Case-Based Gesture Interface for Multiagent Formation Control

Abstract

Current multiagent systems require human operators to communicate in real-time with one another. A better option would be to have a single human operator control a swarm of agents using a natural interface, such as something worn by the operator. This includes being able to control the shape and movement of the agents in a dynamically shifting environment using an intuitive interface. Prior research has focused on moving the collective of agents as a whole, and has not used a case-based approach for this purpose. We show that a swarm can be shaped into different formations via a case-based gesture control strategy using a wearable (smart watch) interface. This approach enables a swarm to complete a task by taking high-level commands from an operator/user. We describe our approach, our initial results and demonstrations, and discuss next steps.