The implementation I made is very rough and I'm not sure whether it is really working right lmao. I think we should create a dataset of recordings to actually test whether the implementation is accurately measuring the angle and distance of the target.
In addition, I noticed that the calculation for the sound source localization is not very performant either. I think adding performance improvements would be helpful.
I created an initial implementation of the sound source localization using frequency-domain steered beam-forming approach from this paper. With my implementation, the microphone positions can be configured in the
LocalizationConfigclass.The implementation I made is very rough and I'm not sure whether it is really working right lmao. I think we should create a dataset of recordings to actually test whether the implementation is accurately measuring the angle and distance of the target.
In addition, I noticed that the calculation for the sound source localization is not very performant either. I think adding performance improvements would be helpful.