This example shows an example where a local minima is not so easy to escape from. In fact, the random walk occurs so many times, that the robot attempts to BACKTRACK several times without success. A random walk eventually leads the robot to an area where DESCEND leads the robot to the goal. An assumption comes into play here, because it appears as if the robot grazes an obstacle as it goes around it. This is because of the size of the random steps, and not checking to see if the path between configurations lies completely in CFree.