An Intent-Based Reasoning System for Automatic Generation of Drone Missions for Public Protection and Disaster Relief

Abstract

The utilization of drones for search and rescue operations has become more prevalent over the years. Drones can provide an aerial perspective which can aid first responders in gaining an overview of a situation. Autonomous drones can automate search and rescue operations by removing the human pilot, which can increase efficiency and lower costs. The increased development of machine learning models and techniques has paved the way for intent-based reasoning systems that can understand users' intent. This can allow users to control autonomous drones by expressing their intent. Which can be utilized for search and rescue operations.

However, machine learning models require vast computational power and data storage. In addition, autonomous drones have high-performance requirements. The development of 5G can provide the infrastructure required to meet the stringent performance requirements of machine learning models and autonomous drones. By leveraging the advanced features of 5G, such as network slicing, high-speed communication, and low latency, it provides the infrastructure that supports the use of machine learning models in coordination with drones.

This thesis proposes a system prototype that can generate drone missions based on user intent which can be used for rescue operations. The system utilizes a large language model and automatic speech recognition model to capture the intent of the user and generate drone missions that integrate with a 4G-enabled drone. The evaluation of the system reveals that the system can reliably capture the user's intent with simple commands, but struggles with more complex commands. The prototype demonstrates that intent-based reasoning systems for controlling autonomous drones using 5G technology can aid first responders during PPDR missions.