ROBOT IN A RECONFIGURABLE MAZE

Autonomous vehicles have existed since several decades and they continue to evolve
promptly. Their usage in different domains made them an interesting area for
academic researchers as well as governments’ projects. Issues that are still holding
back the development of autonomous vehicles are the accurate mapping and
localization of the surrounding that enable these vehicles to perform independently
in a precise manner.
Using either the left-wall following or right-wall following algorithm alone will
sometimes result in the robot being stuck in a loop and failed to solve the maze. This
report describes a hybrid method where one of the two algorithms is selected based
on the first opening of a reconfigurable maze. It has been demonstrated that by
combining the two algorithms, unless the maze was purposely configured
containing a loop, the rate of success is more than 90 percent.