An Evolution of Assistive Robot Control to Meet End-User Ability

This research focuses on improving the control of assistive robotic arms for individuals with upper limb paralysis by using a Body-Machine Interface (BoMI). The goal is to translate small body movements into precise commands for the robotic arm. The study investigates whether traditional methods of creating control maps from body movements are sufficient for individuals with limited range of motion. They analyze the dimensionality of movements in both unimpaired and neuromotor-impaired individuals. The system involves a supervised map that predicts motion prompts and converts body kinematics into control signals for the robot. The system also uses a training paradigm that gradually increases the controllable dimensions and incorporates robot autonomy to assist the user.

• 1
  Traditional methods of BoMI map generation may not be suitable for controlling devices with many degrees of freedom (DoFs), especially for those with neuromotor impairments.
• 2
  The intrinsic dimensionality (ID) of free exploration data from unimpaired individuals is significantly lower than the number of DoFs required to position the end-effector of a robotic arm.
• 3
  The range of motion of cSCI participants during shoulder and arm prompts is comparable to that of the uninjured population.