Collected molecules will appear here. Add from search or explore.
A classical motion planning stack for the CARLA simulator, handling lane centering, obstacle avoidance, and basic traffic rules like stop signs.
Defensibility
stars
53
forks
15
This project is a classic example of a curriculum-based implementation, likely originating from the Coursera 'Self-Driving Cars' Specialization (University of Toronto). With 53 stars and 15 forks over 6.7 years, and zero current velocity, it serves as a historical reference rather than a living codebase. The defensibility is near zero as the techniques used (likely basic PID controllers, state machines, and simple path planners) have been commoditized by more robust open-source stacks like Baidu Apollo or Autoware. Frontier labs and specialist AV companies (Wayve, Tesla, Waymo) have moved far beyond these rule-based heuristics toward end-to-end neural planners or more sophisticated MPC/Trajectory optimization. The project has high platform risk because it is tethered to an older version of the CARLA simulator, and the displacement horizon is '6 months' only in the sense that any modern researcher would choose a more recent and maintained framework immediately upon starting a new project.
TECH STACK
INTEGRATION
reference_implementation
READINESS