As assistive machines increase in dimensional complexity, they often become more difficult to control. This can often mean that controlling these complex machines becomes  inaccessible to people with motor impairments. We explore the use of Body Machine Interfaces (BoMI), a novel interface that transforms body motions to a high-dimensional joystick-like controller. We aim to determine whether the BoMI can effectively control complex assistive robotic machines, and how we can leverage robot learning for human motor learning. The proposed solution is to leverage robotics autonomy to gradually transfer more control to the human operator as they gain competency. 

Our studies revealed key insights into user learning with BoMI, including differences between joint-space and task-space control, as well as strategies for scaling autonomy assistance. These findings inform the development of systems that help users master complex control interfaces, expanding access to assistive technologies.