THE LOADING OF MUNITIONS, PARTICULARLY DURING BATTLE, IS A MAJOR APPLICATION OF ROBOTICS.

THE LOADING OF MUNITIONS, PARTICULARLY DURING BATTLE, IS A MAJOR APPLICATION OF ROBOTICS. THE LOADING PROCESS, ALTHOUGH SEEMINGLY SIMPLE, IS MOST DIFFICULT UNDER ATTACK FOR A HUMAN OPERATOR IN THE PRESENCE OF PHYSICAL DISTURBANCES SUCH AS CHEMICAL/BIOLOGICAL WARFARE, RAIN, HEAVY WIND AND LACK OF LIGHT. SUBSEQUENTLY, FATIGUE IS COMMON AND BATTLE EFFECTIVENESS VARIES WIDELY AND IS EXTREMELY OPERATOR DEPENDENT. TO ENHANCE PRODUCTIVITY AND IMPROVE REPEATABILITY, ROBOT LOADING SYSTEMS NEED TO BE DEVELOPED. THE UNUSUAL ENVIRONMENT ASSOCIATED WITH MILITARY APPLICATIONS NECESSITATES THE USE OF "INTELLIGENT" ROBOT LOADING SYSTEMS EMPLOYING EXPERT SYSTEM AND ADAPTIVE CONTROL TECHNOLOGY. ON-LINE ADAPTIVE CONTROL TECHNIQUES FOR PRECISE PLACEMENT NEED TO BE DEVELOPED TO ADDRESS THIS PROBLEM. THE LARGE REAL-TIME COMPUTATIONAL REQUIREMENTS OF ADAPTIVE TECHNIQUES, NECESSITATES THE USE OF PARALLEL PROCESSING - PARTICULARLY WHEN CONSIDERING NONLINEAR ROBOT DYNAMICS. A HIGH-SPEED NUMERIC SYSTEM AND KNOWLEDGE BASE TO SET AND CONTINUOUSLY ADJUST LOADING PARAMETERS BASED ON SENSOR FEEDBACK IS PROPOSED. THE DEVELOPMENT OF PRACTICAL MODELS AND CONTROL OF THE ROBOT LOADER, STOCHASTIC ENVIRONMENT, THE ADAPTIVE CONTROL SYSTEM AND THE GENERATION OF THE KNOWLEDGE BASE ARE THE FOCUS OF THIS PROPOSAL.