A NEW OBJECT REPRESENTATION FOR ROBOTICS AND ARTIFICIAL-INTELLIGENCE APPLICATIONS

By combining the simplicity of the sphere with the power of the notion of hierarchy of detail, a new practical model is proposed for applications in robotics and artificial intelligence. The new model is based on a double spherical representation of solid bodies. First, each element making up a given object is approximated by means of a set of exterior spheres which are automatically defined. Second, another set, composed of interior spheres, is defined; this inner representation is also generated in an automatic way. These representations define a hierarchy, since they can be redefined as many times as necessary; starting with two spheres per element, an outer sphere and an inner one, the approximation may be improved until it contains hundreds of spheres. In this way, both representations converge to a zero-error model. The proposed spherical model leads to a simpler treatment of several important problems in robotics and artificial intelligence. Namely, the problems of collision detection, motion planning and spatial reasoning can be dealt with in a simpler manner by using only spheres, while these problems become much harder, or even unsolvable, when objects with a more complex geometry are involved.