Published Paper Details:

The main aim of this paper was to compare machine learning algorithms used in controlling a wall-following robot. A robot seldom has access to a complete and precise representation of its surroundings; thus, it must often operate under uncertainty. Even the physical properties of a robot might be questionable when used in real-world applications since such attributes can change significantly in different ways [1]. The motors and the limbs may be destroyed or damaged, the sensors may be obstructed, or the components may be reassembled improperly after repair. These are all possible outcomes [1]. Following the execution of a given motor order, a robot may be unable to make accurate predictions about its location. For want of a better term, it does not understand the plant equation, which is the equation that explains the consequence of acts that have been performed on it. When it comes to robot recovery after its physical design has been modified, we are particularly interested in establishing an effective technique.