This repository contains code to compute the dechirping efficiency (Margot et al. 2021) of radio technosignature data-processing pipelines such as those used by the Breakthrough Listen team in Enriquez et al. (2017) and Price et al. (2020). The dechirping implemented in this notebook comes from the turboSETI package.
In this notebook, we provide the code to replicate the Breakthrough Listen (BL) portion of Figure 7 of Margot et al. (2021):
Additionally, we provide the code to generate Table 4, which lists the dechirping efficiency statistics of the tree algorithm for various array dimensions:
| Rows | Min (%) | Max (%) | Mean (%) | Median (%) | STD (%) |
|---|---|---|---|---|---|
| 2 | 100.00 | 100.00 | 100.00 | 100.00 | 0.00 |
| 4 | 100.00 | 100.00 | 100.00 | 100.00 | 0.00 |
| 8 | 75.00 | 100.00 | 93.75 | 100.00 | 11.57 |
| 16 | 75.00 | 100.00 | 90.62 | 93.75 | 10.70 |
| 32 | 68.75 | 100.00 | 85.16 | 81.25 | 11.20 |
| 64 | 68.75 | 100.00 | 81.64 | 78.12 | 9.83 |
| 128 | 64.06 | 100.00 | 77.93 | 75.00 | 8.88 |
| 256 | 64.06 | 100.00 | 75.17 | 73.44 | 7.68 |
| 512 | 60.16 | 100.00 | 72.42 | 71.09 | 6.84 |
| 1024 | 60.16 | 100.00 | 70.08 | 69.14 | 6.10 |
| 2048 | 56.84 | 100.00 | 67.92 | 66.60 | 5.53 |
| 4096 | 56.84 | 100.00 | 66.01 | 64.94 | 5.06 |
Notebook author: UCLA graduate student Paul Pinchuk