|
Path Planning Overview
Path planning is the basic process by which our vehicle decides on what path to take through the world. The A.I. uses the world model created by the sensors, the GPS waypoints provided during the race by DARPA, and a road following algorithm to pick a best path.
Road Following
The road following algorithm uses color differences, shadowing, and edge-detection algorithms to detect the sides of a road (if there is a road) and then decides if the road is turning, going straight, which direction, how sharply, etc. The road following algorithm uses input from most of the vehicle sensors, and provides the A.I. with probable road characteristics.
|
The red line represents the ideal path picked by our A.I., and the turquoise path represents the actual path traveled by our vehicle. The differences arise because 1) we did not start our vehicle on the ideal path, and 2) our vehicle must, without exceeding its performance limits, avoid small obstacles such as boulders.
|
