ADS-B Reception Error Correction Based on the LSTM Neural-Network Model

Standard automatic dependent surveillance broadcast (ADS-B) reception algorithms offer considerable performance at high signal-to-noise ratios (SNRs). However, the of ADS-B in applications can be problematic low SNRs and interference situations, as detecting decoding techniques may not perform correctly such circumstances. In addition, conventional error correction have limitations their ability to correct errors messages, bit confidence values declared inaccurately event interference. The principal goal this paper is deploy a Long Short-Term Memory (LSTM) recurrent neural network model for conjunction with algorithm. data various flights are collected cleaned an initial stage. clean divided randomly into training test sets. Next, LSTM trained based on dataset, then evaluated dataset. proposed only improves packet rate (PECR), but it also enhances terms sensitivity. evaluation results reveal that scheme achievable efficient avionics industry. It worth noting algorithm algorithms’ prerequisites.