King's students build, code and race self-driving cars at the 11th F1TENTH Autonomous Grand Prix

Part of the 2023 IEEE International Conference on Robotics in London, the race was an opportunity for the King’s team to explore the latest innovations in robotics and automation, as well as a chance to put what they had learnt in the classroom into action.

First founded in 2016, the F1TENTH Autonomous Grand Prix is a set of regional and international gatherings providing more than 20 groups across the USA, Europe and Asia with an accessible open-source platform on which to develop algorithms for self-driving cars. The organisers provide 150 pieces of hardware built on top of a remote control car and a software programme to all participating teams; leaving them to battle it out with their code to see who can build the most successful autonomous vehicle.

Post-doctoral researcher Milad Kazemi Mehrabadi, PhD student Tom Kuipers and final year Computer Science undergraduate Karina Nazarova, worked on creating a racer for F1TENTH for over two months. For them, working on this project has been a great way to break free from the siloes of standard approaches to computer science and engineering education, while also providing an opportunity for hands-on learning experience with industry standard robotic frameworks outside the classroom.

Autonomous vehicles by their nature combine a number of disciplines across robotics and engineering, with the team working across each to ensure the success of their car. The autonomous vehicle the students used work using LiDAR sensors, light detection and ranging sensors, that operate in tandem with an on-board AI-enabled computer to build a virtual map of the track. This allows the students to create optimisation algorithms to help find the optimal racing path and cross the finish line first.

Alongside LiDAR and AI, groups must also use automatic planning and control techniques that processes live data to constantly update the car’s path by predicting the future behaviour of other racer’s vehicles. Students then utilise simulation software to test the code they have created, faithfully recreating how their autonomous vehicle will run before it even gets out on the track.

As founder of the Grand Prix, Dr Rahul Mangharam, explains “The problem with engineering education today is that it’s fragmented. You study LiDAR one day and simulation software the next, but you never make the connection between the two. If you can’t make the connections between these different fields of study, you can’t be a good engineer.

F1TENTH empowers students to see how these different disciplines are connected and how they all work together in an autonomous vehicle. This a platform where students are free to experiment and fail without the consequences of crashing a full-scale self-driving car; where they can bridge the gap between the theory they learn in the classroom and the practical skills they need in industry”.

Between 50 to 60 research papers a year are written a year by the 80+ universities using F1TENTH to demonstrate theoretical work, and Milad and Tom will be finalising a research paper over the summer using their experience with the event as a case study. The team hope to be out on the circuit at the next F1TENTH, with Tom saying “As soon as we get back to King’s, we’ll be rolling out the test track and trying again.”

In addition to participating, the team from King’s also played an integral part in bringing the event to life in the heart of London, planning logistics for the three days and laying the track for the vehicles to race on.

Dr Nicola Paoletti, Senior Lecturer in Computer Science, Department of Informatics, who supported the student team from King’s, said:

“F1tenth is a great platform for learning about robotics, AI and autonomous systems. Alongside helping to instill some of the fundamentals of these disciplines, it’s also a great way to develop strong collaboration skills and team spirit. From fine-tuning hardware to working out the car’s racing strategy, each member must play their part, trouble shooting and making key decisions.”

He also congratulated the student team:

“The merit goes to Karina, Tom and Milad who did most of the hard (but fun) part of the job. Moving forward, I hope we’ll get more students involved in this project to help us to develop more advanced algorithms. Of course, we’d also love to participate in future editions of the F1TENTH Autonomous Grand Prix as well”.

Three undergraduate students will be carrying out F1TENTH related research throughout the course of the summer as part of the King’s Undergraduate Research Fellowship, looking at the cross-section of robotics and computer science. To find out more about F1TENTH and keep up to date with their latest news, check out their website.