"The World Is Its Own Best Model": Robust Real-World Manipulation Through Online Behavior Selection

TL;DR

This paper introduces a probabilistic filtering approach for robotic manipulation that monitors environment states to dynamically select controllers, enhancing robustness against disturbances during tasks like opening drawers and grasping objects.

Contribution

It presents a novel online behavior selection framework based on environment state monitoring, improving robustness in real-world robotic manipulation tasks.

Findings

Robust drawer opening demonstrated in real-world tests

Effective environment state monitoring for controller selection

Enhanced task resilience to disturbances

Abstract

Robotic manipulation behavior should be robust to disturbances that violate high-level task-structure. Such robustness can be achieved by constantly monitoring the environment to observe the discrete high-level state of the task. This is possible because different phases of a task are characterized by different sensor patterns and by monitoring these patterns a robot can decide which controllers to execute in the moment. This relaxes assumptions about the temporal sequence of those controllers and makes behavior robust to unforeseen disturbances. We implement this idea as probabilistic filter over discrete states where each state is direcly associated with a controller. Based on this framework we present a robotic system that is able to open a drawer and grasp tennis balls from it in a surprisingly robust way.