Troy is Going Places with NeWheels
At 16 Troy Cordie left school to start a trade as a steel draftsperson, turning architectural and engineering drawings into fabrication plans, but a slow in work prompted a future-proofed career change.
Troy enrolled at QUT as a mature-aged student and found a love for robotics in his first year of undergraduate engineering studies.
Now, a PhD student, he is completing his thesis in modular field robotics working on NeWheel (pronounced ‘any wheel’) with CSIRO Data61, while under the supervision of Dr Tirthankar Bandyopadhyay and world-renowned QUT roboticist Professor Jonathan Roberts.
NeWheel modules are self-contained, two-degree-of-freedom robot wheels capable of individual operation or collaborative work.
Each wheel can achieve a desired velocity and steering angle using onboard motors, a battery and computing.
“My field of research is modular field robotics, specifically the creation of bespoke robots that make working with robotics simpler — so you don’t need to be an engineer,” Troy said.
“The goal is to make robots more accessible and customisable.”
Troy will have an operational NeWheel on display at Robotronica — QUT’s free robotics festival at the Gardens Point campus on 18 August, 9am — 4pm.
A modular field robot is created from a group of independently working modules as opposed to a robot designed for the specific task.
“Modular robotics allows us to customise the robot in the field based on the environment,” Troy said.
“The customisation would be limited to the number of available modules. We currently have four modules so we can make robots with up to four robotic wheels.
“Simulations have shown as long as we have computing power, we can add more wheels. So it is size and shape invariant — a modular robot can be any size or any shape.
“At this stage for practicalities, we are restricted to robots that need wheels roughly 150 millimetres in diameter, though the controller would be scalable to wheels of different sizes.”
Troy works specifically on the controller — the computer power that governs direction and speed. He investigates ways to join robot modules to allow them to communicate with each other and operate collaboratively in the field, as well as solutions that allow the platform to be reconfigured in the field.
“The controller is told how fast the robot needs to move then works out how fast each of the wheels should move and gives those separate commands,” Troy said.
“The wheels achieve that speed so that, as a whole, the robot moves as one. But really, individually, each wheel is moving on its own.
The key difference between a modular robot built by NeWheels and an all-wheel-drive car is that each wheel has its own motor, says Troy.
“The wheel speeds aren’t coupled allowing better terrain following. This is starting to appear in electric four-wheel drive concept cars.
“Also, we can change the shape and number of wheels. I’m yet to see a car that can do that.
“The big advantage is that this allows us to create custom robots to fulfil existing tasks, or it allows us to create custom robots in the field.
“If we arrive somewhere and the robot we had in mind doesn’t work — modules allow us to customise the robot quickly to overcome challenges.
“Or, it allows us, in some examples, to customise existing infrastructure — such as trollies.
“A person could push a trolley but, in theory, it could also be retrofitted with a set of wheels and turned into a robot without the need to redesign the trolley as a robot.
“You could connect the set of wheels and have the trolley drive itself.”
The NeWheel concept grew out of an interest in looking for unconventional modes of locomotion, and Troy would like to see an industrial NeWheel adopted.
“I think robotics has the potential to be really assistive.
“There are lots of jobs where people get injured pushing heavy trolleys and carrying heavy objects. So, I think robotics has the ability to remove people from dangerous jobs.”
Logistics within factory settings is one area of potential application but Troy says any task which puts a human in danger could be a candidate for modular robotics, including search and rescue, and extra-terrestrial missions.
“We’ve recently looked at using modular wheels for extra-terrestrial rover-type missions.
“Some of the inherent robustness that you get through modularity could be used to prolong the life of a rover through adapting its control and its mode of operation to failures within the robot.
“Each wheel has two motors. If we had a steering motor failure we could adapt the rest of the platform so none of its manoeuvres would require the failed wheel to steer.
“If there’s enough redundancy already within the system, you can potentially just abandon a module and adapt the controller to work with fewer wheels, which is something that a standard robot can’t do.”
In a team of five researchers, Troy says it took about 12 months and a few iterations of the hardware and controller to get to where they are now.
Troy started his PhD with CSIRO in 2017 and aims to finish in 2020, however, his relationship with CSIRO began during his undergraduate studies.
“I actually came to know about the robotics group at CSIRO in Brisbane through a field trip with one of the robotics units I did at QUT — Introduction to Robotics.
“I joined CSIRO for my vacation work and my PhD topic grew from the work I did as an undergrad.
Troy studied mechatronic engineering, a combination of electrical and mechanical engineering.
“As part of mechatronics we do electrical engineering, which helped with the design of the electronics and the selection of the motors.
“But, we also do mechanical engineering, which helps with the design of the physical wheel itself, and also with the controller that’s used to instruct the wheels where to go.
“There is quite a lot of overlap there with control from an electrical engineering point of view, but the fundamental mechanical equations of motion stem from mechanical engineering.”
Troy says there was also some overlap between his past and future careers, but he chose to start again from scratch.
“I was a steel detailer — the person who sits between an engineer and an architect designing buildings and turns their drawings into fabrication drawings to be constructed.
“That sort of work was slowing down in Australia and I wanted a career that would be fairly future proof — so I didn’t have to retrain again.
“So, I figured I’d do engineering. I came to QUT and I did my first year, which was a generic first year, and I decided I liked the idea of building robots.
“There’s a little bit of overlap with my old career, which included a lot of design and CAD work, and I guess a lot of maths as well. That sort of helped me a little bit. But other than that, it was pretty well starting over from scratch.
“It was quite daunting coming back as a mature-aged student. I hadn’t finished high school. I had gone and done a cadetship when I was 16, so it was daunting coming back — learning and studying again after so long.
“I got through it and did okay. It was hard but I’ve managed to pick something that I enjoy.
“I almost feel like I get paid to play with toys now. It definitely does feel like work some days but when it all works — the payoff is good.
