Analysis of Social Robotic Navigation approaches: CNN Encoder and Incremental Learning as an alternative to Deep Reinforcement Learning

TL;DR

This paper explores alternative approaches to deep reinforcement learning for social robotic navigation, focusing on CNN encoders and incremental learning to better handle social tasks involving humans.

Contribution

It analyzes the potential of CNN encoder and incremental learning methods as viable alternatives to deep reinforcement learning in social navigation tasks.

Findings

Incremental learning models can adapt to social navigation scenarios.

CNN encoders offer a promising approach for social robotic tasks.

Deep reinforcement learning faces challenges with human-in-the-loop scenarios.

Abstract

Dealing with social tasks in robotic scenarios is difficult, as having humans in the learning loop is incompatible with most of the state-of-the-art machine learning algorithms. This is the case when exploring Incremental learning models, in particular the ones involving reinforcement learning. In this work, we discuss this problem and possible solutions by analysing a previous study on adaptive convolutional encoders for a social navigation task.