Autonomous and Rational Behavior of Sony AIBO Robot
Yee Sern Wei · 2005
Autonomy has been a key research focus in the field of robotics approach. In particular, there are two main streams on how intelligence is engineered into autonomous robots. One is the behavior-based approach. The other is the knowledge representation based-approach. Intelligence emerges from behavior-based approach and autonomy is achieved by interactions between the dynamics of the environment and a large repository of simple robot behaviors. The behaviors are run in parallel and governed by motivational attitudes. For instance, an AIBO robot has more than a thousand behaviors built into it. Task can be accomplished through a series of behaviors run concurrently. However, behavior-based robots exhibit simple creature like intelligence. In other words, behavior-based robots are only equipped with a brain as small as insects (i.e. ants or bugs), which implies that they are not able to perform more sophisticated reasoning or decision making. In particular, they lack rationality on their behaviors. For instance, take opportunities, be persistent with goals and yet be flexible enough to drop un-achievable goals. This research is the first attempt to design and program rational intelligence into robot. The program and testing is done using a Sony AIBO robot and the reasoning framework is by Belief, Desire and Intention- BDI.