Hardware-efficient neural networks for FPGA-based radio triggering of extensive air showers
Abstract
We present a hardware-efficient hybrid trigger for FPGA-based radio detection of extensive air showers. The hybrid design consists of a lightweight denoiser that cleans raw ADC traces and a compact classifier that operates on the denoised output, enabling robust near-threshold pulse detection in high-interference environments. Both neural networks are trained quantization-aware. Signals are generated from detector-folded CoREAS/CORSIKA simulations and embedded into measured noise to form a realistic benchmark. The trigger reaches an AUC of 0.992 while fitting comfortably within the resource budget of a Zynq-7000 Z-7020, with microsecond-scale latency and sub-watt power consumption. RTL validation confirms agreement between the fixed-point hardware and the quantized software model, demonstrating that neural denoising combined with classification provides reliable, low-cost radio triggering in noisy environments.
Versions
Related
- Hybrid neural denoising for resource-efficient near- and sub-threshold radio triggering of extensive air showers
- A dual-polarization whitened-template trigger for real-time radio detection of extensive air showers
- Macrocanonical Generator Networks: Data-Efficient Neural Surrogates for Amortized Physics Simulation
- Radio measurements of air showers with the IceCube-Gen2 surface prototype station at the Pierre Auger Observatory
- Reconstruction of very inclined air showers with radio data and application to the first GRANDProto300 cosmic-ray candidates
BibTeX
@misc{dimitrov2026hardwareefficient,
title = {{Hardware-efficient neural networks for FPGA-based radio triggering of extensive air showers}},
author = {Vesselin Dimitrov and Alperen Aksoy and Ilja Bekman and Markus Cristinziani and Eric-Teunis de Boone and Qader Dorosti and Chimezie Eguzo and Stefan Heidbrink and Stefan van Waasen and Andre Zambanini},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.16973},
eprint = {2609.16973},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.16973},
}
RIS
TY - UNPB
TI - Hardware-efficient neural networks for FPGA-based radio triggering of extensive air showers
AU - Vesselin Dimitrov
AU - Alperen Aksoy
AU - Ilja Bekman
AU - Markus Cristinziani
AU - Eric-Teunis de Boone
AU - Qader Dorosti
AU - Chimezie Eguzo
AU - Stefan Heidbrink
AU - Stefan van Waasen
AU - Andre Zambanini
PY - 2026
DA - 2026/09/15
PB - arXiv
DO - 10.48550/arxiv.2609.16973
UR - https://arxiv.org/abs/2609.16973
AB - We present a hardware-efficient hybrid trigger for FPGA-based radio detection of extensive air showers. The hybrid design consists of a lightweight denoiser that cleans raw ADC traces and a compact classifier that operates on the denoised output, enabling robust near-threshold pulse detection in high-interference environments. Both neural networks are trained quantization-aware. Signals are generated from detector-folded CoREAS/CORSIKA simulations and embedded into measured noise to form a realistic benchmark. The trigger reaches an AUC of 0.992 while fitting comfortably within the resource budget of a Zynq-7000 Z-7020, with microsecond-scale latency and sub-watt power consumption. RTL validation confirms agreement between the fixed-point hardware and the quantized software model, demonstrating that neural denoising combined with classification provides reliable, low-cost radio triggering in noisy environments.
ER -