McMaster University's Artificial Computing System

The developers of this robot intend to simplify the complex task of making a robot intelligent by classifying the types of problems faced by the robot. In that way, the robot can determine the best course of action depending upon the task required. Obviously, a lot of time will be spent determining the key factors necessary to identify each type of problem. To simplify the task of classifying the problems faced by the robot, we intend to use our engineering physics background to develop an approach. By understanding the key underlying physical factors specific to a situation, and using these to develop a solution, the amount of processing required of the system will be greatly reduced. Only essential data will be collected, with the actual system determining what data and how much data is needed to solve the problem. In keeping a general approach to problem solving, the robot will need to identify situations that have been previously encountered. The key to identifying similar problems is collecting the proper data to determine the robot’s situation. For example, finding four solid objects enclosing the robot could be interpreted as a room. What to do in this room would then be the new problem faced by the robot. The initial task of finding the four walls would be simplified by first getting the correct data to identify the objects as walls and not obstacles to be avoided. We hope to avoid relying on properties specific only to one goal to keep the algorithms general.