A Simulation-Based Approach to Decision Support for Robot-Human Team Configuration

Abstract

Missions that involve tasks such as search-and-rescue or reconnaissance have traditionally involved humans, perhaps with the assistance of one or more robots. The robots, or unmanned systems, are typically teleoperated - operated by remote control to inspect a suspicious object, for example. With the advent of newer and less expensive forms of autonomy and one-on-one communication, the robots are becoming more capable of acting as peers, rather than tools, alongside the humans. As the diversity of mixed human-robot teams is increased, so is the complexity of trying to decide on questions regarding configuration: what robots should be used, how many, and how many humans should be employed in the teaming process? We propose a decision support approach that builds a large number of discrete event simulations, with dynamic queuing models partially induced from CAD layout specifications in an urban environment. This work describes a detailed design with an implementation plan for the decision support.