Cable Driven Elephant Trunk-Inspired Robot: Analytical Derivatives and Computational Insights

Soft robots are becoming popular due to advantages such as flexibility, the capability to maneuver in complex environments, adaptability, and many more. The dynamic model of the robot and its solution speed is crucial from a control point of view for tasks such as path planning and optimization. In this work, we derived the dynamic model of a cable-driven elephant trunk-inspired articulated soft robot. The robot comprises smaller rigid links connected via universal joints with mechanical impedance. We further derived the dynamic model's analytical derivative, exploiting the Recursive Newton-Euler Algorithm (RNEA) algorithm via chain rule. The model is implemented in MATLAB and simulated using ode15s solver, including analytical derivatives. When implemented along with its analytical derivative, the dynamic model enhanced the prediction accuracy; however, the computation times varied depending on the system configuration. With more minor and slow decaying oscillations, the analytical derivative increased the computation time compared to numerical differentiation, yet at the cost of enhanced accuracy.