Real-time Mapping of Physical Scene Properties with an Autonomous Robot Experimenter

TL;DR

This paper demonstrates that an autonomous robot can efficiently map physical scene properties in real-time by actively interacting with the environment, using entropy-guided sparse interactions to achieve dense annotations quickly.

Contribution

It introduces a method for robots to autonomously and efficiently map diverse physical scene properties through active, entropy-guided interactions during scene scanning.

Findings

Dense scene property maps generated in minutes

Sparse interactions guided by entropy improve efficiency

Multiple physical properties can be mapped simultaneously

Abstract

Neural fields can be trained from scratch to represent the shape and appearance of 3D scenes efficiently. It has also been shown that they can densely map correlated properties such as semantics, via sparse interactions from a human labeller. In this work, we show that a robot can densely annotate a scene with arbitrary discrete or continuous physical properties via its own fully-autonomous experimental interactions, as it simultaneously scans and maps it with an RGB-D camera. A variety of scene interactions are possible, including poking with force sensing to determine rigidity, measuring local material type with single-pixel spectroscopy or predicting force distributions by pushing. Sparse experimental interactions are guided by entropy to enable high efficiency, with tabletop scene properties densely mapped from scratch in a few minutes from a few tens of interactions.