| Definition | A simulation benchmark for autonomous racing with large-scale human data |
| No of environments | 109 |
| No of transitions | ~2.3M |
Despite the availability of international prize-money competitions, scaled vehicles, and simulation environments, research on autonomous racing and the control of sports cars operating close to the limit of handling has been limited by the high costs of vehicle acquisition and management, as well as the limited physics accuracy of open-source simulators. Assetto Corsa is a racing simulation platform based on the simulator to test, validate, and benchmark autonomous driving algorithms, including reinforcement learning (RL) and classical Model Predictive Control (MPC), in realistic and challenging scenarios. The dataset is collected from human drivers.
There are four tracks and three cars used in the dataset. The tracks include Indianapolis (IND), an easy oval track; Barcelona (BRN), featuring 14 distinct corners; Austria (RBR), a balanced track with technical turns and high-speed straights; and Monza (MNZ), the most challenging track with high-speed sections and complex chicanes. The cars are the Mazda Miata NA (Miata) with a top speed of 197 km/h, the Dallara F317 (F317) with a top speed of 250 km/h, and the BMW Z4 GT3 (GT3) with a top speed of 280 km/h. This diverse array ensures a comprehensive dataset for evaluating driving algorithms.
An environment corresponds to a human driver, a car and a track. Data are sampled at 25Hz.
| Example observations | Car and wheels speed Car and wheels acceleration Slip angle RPM Grip |
| Example actions | Steer angle Acceleration status Brake status |
| Example reward | Trajectory |