valigatotus / lidar-fsae Goto Github PK

Following could be implemented to improve the driving logic:

• PID control
• Obstacle avoidance
• Path planning algo
• Using a higher-level controller
• Testing and tuning parameters

Maybe doing some kind of combination of clustering and circle fitting in order recognize cones ?

• first clustering scan
• then circle fitting on clusters to detect if it is a cone or not (algos: RANSAC, Least Squared Method or ...)

make the blocking processes (see: scan, think: clustering & target definition, act: robot steering (turn & move)) non-blocking and working concurrently.