search

project8 / katydid

Project 8 data analysis package
Other
5 stars 4 forks source link

Extend Digital Beamforming from fixed grid #182

Closed buzinsky closed 3 years ago

buzinsky commented 4 years ago

In KTChannelAggregator, the points on which the DBF is evaluated forms a fixed, rectangular grid. One question we were interested in was to find the maximum of the DBF, to the extent that the error in the position reconstruction is limited by noise. Of course, one could just use finer and finer grids to estimate the maximum, but this is going to be inefficient. Although, its not under the scope of this request, I can imagine a maximizer function, finding a region at which the DBF is large, and calling the ChannelAggregator with a finer mesh in that localized area, without redoing the entire transverse plane.

The request is this: I was wondering if we could generalize the rectangular grids to arbitrary arrays of points, so some other processor can do this search in the future. I think this entails a) defining a data structure like vector<pair> for the nonuniform "grid" (+ interface to it) b) adjusting the loops accordingly. The data structure should default to a uniform grid. I also think it will make the code a bit cleaner.

psurukuc commented 3 years ago

Commit 4e54684211557d25b716530e7529c1e2e2322e37 should implemented this.