Optimal Path Planning for a Robot in an Unknown Environment

The Micromouse competition has been running since the late 1970s around the world. The goal

is simple: get to the end of the maze as fast as possible. To solve the problem, a differential

drive robot was built. The robot was programmed to navigate the maze with a given map of the

maze’s walls. The proposed solution combines several control algorithms to ensure the robot

can navigate the maze without losing track of its location. First, a map of the maze and a route

from start to finish with all the points along it were coded in the robot’s memory. The route

consists of two main actions: driving forward and turning. Second, the robot was fed with an

initial position and a destination point in the maze. When navigating in the maze, the robot

follows the route while keeping track of its position using encoders and distance sensors. The

robot is programmed to reach the next point in the route using a PID control scheme over the

distance it needs to travel. Combining encoder readings and the robot’s kinematic model allows

it to determine the distance it has traveled. Distance sensors are used to validate the robot’s

position against the known map of the maze. The measurements from both the encoders and

the distance sensors are fused using a Kalman filter to improve the robot’s estimation of its

position in the maze. The proposed solution shows a way to implement the navigating phase

from the start to the goal after the robot has learned the layout of the maze while establishing

localization to navigate the maze.