RoboRacer
Formerly F1TENTH. Teams build a small autonomous race car to a reference spec and write software to avoid crashes, cut lap times and overtake. Used at dozens of universities, with an annual Grand Prix.
- Mapping
- Time trials
- Head-to-head
We race to find what breaks.
A race gives a lap time and a finishing position, so everyone can see whether the software got better.
Competitions we're targeting
We plan to climb in scale, from 1:10 cars to full-size race cars, entering the series where university teams race autonomy against each other.
Autonomous Karting Series
A collegiate series where university teams design, build and race fully autonomous electric go-karts, including the AKS Purdue Grand Prix.
- Vehicle build
- Perception at speed
- Racing
Formula Student Driverless
The driverless class of Formula Student, where student-built race cars run dynamic events on their own around a coned track.
- Cone detection
- SLAM
- Dynamic events
Indy Autonomous Challenge
University teams program fully autonomous race cars and compete at major tracks. In September 2026 it ran a head-to-head passing competition at Laguna Seca.
- High-speed control
- Passing
- Road courses
A2RL
The Abu Dhabi Autonomous Racing League, run by ASPIRE. It races fully autonomous cars at Yas Marina Circuit and has expanded to Europe.
- Multi-car racing
- Strategy
- Endurance
One lap, six tests.
Racing pushes every part of the autonomy stack at once. Each sector of a lap leans on a different one.
- S1
Perception at speed
Seeing the track and other cars clearly while moving fast.
- S2
Localization under pressure
Knowing exactly where the car is at racing pace.
- S3
Opponent prediction
Guessing what the other cars will do next.
- S4
Real-time planning
Choosing a racing line many times a second.
- S5
Aggressive control
Driving at the edge of grip without losing it.
- S6
Low-latency inference
Running models on the car with almost no delay.
How a race is scored
Most series run some version of these three formats. Each one stresses a different part of the stack.
01
Time trial
One car, an empty track, the fastest clean lap. Tests localization and how hard the controller can push.
02
Head-to-head
Two or more cars on track at once. Tests opponent detection, prediction and safe racing lines.
03
Passing
Overtake at speed without contact. Tests planning and control at the edge of grip.
What you'd work on
Racing has room for software and hardware people. Most members own one of these.
Perception
SoftwareDetecting the track, cones and other cars from lidar and cameras at speed.
Localization and mapping
SoftwareBuilding a map of the track and knowing where the car is on it, many times a second.
Planning
SoftwareRacing lines, overtakes and strategy against opponents.
Controls
SoftwareSteering and throttle that hold the line at the limit of grip.
Vehicle and electronics
HardwareBuilding and maintaining the cars, sensors, compute and wiring.
Simulation
SoftwareVirtual tracks for testing every change before it touches a car.
The objective isn't participation.
We race to win.
Faster lap times are the clearest proof that a perception or planning change worked.