Modeling and Kinematic Analysis of Human Finger, Wrist, and Forearm Mechanisms using Parallel Linkages

Abstract

The human forearm and wrist are in charge of spatial orientation of the hand, which frequently interacts with various objects or environmental constraints. Systematic movements of complex skeletal structures widen the range of motion (ROM) of the wrist and physical adaptation of joints by elastic relaxation of the muscle plays a major role in the interaction of the finger. In robot design, it is essential to equip mechanical systems for the orientation and interaction of an end-effector for dexterous manipulation. In this thesis, the following characteristic behaviors of the human finger, wrist, and forearm are investigated and developed by using parallel linkages:

In the first place, the importance of the pre-shaping motion, that each phalanx approaches an object to be grasped at the same time, is emphasized in the design of the underactuated finger mechanism. At that time, the finger mechanism able to realize this motion as well as shape-adaptation to arbitrary objects is proposed. Especially, the proposed mechanism makes no deformation of an equipped spring during the pre-shaping motion, thus reducing the load which an input actuator copes with. A method of the dimensional synthesis for the pre-shaping motion is dealt with. Besides, transmission qualities of the mechanism are evaluated to reduce input-output sensitivity in the synthesis procedure. Some quantitative simulations and qualitative experiments are conducted to confirm the design feasibility.

Secondly, lots of conventional underactuated fingers have been designed with the emphasis on the shape-adaptation to the object. However, most target objects lie on the rigid tabletop or around obstacles, thus expecting the fingertip contact with them. The design method of the finger mechanism to overcome the contact situation is addressed. The proposed method makes the fingertip slides on the surfaces towards objects to be pinched. For this purpose, the fingertip force direction is investigated via geometric and analytic approaches. The method of the dimensional synthesis to design the desired force direction is proposed with several design criteria. The synthesized finger mechanism is simulated to confirm the validation of the design criteria. A gripper with the synthesized finger mechanism designed by the criteria is developed and experimented to show the availability of the synthesis method. Some demonstrations are conducted to confirm its practical effectiveness as well.

Thirdly, a wrist mechanism with two rotational degrees of freedom (DOFs) is analyzed to figure out kinematic characteristics and to improve its performance. For the improvement, an overconstrained anti-parallelogram linkage as the structural design is introduced to realize the large ROM and reduce parasitic motions. In this study, a process of design modification is addressed. At that time, the performances of the original and modified ones are compared via an axode analysis. Besides, the screw pitch and global isotropy under the workspace are evaluated with the comparison between them. The ultimate design is constructed with a flexible hinge joint which leads to a minimal displacement of the joint.

Lastly, a general skeletal model describing the behavior of the radius and ulna during forearm rotation is proposed. Most researchers assume that the ulna is stationary in the kinematic analysis of the forearm. Only, a few researchers have tried to disclose the ulnar movement, which is described as a conical circumduction. They believe that this ulnar movement compensates a transposition and slight tilt of the hand. However, constructing a kinematic model to describe the ulnar movement is a challenge. Thus, a general forearm model is established and analyzed in this research. The proposed model is developed by starting from analyzing and modifying a conventional model. Ultimately, the proposed model can describe elliptical circumduction.

The aim of this thesis is to establish kinematic models of the human finger, wrist, and forearm, thus analyzing and synthesizing them to mimic their characteristic behaviors.