Towards a Geometry-Programmed Hand–Wrist Complex with Human-Hand-Inspired Soft Cover

Robotic hands and wrists intended for interaction with uncertain or delicate objects must combine compact motion, adaptive grasping, safe contact, and stable frictional behavior. Conventional rigid mechanisms provide accuracy and load capacity, but often require precise sensing and control to avoid damaging objects. Fully soft systems improve safety and conformability, but can be difficult to model, control, and integrate with predictable mechanical behavior. This thesis investigates an intermediate design philosophy in which geometry is used as a programming variable for passive mechanical behavior in robotic hand--wrist subsystems. The thesis is organized around the idea that robotic function can be partially embedded into the mechanical structure itself. Rather than relying only on actuation and feedback control, the work uses geometry to shape how a component moves, deforms, stiffens, forms contact, and resists slip. This principle is applied across three physical layers: wrist articulation, hand and gripper morphology, and soft contact interfaces. At the wrist level, Tetra-based compliant spherical joints are studied as compact flexure-based mechanisms capable of remote-center rotational motion. Their geometry-dependent behavior is analyzed through parametric modeling, finite-element simulation, and morphology refinement. Origami-inspired wrist mechanisms are then investigated as a complementary route in which rotational motion is encoded through crease geometry and rigid-foldable kinematics. At the hand level, monolithic soft fingers are designed with prescribed relative stiffnesses by optimizing compliant-joint geometry while keeping material and overall architecture fixed. This enables a passive stiffness hierarchy inspired by the nonuniform contribution of human digits during grasping. A second hand-side case study, the BIFLEX gripper, demonstrates how thumb-index-inspired grasping functions can be realized using soft-finger morphology, tendon routing, linkage geometry, and simple underactuated transmission. At the contact level, the thesis develops human-fingertip-inspired soft pads whose internal geometry is optimized to reproduce nonlinear fingertip-like compliance. The optimized pads are fabricated from silicone materials and experimentally validated through normal and tangential force measurements. Analytical stiffness models and contact-area models show that the replicated compliance also supports contact-area spread behavior relevant to tactile sensing. Finally, textured soft pads are experimentally evaluated under different loads, speeds, materials, and dry/wet conditions to assess friction magnitude and frictional stability. The thesis does not present a fully integrated hand-wrist-cover platform. Instead, it establishes subsystem-level scientific foundations for such a system. The results show that geometry can be used systematically to program rotational motion, finger stiffness, grasping behavior, contact compliance, contact-area evolution, and frictional stability, thereby reducing the burden on actuation and control in future adaptive robotic hands.