# Privacy protection method for ADS-B air traffic control data based on convolutional neural network and symmetric encryption

## TL;DR

This paper introduces a new method to protect privacy in ADS-B air traffic data using a combination of convolutional neural networks and symmetric encryption.

## Contribution

The novel integration of deep learning and symmetric encryption for ADS-B privacy protection is the key innovation.

## Key findings

- The proposed method effectively scrambles sensitive ADS-B data without instances of theft or damage.
- Encryption times scale linearly with data size, showing high efficiency for 10GB to 40GB datasets.
- The method outperforms existing approaches in encryption speed while maintaining strong privacy protection.

## Abstract

ADS-B (Automatic Dependent Surveillance-Broadcast) is a key surveillance technology in modern air traffic management, which broadcasts real-time aircraft information such as position, speed, and altitude for enhanced flight tracking and safety. However, the open broadcast nature of ADS-B communication raises significant privacy concerns, as sensitive data can be easily intercepted and misused. Research on privacy protection for ADS-B air traffic control data faces significant challenges, making the effective mining and safeguarding of privacy information a critical research focus.

This study proposes a novel privacy protection method that integrates deep learning with symmetric encryption. Specifically, by analyzing the ADS-B air traffic monitoring architecture, we mine and normalize privacy-related data to develop a Convolutional Neural Network (CNN)-based classification model for accurate identification of sensitive information.

Experimental results demonstrate that the proposed method effectively scrambles the original privacy information, with no instances of data theft or malicious damage. For data volumes of 10GB, 20GB, 30GB, and 40GB, the encryption times are 20.36ms, 30.56ms, 40.35ms, and 50.36ms, respectively, showcasing its efficiency.

Compared to existing methods, our approach achieves shorter encryption times while maintaining robust privacy protection. Future work could explore integrating advanced encryption technologies with state-of-the-art deep learning algorithms to further enhance the security of privacy protection in ADS-B systems.