Reconstruction of radio signals from air-showers with autoencoder

TL;DR

This paper presents a neural network autoencoder approach to denoise radio signals from air-showers, enabling improved reconstruction of cosmic ray parameters at lower energies.

Contribution

The study introduces a novel autoencoder-based denoising method that enhances air-shower parameter reconstruction for low-energy cosmic ray events.

Findings

Autoencoder effectively removes noise from radio signals.

Improved reconstruction accuracy for low-energy air-showers.

Enhanced detection efficiency at energies below 10^17 eV.

Abstract

The Tunka Radio Extension (Tunka-Rex) is a digital antenna array (63 antennas distributed over 1km^2) co-located with the TAIGA observatory in Eastern Siberia. Tunka-Rex measures radio emission of air-showers induced by ultra-high energy cosmic rays in the frequency band of 30-80 MHz. Air-shower signal is a short (tens of nanoseconds) broadband pulse. Using time positions and amplitudes of these pulses, we reconstruct parameters of air showers and primary cosmic rays. The amplitudes of low-energy event (E<10^17 eV) cannot be used for successful reconstruction due to the domination of background. To lower the energy threshold of the detection and increase the efficiency, we use autoencoder neural network which removes noise from the measured data. This work describes our approach to denoising raw data and further reconstruction of air-shower parameters. We also present results of the low-energy events reconstruction with autoencoder.