GENERAL PURPOSE, OPTICAL, VECTOR-MATRIX MULTIPLIER

A GENERAL PURPOSE, OPTICAL, VECTOR-MATRIX MULTIPLIER WILL BE DEVISED TO COMPUTE 256X256 ELEMENT MATRICES TIMES 256-ELEMENT VECTORS AT 8-BITS PRECISION IN LESS THAN 100 MICROSECONDS. THIS PROJECT WILL TAKE AN INNOVATIVE ARCHITECTURAL APPROACH THAT USES TWO 64-CHANNEL ACOUSTO-OPTIC DEVICES OPERATING AT SEVERAL HUNDRED MEGAHERTZPER CHANNEL IN A TELECENTRICALLY IMAGED CONFIGURATION. THE USE OF LASER DIODE ARRAYS, COMPUTER-OPTIMIZED OPTICAL SYSTEMS, MULTICHANNEL GALLIUM-PHOSPHIDE AO DEVICES, AND 128-CHANNEL HYBRID AVALANCHE PHOTODIODE ARRAYS WITH PROCESSING ELECTRONICS, COUPLED WITH AN INNOVATIVE "SYSTOLIC" DATA FLOW OPTIMIZED FOR THE HARDWARE, ALLOWS PERFORMANCE PROJECTIONS WELL WITHIN NASA'S REQUIREMENTS. THE MACHINE ARCHITECTURE IS FOCUSED TO THE SOLUTION OF SIMULTANEOUS EQUATIONS FOR THE MATRIX-INVERSE PROBLEM. ANTICIPATED USE IS ULTIMATELY IN A FEEDBACK CONTROL SYSTEM TO CALIBRATE NASA MULTIPLE 256-ELEMENT MIRROR ARRAYS. A GENERAL PURPOSE, OPTICAL, VECTOR-MATRIX MULTIPLIER WILL BE DEVISED TO COMPUTE 256X256 ELEMENT MATRICES TIMES 256-ELEMENT VECTORS AT 8-BITS PRECISION IN LESS THAN 100 MICROSECONDS. THIS PROJECT WILL TAKE AN INNOVATIVE ARCHITECTURAL APPROACH THAT USES TWO 64-CHANNEL ACOUSTO-OPTIC DEVICES OPERATING AT SEVERAL HUNDRED MEGAHERTZPER CHANNEL IN A TELECENTRICALLY IMAGED CONFIGURATION. THE USE OF LASER DIODE ARRAYS, COMPUTER-OPTIMIZED OPTICAL SYSTEMS, MULTICHANNEL GALLIUM-PHOSPHIDE AO DEVICES, AND 128-CHANNEL HYBRID AVALANCHE PHOTODIODE ARRAYS WITH PROCESSING ELECTRONICS, COUPLED WITH AN INNOVATIVE "SYSTOLIC" DATA FLOW OPTIMIZED FOR THE HARDWARE, ALLOWS PERFORMANCE PROJECTIONS WELL WITHIN NASA'S REQUIREMENTS. THE MACHINE ARCHITECTURE IS FOCUSED TO THE SOLUTION OF SIMULTANEOUS EQUATIONS FOR THE MATRIX-INVERSE PROBLEM. ANTICIPATED USE IS ULTIMATELY IN A FEEDBACK CONTROL SYSTEM TO CALIBRATE NASA MULTIPLE 256-ELEMENT MIRROR ARRAYS.