Talk by Stephen McGill on 19.03.2014 at 15:00h - Human Interface Control of Next Generation Humanoids


Human Interface Control of Next Generation Humanoids


Until very recently, most humanoid robotics researchers have been far behind their ultimate goal of building a robot that can work like humans in unmodified environments. From the 2050 goal of seeing humanoids play FIFA soccer to the recent DARPA Robotic Challenge (DRC), robotics experts are beginning to adapt and select the most effective algorithms and hardware for the job of putting humanoids into the wild. One crucial aspect of this deployment is involving a human to aid the robot during difficult tasks. This talk will focus on the different levels of autonomy used in controlling a high degree of freedom humanoid, specifically with experiences from the recent DRC trials. Additionally, some details will be discussed about how the underlying modular software has evolved from RoboCup's kid sized turf to Team THOR’s cutting edge rescue robot.


Stephen McGill is a doctorate student at the University of Pennsylvania, researching with advisor Prof. Daniel D. Lee of the GRASP Lab. His main focus is adding intelligence to humanoid robots via sliding autonomy, from full autonomy to teleoperation. He is a leading member of Team DARwIn of RoboCup's kid-sized league and Team THOR of the DARPA Robotics Challenge, and is currently performing research under Prof. Daniele Nardi at Università di Roma "La Sapienza".