The evolution in humanoid robots, aiming for extensive human-robot co-existence and collaboration, calls for an enhancement in the humanoid manipulation dexterity. Dexterity, in simpler term, is the ability to perform a difficult task quickly and skillfully. When it comes to manipulation dexterity, the emphasis is often on design of anthropomorphic hands. However, this doctoral dissertation endeavors in enhancing the wrist and forearm dexterity from a mechanism design perspective. Light weight design, backdrivability and high payload-to-weight ratio form essential characteristics of a dexterous robotic arm. These objectives can be achieved by reducing inertial loads on the arm by relocating the distal actuators to the proximal links and transmitting power through parallel linkages or tendons. This thesis explores state-of-the-art parallel architectures for wrist implementations. It presents a detailed comparison analyses for promising candidate mechanisms considering their workspace parameters and mechanism isotropy. Based on this study, a two degrees of freedom parallel orientational mechanism is proposed for wrist flexion/extension and radial/ulnar deviation. This mechanism has hemispherical workspace, high range of motion, near to full isotropic behavior and a compact design; the ideal characteristics for a humanoid wrist. Alternatively, for tendon-based transmissions, one of the critical issues is to have appropriate tendon routing so as to avoid any kinematic couplings. To this effect, a novel constant length tendon routing mechanism through an axial joint is also proposed for forearm pronation/supination. This mechanism produces a decoupled full circle rotation while allowing simultaneous routing for four distal joint tendons. The concept idea, computer-aided design, prototyping and preliminary tests are demonstrated for both the wrist and forearm mechanisms that aim towards enhancing humanoid dexterity.
Design of Wrist and Forearm Mechanisms for Enhanced Humanoid Dexterity
SHAH, DIVYA HARESH
2021-06-18
Abstract
The evolution in humanoid robots, aiming for extensive human-robot co-existence and collaboration, calls for an enhancement in the humanoid manipulation dexterity. Dexterity, in simpler term, is the ability to perform a difficult task quickly and skillfully. When it comes to manipulation dexterity, the emphasis is often on design of anthropomorphic hands. However, this doctoral dissertation endeavors in enhancing the wrist and forearm dexterity from a mechanism design perspective. Light weight design, backdrivability and high payload-to-weight ratio form essential characteristics of a dexterous robotic arm. These objectives can be achieved by reducing inertial loads on the arm by relocating the distal actuators to the proximal links and transmitting power through parallel linkages or tendons. This thesis explores state-of-the-art parallel architectures for wrist implementations. It presents a detailed comparison analyses for promising candidate mechanisms considering their workspace parameters and mechanism isotropy. Based on this study, a two degrees of freedom parallel orientational mechanism is proposed for wrist flexion/extension and radial/ulnar deviation. This mechanism has hemispherical workspace, high range of motion, near to full isotropic behavior and a compact design; the ideal characteristics for a humanoid wrist. Alternatively, for tendon-based transmissions, one of the critical issues is to have appropriate tendon routing so as to avoid any kinematic couplings. To this effect, a novel constant length tendon routing mechanism through an axial joint is also proposed for forearm pronation/supination. This mechanism produces a decoupled full circle rotation while allowing simultaneous routing for four distal joint tendons. The concept idea, computer-aided design, prototyping and preliminary tests are demonstrated for both the wrist and forearm mechanisms that aim towards enhancing humanoid dexterity.File | Dimensione | Formato | |
---|---|---|---|
phdunige_4264493.pdf
accesso aperto
Tipologia:
Tesi di dottorato
Dimensione
19.34 MB
Formato
Adobe PDF
|
19.34 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.