Internal Monthly QCR Newsletter April 2023

View this email in your browser

What's Inside

Directors Message

Centre News

Member Achievements

Feature Story

Events - Past and Coming Up

New Faces

Member Publications

Feature Story

Directors Message

Another action-packed newsletter with plenty of new faces, accolades and achievements for centre members, and ongoing activities within the centre including the now almost institutional board games night – great to see.

I have just returned from my first extensive tour of the United States in four years, using the invite to give a keynote at the 2023 Neuro-Inspired Computing Elements as an excuse to catch up with a lot of colleagues, organisations and meet new ones. I visited 7 universities, gave 6 talks, caught up with researchers at global tech orgs, national research labs, and exciting startups, and saw more autonomous vehicles driving around Californian streets and university campuses delivering coffee.

This was my fourth recent trip overseas after a long hiatus due to the pandemic, and has served a couple of key purposes. The first is to continue to showcase the QUT Centre for Robotics: while we have still have good profile internationally it’s vitally important to be “in the room” so to speak on a regular basis, and trips like this, as well as the centre’s representation at major upcoming conferences like ICRA2023 in London, are vital.

The second purpose is to get a snapshot of the “lay of the land” so to speak in terms of what is happening on a global scale, and where we fit in that landscape. I’m able to report that everyone is as bemused by new developments in the generative AI space as we are, and everyone is scrambling to try to respond: from leading tech companies to individual research labs. It has also become clear that our signature offering in this global landscape as a centre is the fact we are still a force in terms of high impact fundamental research at the highest levels, but also engage in the nitty gritty of translating this research into actual deployments, tech and policy input with stakeholders from industry and government.

Whilst researchers in Australia may complain on occasion about how hard it is to get funding from industry, it does give us one superpower: compared to many of our peers, we are much more experienced at working with industry partners over long periods of time, deeply understanding their challenges and goals, and working up appropriate solutions that don’t just consider the core technological challenges but the wider implications of the solution, around acceptance, societal impact and other factors. If we don’t do this we don’t get work - industry typically doesn’t just turn up and throw money at us. This is a hallmark of what we do, and something we should be proud of, and leverage and promote relentlessly when going after new opportunities.

In other news: the centre is working on its submission to the National Robotics Strategy Discussion Paper – individuals are also encouraged to submit proposals, the more the merrier. We’re preparing for our first external review in June, and continuing to be active in the AI space at a federal level including being part of the Kingston AI group. We also have our annual mid-year offsite retreat coming up, which will be a great event.

Forward

Centre News

Australian Space Awards

QUT finalist of the Australian Space Awards 2023 in the category: Academic Institute of the Year

Group nomination based on multiple pieces of work including:

David Flannery & team’s work on Mars Perseverance mission
Peter & team’s Space Logistics ASA demonstrator feasibility project
The Moon Yard / Lunar Testbed project

Member Achievements

PhD

Dr Krishan Rana has been awarded his PhD

Topic: Leveraging Control Priors in Deep Reinforcement Learning for Robotics

New Baby

Baby Freddie (Frederik Lamont) was born 4 April. Everybody’s happy and healthy.
Congrats to Riki and his wife!

Feature Story

Top Ranking for Peter Corke

Peter Corke has been ranked in the top 10 best computer science scientists in the country

>READ MORE

Events

Upcoming Events

Robotic Vision Scene Understanding Challenge

The Robotic Vision Scene Understanding Challenge evaluates how well a robotic vision system can understand the semantic and geometric aspects of its environment. There are two tasks in this challenge: Object-based Semantic Mapping/SLAM and Scene Change Detection.

2023 RVSU Challenge running in CVPR 2023 Embodied AI workshop

Submissions by end of May 2023

Direct tech support via David Hall: david.hall@csiro.au

>READ MORE

Upcoming Events

Fulbright Scholarships Information Session – QUT

Friday 5 May 2023 12:00 PM

Fulbright Scholarships create life-changing opportunities for students and researchers, offering an unparalleled experience of U.S. academic and cultural exchange built upon a 75-year foundation of bilateral collaboration, knowledge transfer, and diplomacy.

This information session will cover Fulbright Scholarship opportunities for Australians who are seeking funding to undertake a master’s degree, PhD research, postdoctoral research or short-term research program in the U.S.A in 2024/25.

After registering, you will receive a confirmation email containing information about joining the webinar.

Past Events

QCR Board Games Night

Please contact our HDR Reps if you'd like to be invited to future HDR Events:

Connor Malone - connor.malone@hdr.qut.edu.au
Jordan Gleeson - jordan.gleeson@hdr.qut.edu.au

New Faces

Welcome to Gokul Nair

A warm welcome to Gokul Nair, who has started his PhD with Tobi and Michael on the Intel-sponsored project “Active, adaptive and predictive robot localisation”.

Gokul joins us with a BTech + M.S. by Research from IIIT-Hyderabad, and work experience at Addverb Technologies.

Welcome to Alicia Mitchell

We recently had Alicia start with us as a PhD student under the supervision of Laurianne Sitbon.

Alicia specialises in human-robot interaction (HRI) and her current research focuses on designing social robots to support people with intellectual disabilities in work environments.

Welcome to Adam Tow

A warm welcome to Adam Tow, who rejoins QUT as a Research Engineer in the REF Robotics & Autonomous Systems team.

Adam joins us from Dorabot, where he most recently was the Lead Robotics Engineer working on their technology stack for improving smart warehouse, distribution and fulfillment operations. This experience has been built upon his time as a PhD student within the ACRV, where he was a member of the winning team at the 2017 Amazon Robotics Challenge.

Welcome to Andrew Lui

A warm welcome to Dr Andrew Lui, who joins QUT as a Senior Research Engineer in the REF Robotics & Autonomous Systems team.

Andrew joins us most recently from Hong Kong Metropolitan University, where he is currently an Honorary Professor, having previously held roles as Associate Dean (School of Science and Technology), and Head of Department (Technology). With over 20 years of programming experience incorporating C/C++/C*/Python/Java/Matlab/PHP/SQL, as well as experience in the training and use of deep learning models, Andrew will provide additional depth to the software focused elements of REF.

Behind the Scenes

What do you do in the Centre and what are you currently working on?

Within the Centre I have a few roles including PhD student, being one of three HDR student reps, as well as also working as a research assistant. Currently my PhD work focuses on the use of computer vision in autonomous vehicles for scene understanding. Specifically, how image/video data can be used with visual place recognition to identify places a vehicle has previously visited which are similar to its current location and how this information can be used to improve its interpretation and understanding of its current environment. My research assistant work, which is also based around visual place recognition, looks at implementing these types of localisation techniques on a physical robot/vehicle and building a system to monitor the integrity and explainability of their results. As a HDR rep, I try to help students with any issues that pop up, organise social and professional development sessions/events for students and try to make sure they are having a pleasant experience in the Centre. Currently I am helping to organise a philosophy night for QCR, a HDR showcase event to demonstrate our student’s research to industry and our continuing board game nights!

What’s on your desk?
LOTS of previously read or yet to read papers along with old drafts from a personal paper, a notepad full of scribbles, water bottle, mug, wireless phone charger and a little Lego Pokémon figurine (Charizard)

How would you describe your work to a child?
I use images to tell a self-driving car both where it is, like a GPS, and where the road, trees and buildings are around it so it knows where to drive.

Three words to describe your work?
Challenging, practical, exciting

What’s your dream travel destination?
Fiction category: Earth’s parallel in the Pokémon universe, the Kanto region
Non-Fiction category: Scotland

Please finish these sentences:

I’m currently reading: ‘Toward accurate visual localization with continuous place-descriptor regression’ – or ‘Cobalt Blue’ by Matthew Reilly depending on the time of day

And currently wishing: To finish off this PhD and join the real world!

My last meal would be: A sizeable portion of Mum’s apricot chicken, Milo on ice-cream for dessert and a bottle of Ardbeg scotch (better make it the 25 year if its my last meal)

Final question! Tea, coffee or no caffeine? Tea of course!

Member Publications

PhD thesis

Leveraging control priors in deep reinforcement learning for robotics

K Rana - 2023

https://eprints.qut.edu.au/238632/1/Krishan_Rana_Thesis.pdf

Papers

Exactly Optimal Quickest Change Detection of Markov Chains
JJ Ford, JM Kennedy, C Tompkins, J James, A McFadyen - arXiv preprint arXiv …, 2023

https://arxiv.org/pdf/2303.13778.pdf

Autonomous Detection of Mouse-Ear Hawkweed Using Drones, Multispectral Imagery and Supervised Machine Learning

N Amarasingam, M Hamilton, JE Kelly, L Zheng, J Sandino, F Gonzalez, R Dehaan, H Cherry - Remote Sensing, 2023

https://www.mdpi.com/2072-4292/15/6/1633

A-MuSIC: An Adaptive Ensemble System For Visual Place Recognition In Changing Environments

B Arcanjo, B Ferrarini, M Milford, KD McDonald-Maier, S Ehsan - arXiv preprint arXiv …, 2023

https://arxiv.org/pdf/2303.14247.pdf

Addressing the Challenges of Open-World Object Detection

D Pershouse, F Dayoub, D Miller, N Sünderhauf - arXiv preprint arXiv:2303.14930, 2023

https://arxiv.org/pdf/2303.14930.pdf

The Need for Inherently Privacy-Preserving Vision in Trustworthy Autonomous Systems

AK Taras, N Suenderhauf, P Corke, DG Dansereau - arXiv preprint arXiv:2303.16408, 2023

https://arxiv.org/pdf/2303.16408.pdf

Motivation – when Optimisation Fails