This is the video of the paper "A multi-mode teleoperation framework for humanoid loco-manipulation" published in IEEE Robotics & Automation Magazine, and selected for presentation at ICRA 2020.
Despite the many successful cases of teleoperating mobile robots and manipulators, even in space, the teleoperation of humanoid robots still presents major challenges. While humanoid robots are designed with the ambition of mimicking the human body’s capabilities, differences in kinematics (e.g., structure and joint limits) and dynamics (e.g., mass distribution and inertia) are still significant. Another crucial issue is the necessity of ensuring the dynamic balance of the robot while trying to imitate human motion. This is not straightforward during locomotion tasks, in which the dynamics are highly involved.
A possible solution is, therefore, to use two forms of tele operation: a lowlevel one for manipulation, realized via whole-body teleoperation, and a high-level type for locomotion, based on the generation of reference velocities that are then tracked by the robot. We believe that this combination of different modes of teleoperation will considerably ease the burden of controlling humanoid robots, ultimately increasing their adaptability to complex situations that cannot be handled satisfactorily by fully autonomous systems.
We show our framework in action for the teleoperation of the humanoid robot iCub. The operator is equipped with a Xsens motion tracking suit and an Oculus VR system.
Paper here: https://hal.inria.fr/hal-02291907/document