Robotic Conceptual Frameworks
From RobotCurwikilums
There are three autonomous robotic conceptual frameworks in prevalent use today. A conceptual framework, used in this context, is the highest-level, overarching hardware and software architecture used to implement an autonomous robot.
Hierarchical
The Hierarchical framework is the "classic" architecture for autonomous robots that emerged in the 1960's and refined over the next two decades. (See Figure 1.) A hierarchical implemented robot would sense its enviroment, build and/or edit an internal world representation from the sensed data and current state information, think or plan its next set of goals derived from its internal world, and then act on those goals.
Behavioral
The behavioral or reactive framework was a 1990's roboticists' 'reaction' to the Hierarchical approach. Inspired by simple biological organisms, such as worms and insects, that do very well in the very complex world (the number of insects species alone are estimated to be approximately half of the total number of described species.) A behavior based robot senses it immediate environment and reacts directly with its sensed data to move, grab, etc. (See Figure 2.) Hence the robot's world model is grounded to the physical environment and the intelligent system emerges from the coupled robot - environment.
Hybrid
The hybrid framework, as the name suggests, attempts to blend the strengths of the Hierarchical and Behavioral models. (See Figure 3.) It is interesting that the Science or Biology of Mind has a principle that different areas of the human brain carry out specialized mind functions. Perhaps, as a corollary, the human sense of a single unified self is illusionary.
