Vine Robot Localization Via Collision

Abstract

Localization is critical for robots performing tasks in unstructured environments, yet traditional methods struggle with occlusions and sensor limitations. This paper introduces a novel localization strategy for Vine Robots, leveraging contact-based sensing to estimate the tip position via collision detection. Single- and multi-obstacle experiments demonstrated less than 5% error relative to the robot’s length, establishing the approach as a robust method for soft robot localization.