Automated Body Control for Teleoperated Robots

Summary & Methodology Analysis

The researchers addressed the high cognitive load of whole-body teleoperation by moving away from hybrid interfaces that require independent manual control of robot components. The core methodology involves shifting from free-form control of heads and arms with separate touchpad inputs to a coupled approach. In this system, torso height and base rotation are automated through perception-centered coupling, which uses head pitch and head pan as signals. Simultaneously, manipulation-centered coupling adjusts the base and torso based on end-effector height and workspace boundaries. This integration effectively turns the robot into an extension of the operator body.

To manage these incoming data streams, the system utilizes a command synthesis layer. This component employs a threshold-based function and saturation to merge perception and manipulation inputs into final motor commands. To ensure predictable behavior, the system enforces a strict priority order for base translation, prioritizing backward, then sideways, and finally forward movements. The implementation relies on the TIAGo OMNI++ robot platform, the HTC Vive Pro 2 for motion tracking, and TRAC-IK for inverse kinematics, which is a solver that maps the desired end-effector position to robot joint configurations.

The research evaluated this framework using the NASA-TLX workload assessment tool. The results showed consistent improvements in performance, specifically in reducing the time required for complex grasping and placing tasks. By automating the secondary body movements, the system not only improved efficiency but also successfully minimized the robot reaching problematic configurations, such as joint limits, compared to traditional hybrid control methods. The paper does not specify the latency or processing overhead of the command synthesis logic.