By Emma Young & Gabrielle Leslie
June 14th 2022
We’ve been keeping a pretty big secret from you, but thought we’d finally let you in on it. For the past three years, Team Swinburne has been developing a driverless vehicle for FSAE-A’s new Driverless Vehicle division. That’s right, we plan to bring an autonomous driven vehicle of our own to competition at the end of this year.
With the project well and truly underway, we thought we’d let you get to know Gabrielle, Lead Engineer for the Autonomous Vehicle in 2022, and have her tell you where we’re at with our progress.
Who are you, what are you studying and how long have you been with the team?
So, I’m Gabrielle. I am 23. I’m in my final year of engineering Mechatronics Engineering in a double degree. I study Bachelor of Engineering (Honours) and a Bachelor of Innovation & Design with a major in Robotics and Mechatronics.
I have been in the team just over a year now, but I’ve enrolled at Swinburne since 2017. Team Swinburne is my first and last University club.
Just so everyone knows, what exactly is Autonomous?
We call it autonomous, but in general, stemming from Formula Student Germany it is called the driverless competition, which, as it is named, there’s no driver controlling the vehicle. This means there is a lot more involvement in selecting and implementing components, especially in the power and electrical architecture and the mechanical components like the steering rack or the braking module, which are directed by software, informed by optical sensors, mounted to the vehicle.
Where are we currently with our own Autonomous Vehicle development?
Currently, we are using a Velodyne LiDAR, it’s a light detection and ranging camera, utilising reflection off of objects by light to create a model based on depth of field of what’s around it, whilst used in conjunction with using an electro optical camera. We transfer the data collected by the optical sensors into an ECU, an electrical computing unit, which processes, collates, and filters, all that data in one place. Then it’s put through an algorithm or a few, which does with the data what we need to control the vehicle, ideally, identifying something like a cone a certain coloured cone, and creating a path through it. Colour cones at competition mean a direction that the vehicle turns in and so when that is fed back into the ECU, the ECU then pushes a command out to through the power and electronics.
We have designed between a level two and three autonomous system, which, according to SAE (Society of Automotive Engineers), means that it’s capable of operating on its own, but a driver can still drive it as necessary. A significant part of the technical scrutineering that we complete at competition and that’s also how the car learns.
That’s about where we are in terms of our progress. Software is probably the furthest along, it has object identification and path planning algorithms well developed , done by final year research students in 2022, guided by our previous Autonomous Vehicle Lead Engineer, Thomas, who invested a great deal of time into making the sensors work with those algorithms.
How did you find picking up as Autonomous Leader this year?
It was difficult, the knowledge gap was great. It would have been easier to go into it knowing more, but you’re provided the information that the previous leader sees fit that you know, and though you may not feel, once you’ve read that information, that’s what you need, in hindsight, it’s not a bad thing that you have to pick up the slack and force yourself to learn these things by getting into the workshop, troubleshooting, and filling in those gaps yourself.
Overall, I think I have changed the approach of autonomous vehicle’s development and I’ve changed hopefully the team’s attitude, which is to value the autonomous vehicle almost as much as the driver vehicle. The all-wheel drive has been a few years in the making, so it is imperative that it goes to competition whereas autonomous has always been deemed as a research project. But it’s now in the development phase and luckily, I have some really fantastic section leads, who have been nothing but eager, keen to be involved in working on both cars and intelligent enough to do that and be capable of working on them both, whilst managing university and work. But yeah, changing that attitude has been my biggest priority. Because we can all learn more, and it’s about what I impart on my members which is most important to me.
One final question. What are your short- and long-term goals for the autonomous vehicle?
Our short time goal is to get to competition and do a demonstrative track drive. It’s not being scored this year, so the expectations are very different to that of the electric driver vehicle.
I think Luke, Chief Engineer, and I have this very nice little dream that we will, at the end of the year, before we get to competition, have two cars sitting there ready to go and we will be the first team or first, you know, team at Swinburne to do that. I think people won’t expect it to happen, but Luke and I sort of thrive off of that doubt and really want to have those cars running, that’s most important.
I want it to go around the circuit, pick up some sort of speed, but I also think that we can learn more from testing autonomous this year than just taking it to competition. It’s why we have the timeline is the way it is, so that we have that time for testing. Because testing is most important for an autonomous system because it has to have a diversity of environments and testing situations in order to learn. And it’s not going to compete at competition unless it experiences those environments.
Long term, after next year, past teams and management have always had the intention of combining the driver and driverless vehicles, into a dual capable vehicle, as have European Formula Student teams have. That’s definitely the intention and we have set up a lot of our architecture to attempt to implement this in future, but this depends on how FSAE-A develop the Driverless competition in the coming years.