Hardware and software architectures for realizing a knowledge based system on eave

Realizing the capabilities of a Knowledge Based System (KBS) in an autonomous underwater vehicle requires the integration of several hardware and software subsystems. This paper details both the VME bus based multiprocessor architecture as well as the LISP environment chosen for implementing a KBS. Overcoming the problems of execution speed and storage requirements in a real time KBS is a focus of this work. This design includes a real time operating system, PSOS, running on multiple 68020s with an interprocessor communication package handling information exchange. The high level software environment is based on Common Lisp and forms the backbone for symbolic processing. This has been implemented with Portable Common Lisp Subset, PCLS, which has been adapted to the EAVE vehicle. A unique aspect of this development effort is the utilization of several conventional technologies to realize a high performance symbolic processing system.