Wiggle and Go! System Identification for Zero-Shot Dynamic Rope Manipulation
This work addresses the problem of dynamic rope manipulation for robotic tasks, offering a zero-shot solution that avoids large real-world datasets or iterative trial-and-error, which is significant for efficient and reliable task execution.
Wiggle and Go! introduces a two-stage system identification framework that enables zero-shot dynamic rope manipulation by learning simulation priors to predict physical parameters, achieving 3.55 cm average accuracy on 3D target striking (vs. 15.34 cm without system parameters) and a 0.95 Pearson correlation between predicted and real rope frequencies.
Many robotic tasks are unforgiving; a single mistake in a dynamic throw can lead to unacceptable delays or unrecoverable failure. To mitigate this, we present a novel approach that leverages learned simulation priors to inform goal-conditioned dynamic manipulation of ropes for efficient and accurate task execution. Related methods for dynamic rope manipulation either require large real-world datasets to estimate rope behavior or the use of iterative improvements on attempts at the task for goal completion. We introduce Wiggle and Go!, a system-identification, two-stage framework that enables zero-shot task rope manipulation. The framework consists of a system identification module that observes rope movement to predict descriptive physical parameters, which then informs an optimization method for goal-conditioned action prediction for the robot to execute zero-shot in the real. Our method achieves strong performance across multiple dynamic manipulation tasks enabled by the same task-agnostic system identification module which offers seamless switching between different manipulation tasks, allowing a single model to support a diverse array of manipulation policies. We achieve a 3.55 cm average accuracy on 3D target striking in real using rope system parameters in comparison to 15.34 cm accuracy when our task model is not system-parameter-informed. We achieve a Pearson correlation coefficient of 0.95 between Fourier frequencies of the predicted and real ropes on an unseen trajectory. Project website please see https://wiggleandgo.github.io/