Optimizing the Topology of Tendon-Driven Fingers: Rationale, Predictions and Implementation

Tendon-driven mechanisms in general, and tendon-driven fingers in particular, are ubiquitous in nature, and are an important class of bio-inspired mechatronic systems. However, the mechanical complexity of tendon-driven systems has hindered our understanding of biological systems, and the optimization of the design, performance, control, and construction of mechatronic systems. Here we apply our recently-developed analytical approach to tendon-driven systems [1] to describe a novel, systematic approach to analyze and optimize the routing of tendons for force-production capabilities of a reconfigurable 3D tendon-driven finger. Our results show that these capabilities could be increased by up to 277% by rerouting tendons, and up to 82% by changing specific pulley sizes for specific routings. In addition, we validate these large gains in performance experimentally. The experimental results for 6 implemented tendon routings correlated very highly with theoretical predictions with an R value of 0.987, and the average effect of unmodeled friction decreased performance an average of 12%. We not only show that, as expected, functional performance can be highly sensitive to tendon routing and pulley size, but also that informed design of fingers with fewer tendons can exceed the performance of some fingers with more tendons. This now enables the systematic simplification and/or optimization of the design and construction of novel robotic/prosthetic fingers. Lastly, this design and analysis approach can now be used to model complex biological systems such as the human hand to understand the synergistic nature of anatomical structure and neural control.