Abstract
This paper presents the concept and implementation of EusLisp developed for describing 3D geometric models on which environment recognition functions and intelligent task planning are performed. Based on an idea that object-orientation is best suited for intelligent robot programming, object oriented functions are implmented at the basis and Lisp functionalities similar to Common Lisp are built on it. For the smooth unification of this object orientation and functional programming, constant-time type discrimination scheme in cascaded class hierarchy is implemented. Memory efficiency for variable-size memory allocation is improved by introducing the Fibonacci buddy memory management. Performance of the multi-thread parallelism with thread local memory management scheme is evaluated. The advantages of EusLisp are demonstrated by its systematic implementation of a solid modeler and two robot applications.
