(Georgia Institute of Technology, 2019-08-28)
Sentis, Luis
Human centered robots provide embodiments whose purpose is to increase human productivity and comfort in complex non-structured environments. On one hand, operating in such environments demands new embodiments that enhance mobility and dexterous manipulation. On the other hand, productivity is an ill-defined term that often ignores key cognitive factors such as mental comfort or perceived workload. We posit that new embodiments consisting of legged manipulators endowed with high performance control and models of human workload behavior could ultimately result in new areas of enhanced productivity compared to simpler set ups. To make this point we need to demonstrate that teams made out of collaborating humans and human-centered robots will ultimately be more effective than teams with humans alone. In order to demonstrate these capabilities in an academic setup, my laboratory has spun-out a company to build some of the most advanced yet easy-to-use, human centered robots that exist today. I will delve in robot control theory and models of workload behavior as well as showcase videos of our new adult-size bipedal robots and full-body iron-man-like suit for heavy-duty logistic operations to emphasize our commitment towards experimental long-term studies.
As the world’s population lives longer, human-centered robotics emerges as a solution for the assistance, augmentation, and representation of humans to their comfort, productivity, and health. In this context, the Human Centered Robotics Lab studies key problems in mobility and manipulation of humanoid robots. In the first part of the talk, I will describe our work with the Office of Naval Research on endowing agile and compliant physical capabilities to bipedal robots. In particular, I will focus on the performance analysis of the whole-body operational space control framework during locomotion. In the second part of the talk, I will describe our work with NASA on physical human-robot interaction involving collisions between humans and mobile platforms and cooperative behaviors in rough terrains. Finally, I will comment on our work on building high performance actuators and software middleware for NASA’s Valkyrie humanoid robot.