UIST '26 (Poster)
Bimanual interaction technique for controlling humanoid robot walking and working
Joon Hyub Lee
Siripon Sutthiwanna
Sang-Hyun Lee
Jin Park
Seok-Hyung Bae
Humanoid robots are difficult for novice users to control intuitively due to their high degrees of freedom. We propose a novel interaction technique that enables users to control the locomotion and manipulation of a humanoid robot by holding and moving a virtual 3D plane with both hands. We implemented our technique using a Unitree G1 humanoid robot model in a VR physics simulation and compared it against gamepad and exoskeleton baselines on tasks from an open-source humanoid robot benchmark in a user study with 24 participants. Our technique was 59% faster and improved benchmark results by 17% in a locomotion task, while it was 41% faster and improved benchmark results by 20% in a manipulation task. It also outperformed the baselines on most subjective workload measures.
Poster
