// Personal Project · Leeds Formula Student AI team
Formula Student AI - Autonomous Racing Systems Development
Joined a team building an autonomous race car, working with ROS2, SLAM and path planning.
I joined the simulation team of Leeds University's autonomous racing team, which builds a self-driving car for the international Formula Student AI competition. The simulation team built the environments used to test the car's autonomous subsystems.
Formula Student AI: The Competition
Formula Student AI is a major international engineering competition in which university teams design, build and race self-driving formula-style cars. On top of the mechanical and electrical work of standard Formula Student, the car has to perceive the track, plan its path, control itself and make decisions in real time.
Competition Challenges
- • Dynamic events: acceleration, skidpad, autocross and endurance racing
- • Static events: engineering design, cost analysis and business presentations
- • Autonomy: the car has to drive itself the whole way
- • Safety: it has to do so safely and reliably
Leeds University FS-AI Team
- • Work: developing an autonomous racing system
- • Target: the international Formula Student AI competition
- • Make-up: students from several engineering disciplines working together
- • Aim: show the car can drive itself on track
My Role: Simulation & Autonomous Systems Exploration
Strategic Position
I joined the simulation team because it touches every part of the car. Simulation sits between perception, planning, localization and control, so the role was a way to see how all of those subsystems fit together, even though my own work was on the simulation side.
Primary Responsibility: Simulation
- • Proof of concept: show in simulation that the car could move and be controlled
- • Integration testing: check that the subsystems worked together before going onto hardware
- • Environment: simulation infrastructure built on ROS2
- • Safer testing: try autonomous behaviour in simulation before real-world tests
Broader Involvement Strategy
- • Other subsystems: the role was meant to give me exposure to perception, planning and control
- • The whole system: understanding how the parts fit together
- • Learning: picking up the autonomous subsystems one at a time
- • Teamwork: working alongside engineers from other disciplines
2024 Context: Building on Lessons Learned
Simulation mattered more than usual that year. At the previous competition the car had failed to move off the start line, so the team wanted to prove the autonomous systems worked in simulation before competition day.
Technical Infrastructure
Simulation Environment
Hardware Platform
Development Tools & Workflow
ROS2 Ecosystem
Modular nodes that pass messages to each other, spread across a distributed system.
Docker Containers
Reproducible environments with managed dependencies, the same for everyone on the team.
Version Control
Shared code, code review and tracking the project timeline.
Autonomous Systems Components
An autonomous racing car needs several subsystems working together in real time. The simulation work touched all of them:
Perception
- • Camera-based vision
- • LiDAR point cloud processing
- • Sensor fusion
- • Object detection and tracking
- • Track boundary recognition
Planning & Control
- • Path planning
- • Trajectory optimization
- • Vehicle dynamics modeling
- • Steering and throttle control
- • Real-time decision making
Localization
- • SLAM (Simultaneous Localization and Mapping)
- • GPS/IMU sensor integration
- • Position estimation
- • Map generation and tracking
- • Real-time pose updates
Project Timeline & Development Approach
Timeline Structure
Development Philosophy
- • Gantt chart: a timeline with clear milestones
- • Phases: build and validate in small steps
- • Meetings: regular team meetings to work through problems together
- • Testing: repeated validation in simulation
Progressive Contribution Strategy
The plan was for my involvement to widen from simulation into perception, planning and control as the project moved through its phases. I left the team before that happened, so my work stayed on the simulation side.
Why Formula Student AI
Motivation
- • Outside coursework: I wanted real engineering work beyond my degree
- • A big team: a shared project across several engineering disciplines
- • The technology: autonomous vehicles, robotics and AI systems
- • The competition: the chance to be part of an international engineering competition
Learning Objectives
- • Integration: how autonomous subsystems interact
- • ROS2: modern robotics middleware and distributed systems
- • Autonomous vehicles: the basics of perception, planning and control
- • Working across teams: coordinating with mechanical, electrical and software engineers
Key Aspects
- • An international autonomous racing competition run by multidisciplinary teams
- • Testing autonomous systems in simulation so they can be developed safely
- • A view of perception, planning and control from the simulation side
- • ROS2, Docker and autonomous vehicle technology used on a real car
- • Team engineering with a structured project plan
Technology Stack
What I Took From It
I was a member of the Leeds University Formula Student AI team, on the simulation side, and have since left. It gave me a first look at how an autonomous racing car is built and how a large team from different engineering disciplines works towards an international competition.