THIS PROJECT WILL INVESTIGATE THE FEASIBLITY AND BENEFITS OF AN EXPERT SYSTEMS APPROACH TO NOISE STRIPPING.

THIS PROJECT WILL INVESTIGATE THE FEASIBLITY AND BENEFITS OF AN EXPERT SYSTEMS APPROACH TO NOISE STRIPPING. IN A COMMUNICATION ENVIRONMENT THAT CONSISTS OF A LIMITED VOCABULARY AND LIMITED INFORMATION EXCHANGE, ADVANTAGE CAN BE TAKEN OF PRIOR KNOWLEDGE OF THE ENVIRONMENT IN ORDER TO "GUESS" AT THE CONTENT OF A MESSAGE. THE FOCUS OF THIS PROJECT WILL BE TO INVESTIGATE THE FEASIBILITY AND BENEFITS OF UTILIZING AN EXPERT SYSTEM TO PREDICT THE CONTENT OF A MESSAGE AT THE WORD LEVEL AND THEM MAKE USE OF THE PREDICTION IN NOISE ESTIMATION AND REMOVAL. PHASE I WOULD INVESTIGATE THE UTILITY OF SUCH A NOISE REMOVAL METHOD. IF THE METHOD IS FOUND USEFUL PHASE II WOULD SEE A FURTHER EVALUATION OF PHASE I RESULTS OVER A LARGER RANGE OF APPLICATIONS AND THE DEVELOPMENT OF AN ARCHITECTURE AND DESIGN FOR THE IMPLEMENTATION OF THE METHOD.