Demonstrating Superresolution in Radar Range Estimation Using a Denoising Autoencoder
Robert Czupryniak, Abhishek Chakraborty, Andrew N. Jordan, John C. Howell · 2025
Read Demonstrating Superresolution in Radar Range Estimation Using a Denoising Autoencoder by Robert Czupryniak, Abhishek Chakraborty, Andrew N. Jordan, John C. Howell (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-14906.txt.
Description
Robert Czupryniak and co-authors (2025) used a neural network to measure radar distances finer than the usual limit. A radar's range sharpness is normally capped by its bandwidth. The team asked a denoising autoencoder to estimate the gap between two equal reflectors closer than that cap. The network squeezed each returning waveform into a tiny bottleneck code. That code tracked the true gap closely across training runs.
Pulse shape decided how well the trick worked. A pulse built from spherical Bessel functions performed best. A triangle-style pulse came second, and the standard sinc pulse trailed. The bottleneck stayed stable when noise was added, which matters for field use. The authors present the work as a demonstration rather than a finished sensor. It shows that learning the right compression, on the right waveform, can beat the classic resolution limit in tests. Signal design, they stress, is half the method.
Catalog record
| CREATOR | Robert Czupryniak, Abhishek Chakraborty, Andrew N. Jordan, John C. Howell |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 29 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-14906