Getting to the heart of human-robot interactions
Exploring the possibilities for real-world applications
Michael Lewis, Professor
If it’s dusk and a toddler wanders away from her parents in a state park, it’s easy to picture rescue workers frantically searching for the child before daylight fades. What if, instead of sending humans and helicopters, a multi-robot team could search for her? The robots could scatter through the woods, scanning wide geographical areas, increasing the time-efficiency, and, perhaps, the outcome of the search effort. These types of robots teams have potential for use in rescue missions, environmental cleanups, large-scale emergency response, and military operations.
The research conducted by Michael Lewis addresses the number of technical problems that are inherent to working with robot swarms.
Multi-robot systems are called swarm robotics since they operate with similar characteristics to a swarm of insects—they use control laws based on local spatial information about the environment and/or the other members of the swarm within their spatial neighborhood—and they adjust their behavior accordingly.
In order for a human to supervise a swarm, the person must comprehend the swarm’s state and be able to predict the effects of the human’s inputs on the swarm’s behavior. This understanding is essential to achieve the swarm’s optimal performance.
Lewis has been writing papers and conducting research on human-robot interactions since the early 2000s.
“Mobile robots provide a much more difficult problem than stationary robots,” explains Lewis. “Manipulation is its own set of problems in terms of all these degrees of freedom and the problems of motion planning.”
He is currently investigating the influence of culture on the development and maintenance of trust in automation. “This will give us a very good validation sample and let us look precisely at how trust is being expressed subjectively in different populations,” said Lewis.
The author of more than 200 scientific papers in the areas of Human Factors and Autonomous Systems, Lewis is also an associate editor for the leading journals in the field: Human Factors and IEEE Transactions on Human Machine Systems.
Human factors, human computer interaction, visualization
Michael Lewis is a professor at the University of Pittsburgh’s School of Computing and Information.