Balancing Stability and Stiffness Through the Optimization of Parallel Compliance: Using Coupled Tendon Routing

Abstract

Having a wide range of achievable stiffness is essential for a robotic manipulator to robustly and safely interact with unknown environments. However, the achievable controller stiffness is fundamentally bounded by the system's passive stiffness, which introduces problems for compliant robots with series elasticity. Since strong passive stiffness is undesirable in uncertain environments, in this article, we introduce coupled tendon routing (CTR) along with nonlinear parallel compliance (NPC) to effectively shift this boundary with minimal change to the overall stiffness. We present a novel method for optimization and physical implementation to systematically determine the NPC, meet stability goals, and further reduce the overall stiffness through CTR. Experiments are carried out with two tendon-driven two degrees of freedom (2 DoF) planar fingers to demonstrate and validate the effectiveness of the proposed method.