Embedded Computation Architectures for Autonomy in Unmanned Aircraft Systems (UAS).

Area of Science:

• Aerospace Engineering
• Computer Engineering
• Robotics

Background:

• Future Unmanned Aircraft Systems (UAS) require significant onboard computing power for increased autonomy.
• Existing embedded systems face challenges in meeting these escalating computational demands.

Purpose of the Study:

• To analyze the computational resource requirements for future autonomous UAS.
• To propose a framework for classifying UAS tasks and defining autonomy levels.
• To investigate the implications of autonomy on embedded system design.

Main Methods:

• Classification of common UAS tasks (flight, navigation, safety, mission) and executing entities.
• Analysis of task combinations to define autonomy levels.
• Review of current state-of-the-art hardware and software for UAS tasks.

Main Results:

• Established a link between UAS applications, required tasks, autonomy levels, and computing resource needs.
• Identified that current computing architectures are inadequate for future UAS autonomy.
• Highlighted the potential of dynamically reconfigurable hardware for improved performance and energy efficiency.

Conclusions:

• Future autonomous UAS necessitate a re-evaluation of embedded system design.
• Dynamically reconfigurable hardware presents a promising solution for meeting future computational demands.
• The embedded system should be considered a critical component in UAS development.