Contact Control of the Index Finger

Abstract

This thesis aims to develop a tactile-feedback control model of the human index finger that characterizes its bones, tendons, ligaments, muscles, nerves, and tissues using MATLAB/Simulink/Simscape/SimMechanics. The index finger curls from full extension and makes contact with a spherical object located at an arbitrary position. It presses into the object and the compliant tissue of the finger pad deforms until the normal force exceeds its threshold, at which point the relevant joints are prevented from rotating further. Once the magnitude of the external PD-controlled torque that generates the desired full extension-to-full flexion motion falls below the torque applied to the joint to keep its net torque at zero, the finger extends back to its initial position. The 3D simulation of the entire human hand with a predefined motion path provides a visual context for the index finger. The torques generated from this 4-dimensional joint space are optimally inversed to the 7-dimensional tendon space to acquire the required force along each tendon to achieve the contact-controlled motion of the index finger. This work is motivated by future designs of prosthetics and anthropomorphic robotic hands that make use of haptic information as they operate in unknown and unstructured surroundings.