An Adaptive Multi-Sensor Generalised Labelled Multi-Bernoulli Filter for Linear Gaussian Models
Recent development of the multi-sensor generalised labelled multi-Bernoulli (MS-GLMB) tracking algorithm allows joint estimation of target trajectories adjunct to clutter rate and detection probability. Nevertheless, it requires prior knowledge of new birth target distribution which might not be ava...
| Main Authors: | , , |
|---|---|
| Format: | Conference Paper |
| Published: |
2022
|
| Online Access: | http://purl.org/au-research/grants/arc/LP200301507 http://hdl.handle.net/20.500.11937/96500 |
| Summary: | Recent development of the multi-sensor generalised labelled multi-Bernoulli (MS-GLMB) tracking algorithm allows joint estimation of target trajectories adjunct to clutter rate and detection probability. Nevertheless, it requires prior knowledge of new birth target distribution which might not be available in certain tracking scenarios. Conversely, another algorithm has been proposed to handle unknown birth statistics using multi-sensor measurement and a Gibbs sampler, but not be able to estimate clutter rate and detection probability. In this paper, we propose a multi-sensor multi-target tracking algorithm to handle unknown clutter rate, detection profile, and statistics of new birth targets. Our algorithm assumes linear Gaussian property on the dynamic and measurement models for closed-form analytic computation. Experiment with a 3-D tracking scenario demonstrates the robustness of our algorithm. |
|---|