Acoustic beamforming techniques are often used in applications involving microphone arrays such as sonar, binaural hearing aid devices and acoustic indoor localisation. The beamforming methods utilized describe results for a particular array within a given acoustic frequency domain. Software and toolboxes running locally facilitating the theoretical design of such implementations are largely available. However, a multiuser cloud-based beamforming approach to generate results could not be found. Also, generating packages facilitating the implementation on FPGAs has not been carried out to our knowledge. In order to alleviate these shortcomings, we propose a scalable cloud-based accessible multi-user beamforming emulator where users launch emulations based on acoustic beamforming for microphone arrays. Users create, upload, modify and run beamforming emulations. Our platform is also able to generate VHDL packages containing delaying tables which facilitates the implementation of HDL based Delay-and-Sum beamformers. Analysis based on algorithm truncation errors are also possible due to the integration of fixed-point signal processing algorithms. Microphones in a given array can be disabled in order to estimate the response of the remaining microphones during power saving. Results and packages can be downloaded from the user web interface as an archive file. Our Emulator can be configured to take several user defined configurations into account such as microphone arrays, sound source placement and emulation parameters. Graphs such as waterfall diagrams and directivity showing the quality of the beamforming can also be generated.
|2018 4th International Conference on Cloud Computing Technologies and Applications, Cloudtech 2018
|4th international conference on cloud computing technologies and applications
|26/11/18 → 28/11/18