Modeling Birth for the Labeled Multi-Bernoulli Filter Using a Boundary-Value Approach

What is it about?

The traditional Gaussian mixture model (GMM) approximation of constrained admissible region and probabilistic admissible region birth models yield large computational burden to the multi-target tracking filter. This study proposes a new birth model for the Labeled Multi-Bernoulli filter using the approach of boundary value problem (BVP) optimization for more efficient tracking. The BVP birth model determines a reliable orbital state represented by a single Gaussian component, and thus significantly simplifies the computational complexity.

Featured Image