Real-time implementation of an adaptive bayesian beamformer

An implementation of the adaptive Bayesian beamformer is examined for execution on field-programmable-gate-array (FPGA) devices. Using multiple sensor inputs, the Bayesian beamformer can estimate the direction-of-arrival (DOA) of a low-power signal in an environment that is simultaneously populated by high-power interference of limited DOA knowledge. A weighted sum of a discrete set of beamformers with known associated DOAs forms the Bayesian beamformer. Previously observed data provides the basis for the calculation of the a posteriori probability distribution function that renders the sum weights. This paper incorporates further approximations to the derivations to allow for its implementation on FPGA devices, in particular those with lesser gate counts. The feasibility of an all-FGPA implementation versus a heterogeneous implementation is explored