A Distributed Architecture for Human-Drone Teaming: Timing Challenges and Interaction Opportunities.

Area of Science:

• Robotics and Human-Computer Interaction
• Distributed Systems and Autonomous Agents

Background:

• Drones are increasingly expected to operate autonomously while collaborating with humans.
• Existing drone systems often face limitations in computational power for complex tasks and seamless human integration.

Purpose of the Study:

• To propose and evaluate a distributed architecture for civilian human-drone teams.
• To offload complex operations like image recognition and human coordination from the drone to a remote system.
• To explore the trade-offs, specifically communication delays, associated with this distributed approach.

Main Methods:

• Design and implementation of a distributed system architecture.
• Integration of a smartphone as a remote processing and interface unit.
• Experimental analysis of messaging and command response delays using a custom Wi-Fi-connected drone.
• Application demonstration in a bookshelf inventory task.

Main Results:

• The distributed architecture enables enhanced computational power for tasks like image recognition.
• Remote processing allows for better integration of human interfaces for collaborative tasks.
• Experimentally observed communication delays between the remote system and the drone were quantified.
• The system was successfully applied to a practical use case of bookshelf inventory management.

Conclusions:

• A distributed architecture is a viable approach for supporting sophisticated human-drone collaboration.
• Careful consideration of communication latency is crucial for the effective implementation of such systems.
• The proposed architecture offers a flexible and powerful solution for future autonomous and collaborative drone applications.