Adaptive Multi-Bernoulli Filter Without Need of Prior Birth Multi-Bernoulli Random Finite Set

Conventional Multi-Bernoulli (MBer) filter assumes that the birth MBer Random finite set (RFS) is known a priori. However, this is not true for practical scenario. This paper proposes a novel extension of the MBer filter which eliminates the reliance of the prior birth MBer RFS and relaxes the limitation in new-born target appearance volume. The proposed filter classifies the measurements into survival measurements and birth measurements, and adaptively generates the birth MBer RFS using the birth measurements. The novel filtering equations that distinguish the persistent and new-born targets are derived. A Sequential Monte-Carlo (SMC) implementation of the proposed filter is given. Simulations are performed to verify the improvement in the performance of the proposed filter.