AI for Radar in an Air Traffic Management System

该博士研究和开发可应用于安全关键雷达系统的深度学习分类和可解释性方法。民用空中交通管理（ATM）中的人工智能（AI）仍处于起步阶段。无人驾驶汽车（UAV）数量的增加威胁着低空飞行客机和机场的安全。目前，大多数用于空中交通管制（ATC）的机场主监视雷达（PSR）通常不执行飞机雷达特征的实时分类。该博士提出将深度学习（DL）架构纳入飞行器雷达特征的分类中，作为将这些特征映射到离散飞机类别的自动化方式。在这里，该研究将创建实时可解释AI（XAI）解决方案，从基于数据特征到基于符号来解释DL动作。

This PhD investigates and develops Deep Learning classification and explainability methods that can be applied to a safety-critical radar system. Artificial Intelligence (AI) in civilian Air Traffic Management (ATM) is still in its infancy. Increased number of Unmanned Autonomous Vehicles (UAV) threatens the safety of both low-flying passenger jets and airports. Currently, most airport Primary Surveillance Radars (PSR) used for Air Traffic Control (ATC), do not typically perform real-time classification of aircraft radar signatures. This PhD proposes the incorporation of the deep learning (DL) architectures in the classification of air vehicles radar signatures as an automated way of mapping these signatures to discrete aircraft classes. Here, the research will create real-time explainable AI (XAI) solutions ranging from data feature based to symbolic based to explain the DL actions.