Task 010: Self-flying Drones and Visual Analytics

Abstract: The COVID-19 crisis has highlighted the need to bring robotics outside of the lab. Currently, Micro Aerial Vehicles (MAV) swarms cannot easily be tested outside of lab settings without access to external sensing infrastructure. Multi-robot systems until now have been with homogeneous teams of robots. Heterogeneous teams have higher potential to complete a wider variety of tasks more quickly. While there is active research in controlling heterogeneous teams of robots, up until now, heterogeneous multi-robot algorithms have only been tested in simulation. This is mostly due to a lack of good heterogeneous robotic test beds. 

 

In this talk, researchers will describe a framework which will not only enable bridging the gap between simulation and experimentation, but will also address some of the limitations of deploying multi-robot systems. They will present multiple contributions around a heterogeneous multi-robot test bed that is inexpensive and relies on smart-phone grade hardware. The method does not require any external knowledge of relative initial positions between drones or a prior set-up and solely rely on on-board sensors for localization. This work takes the first step towards a quick, easy to deploy heterogeneous swarm system without high overhead costs.

 

Bio: Dinesh Thakur is a Research Scientist in the Department of Mechanical Engineering & Applied Mechanics at the University of Pennsylvania working with Dr. Vijay Kumar. He received BS degrees in Electrical and Electronics Engineering from Birla Institute of Technology and Science, Dubai in 2008. He received his Masters in Robotics from the University of Pennsylvania in May of 2011.