- Liu, Zongxiang; Zhang, Qiquan; Zou, Yanni
- Source
- 2016 IEEE 13th International Conference on Signal Processing (ICSP) Signal Processing (ICSP), 2016 IEEE 13th International Conference on. :1524-1528 Nov, 2016
- Subject
- Signal Processing and Analysis
Filtering algorithms
Adaptation models
Target tracking
Algorithm design and analysis
Bayes methods
Transforms
Gaussian distribution
multiple target tracking
marginal distribution
existence probability
- Language
- ISSN
- 2164-5221
The sequential measurement-driven multiple target Bayesian (SMB) filter is a valid method for multiple target tracking in situation of clutter interference and detection uncertainty. The known SMB algorithm spread the marginal distribution and existence probability of objective, and sequentially handles every receiving measurements. It satisfy closed solution in linear multiple objective models. Nevertheless, the solution is inapplicable to nonlinear Gaussian multiple target models. To handle this problem, we recommended a SMB filter algorithm to adapt nonlinear Gaussian multiple objective models. The recommended implementation applies the unscented transform method to handle the nonlinearity problems. The simulation experiment conclusions show that the recommended filter is more efficient on tracking multi-targets than the traditional PHD algorithm in situation of some clutter interference.