Designing a mutating plan vector for a learning robot
A. Das · 2003
Mutating plan vector design for a learning robot is discussed with emphasis on the need for a Skolemized knowledge base. In the proposed model, a plan is defined as a vector between the start frame and the goal frame. This vector is composed of partially ordered bases consisting of task-level and robot level parameters. When the robot fails to work according to a given plan, mutations of some of the bases are said to have taken place. For further mutations in a correct direction, the robot needs knowledge base support. The knowledge base is Skolemized and a unification mechanism operates.>