Abstract
The study of human motor control using functional magnetic resonance imaging gives rise to many challenges. One of them is the design of haptic interfaces that are compatible with the magnetic field. To achieve this, the existing haptic interfaces employ parallel kinematics. However, they are limited to three degrees of freedom (DOFs). When trying to offer more DOF without floating actuators, parallel kinematics suffer from direct kinematic singularities, and thus, strong mechanical anisotropy. In this paper, we determine an optimal six DOF kinematics that overcomes these limitations. To this end, we use performance indices such as singularity occurrence, worst case output capabilities, sensitivity, and the global isotropy index. The resulting Octopod kinematics avoids a range of direct kinematic singularities by design. Finally, we present and evaluate a non-magnetic-resonance-compatible prototype of this novel type of kinematics.