We have implemented a robotic path planner using a fuzzy controller. Traditionally, controllers are implemented using PID Controllers and are a real pain to calibrate. A fuzzy controller is more intuitive to program, since it uses a mapping table that explores the human knowledge.
For example, this could be a fuzzy rule in a fuzzy controller:
IF distance from car in front of you IS small AND speed IS very fast
THEN brake pressure IS very high.
Traditional fuzzy logic (fuzzy sets, etc) are employed to define the ranges for things like "very fast" and "small" and the corresponding operations (AND, OR, etc).
From my experience, 90% of time a traditional and proven PID controller will be enough, simple and reliable.