F1Tenth
Autonomous racing for 1/10 model cars
January 2023 — May 2023
★ Achieved first place in the reactive methods race using Follow The Gap
★ Achieved third place in the indoor map-based race
The car is equipped with a 2D LiDAR, VESC and Jetson NX as the compute platform.
Follow The Gap
This algorithm is classified as a reactive driving method, as it identifies the widest gap in the upcoming path and moves in that direction. The gap is found by filtering and thresholding the LiDAR readings.
Map based racing
Mapping
A package was used to map the track in th corridor.
Localization
Particle filter was used to localize using just the LiDAR readings.
Path Planning
Waypoints were chosen either by driving the car manually around the track or byhand selecting the desired race line from the generated map.
Race line optimization
The waypoints generated, as mentioned in th above section, are optimzed to achieve the fastest time around the track. We used the TUM repository to generate the waypoints by tuning the parameters for this specific car and track.
Pure Pursuit
The control algorithm used to track the waypoints was a proportional controller using Pure Pursuit.