Autonomous Learning of New Environments with a Robotic Team Employing Hyper-Spectral Remote Sensing, Comprehensive In-Situ Sensing and Machine Learning.

David J Lary,John Waczak,Bharana Fernando,Tatiana Lary,David Schaefer,Lakitha O H Wijeratne,Aaron Barbosa,Matthew D Lary,John Sadler,Adam Aker,Shawhin Talebi

doi:10.3390/s21062240

David J Lary, John Waczak + Show 9 more

Open Access

https://doi.org/10.3390/s21062240

Copy DOI

Abstract

This paper describes and demonstrates an autonomous robotic team that can rapidly learn the characteristics of environments that it has never seen before. The flexible paradigm is easily scalable to multi-robot, multi-sensor autonomous teams, and it is relevant to satellite calibration/validation and the creation of new remote sensing data products. A case study is described for the rapid characterisation of the aquatic environment, over a period of just a few minutes we acquired thousands of training data points. This training data allowed for our machine learning algorithms to rapidly learn by example and provide wide area maps of the composition of the environment. Along side these larger autonomous robots two smaller robots that can be deployed by a single individual were also deployed (a walking robot and a robotic hover-board), observing significant small scale spatial variability.

Highlights

IntroductionThe system that is described here demonstrates a flexible paradigm that is scalable to multi-robot, multi-sensor autonomous teams
This paper describes a robotic team that can rapidly learn new environments
A loop is executed over all the variables that were measured by the robotic that we would like to estimate using the hyper-spectral imagery

Summary

Introduction

The system that is described here demonstrates a flexible paradigm that is scalable to multi-robot, multi-sensor autonomous teams. The inspiration for this autonomous robotic is the automation of what is currently done manually in the production of remote sensing satellite data products. A key part of this substantial time delay is due to the time that is taken for the collection of the relevant training data. Our goal was to reduce this timescale to be near real time by utilising an autonomous robotic team that can both collect the training data, and in real time process and stream the remote sensing data products

Objectives

Methods

Results

Discussion

Conclusion