Learning Relative Interactions through Imitation

TL;DR

This paper presents a neural network trained via imitation learning to enable a robot to perform specific and arbitrary interactions with objects, focusing on pose control and addressing sensor ambiguity issues.

Contribution

The work introduces a neural network approach for robot-object interactions that can handle fixed and arbitrary poses, highlighting the impact of sensor ambiguities.

Findings

High accuracy on fixed-pose tasks with limited data

Challenges in arbitrary-pose performance

Sensor ambiguities affect learned behaviors

Abstract

In this project we trained a neural network to perform specific interactions between a robot and objects in the environment, through imitation learning. In particular, we tackle the task of moving the robot to a fixed pose with respect to a certain object and later extend our method to handle any arbitrary pose around this object. We show that a simple network, with relatively little training data, is able to reach very good performance on the fixed-pose task, while more work is needed to perform the arbitrary-pose task satisfactorily. We also explore the effect of ambiguities in the sensor readings, in particular caused by symmetries in the target object, on the behaviour of the learned controller.