Machine Learning and Array Beamforming Military Boom Microphone

Frontier Audio Labs (Frontier) proposes adapting its commercial ìEar Computerî audio, microphone array, and machine learning algorithm technology for critical industrial, construction, and aviation workers to create the Machine Learning and Microphone Array Beamforming Boom Microphone (MLAB-BM) for aircrews in high ambient noise and wind environments (Figure 1). The MLAB-BM would be in the form factor of a boom microphone (compatible with HGU-68, HGU-56, and HGU-84 and David Clark and other headsets) replacing the existing boom microphones used.ÝÝLike other current ìnoise cancellingî boom microphones, the MLAB-BM would leverage a dynamic microphone capsule with a dual-port design that allows a generally flat 50Hz-15kHz response withÝ a signal to noise ratio above 72dB. The port spacing and internal acoustic chamber would allow 20-35dB of ambient noise rejection in the key speech band. Around the capsule would be a high-density wind screen which would then be surrounded by hydrophobic mesh and a highly durable dome with 50% open area. Together, this would provide slightly better or similar performance to the Ume Cobra andÝM87 microphones.ÝUnlike other current ìnoise cancellingî boom microphones, the MLAB-BM has a chamber (20mm diameter) at its end (where it attaches to the existing helmet/headset) that houses a low-power neural processing chip, four MEMs microphones, an impedance and voltage matching circuit, and a triple AAA battery (which is optional but enables optimal performance), as can be shown in Figure 2. These Ýcomponents power the (1) machine learning based speech denoising algorithms and the (2) directional differential microphone array which enable the MLAB-BM to far exceed the noise cancellation and speech intelligibility of the UmeCobra and M87 microphones.ÝÝThe overall objective of the Phase 1 is validating the ability of the MLAB-BM, and the adaption of Frontierís machine learning denoising and microphone array beamforming technology, to provide significant improvements to ambient acoustic noise cancellation of the transmitted electrical microphone signal.Ý