Learning Autonomous Mobility Using Real Demonstration Data

TL;DR

This paper introduces an LSTM-based learning framework that derives robust feedback control policies from human demonstrations, enabling autonomous obstacle negotiation, stair climbing, and delivery tasks on a tracked robot.

Contribution

It presents a novel architecture that learns reactive control policies from limited real demonstrations, effectively handling complex tasks and non-optimal data.

Findings

Successfully learned obstacle negotiation and stair climbing policies

Demonstrated robustness to real-world uncertainties like slippage

Efficiently handled non-optimal demonstrations and learned new skills

Abstract

This work proposed an efficient learning-based framework to learn feedback control policies from human teleoperated demonstrations, which achieved obstacle negotiation, staircase traversal, slipping control and parcel delivery for a tracked robot. Due to uncertainties in real-world scenarios, eg obstacle and slippage, closed-loop feedback control plays an important role in improving robustness and resilience, but the control laws are difficult to program manually for achieving autonomous behaviours. We formulated an architecture based on a long-short-term-memory (LSTM) neural network, which effectively learn reactive control policies from human demonstrations. Using datasets from a few real demonstrations, our algorithm can directly learn successful policies, including obstacle-negotiation, stair-climbing and delivery, fall recovery and corrective control of slippage. We proposed decomposition of complex robot actions to reduce the difficulty of learning the long-term dependencies. Furthermore, we proposed a method to efficiently handle non-optimal demos and to learn new skills, since collecting enough demonstration can be time-consuming and sometimes very difficult on a real robotic system.