Revisiting Pushchino RRAT search using neural network

TL;DR

This study applied neural networks to search for rotating radio transients in LPA telescope data, successfully detecting four new RRATs and significantly reducing interference, demonstrating an effective method for pulsar detection.

Contribution

The paper introduces a neural network approach for RRAT detection that improves interference rejection and maintains high detection efficiency compared to previous methods.

Findings

Detected 4 new RRATs with low dispersion measures.

Reduced interference by 80 times using neural network.

Loss of real pulsar pulses does not exceed 6%.

Abstract

The search for rotating radio transients (RRAT) at declination from -9o to +42o was carried out in the semi-annual monitoring data obtained on the Large Phased Array (LPA) radio telescope at the frequency of 111 MHz. A neural network was used to search for candidates. 4 new RRATs were detected, having dispersion measures (DM) 5-16 pc/cm3. A comparison with an earlier RRAT search conducted using the same data shows that the neural network reduced the amount of interference by 80 times, down to 1.3% of the initial amount of interferences. The loss of real pulsar pulses does not exceed 6% of their total number.