Finding K-Factor (@Finding_K_Factor)

DoD High Speed Optical Processing 1991 II. WHAT MAKES OPTICAL PROCESSING DESIRABLE? The development of high-speed optical devices and subsystems, and their use in a variety of hybrid applications is a subject of intense research supported by DoD, DARPA, and NSF at many universities and industrial laboratories. These efforts are strongly motivated by the success of key demonstrations of this emerging technology. A synopsis of the technical status of some prominent HSOP devices, and the results of their successful use in selected demonstrations is shown below. It is these demonstrations that have stimulated interest in HSOP and provide the reasons why optical processing is desirable. The following is an overview of these developments. ➡️ Commercially promoted, low-cost 300-Mb/s fiber-optic subsystems are now used in low N applications to interconnect electronic digital computer modules, back-plane, and board-to- board subsystems. Extending this technology to higher data rates for commercial applications has been limited because of the high cost of the optical subsystem. DARPA is trying to stimulate commercial activity by funding the development and demonstration of 1-Gb/s-rate optical interconnection of parallel electronic processors. However, this technology has not yet been transitioned to the electronic digital community. ➡️ DoD and NSF efforts have resulted in significant technical advances for the different classes of SLM devices (Spatial Light Modulator) which include acousto-optic, ferroelectric liquid crystal, magneto-optic, micro mirror, and the recently developed MQW SLM. A three-order-of-magnitude improvement in processing performance is forecast for the next five years. ➡️ DoD support has produced progress in the development of photodetector arrays, and significantly, AT&T manufactured prototypes that integrate MQW SLM and photodetector functions at the pixel assembly level. ➡️ DoD funded programs have stimulated commercial interest in a variety of SEL arrays that are now routinely fabricated in the laboratory. Low power arrays of 1000x1000 MQW SELs on 20-micron centers have been demonstrated. ➡️ The Navy has reported excellent results for a ID channelizer that provides more than 2-GHz instantaneous bandwidth and more than 50-dB dynamic range. ➡️ The Army showed that the use of a 2D optical binary, phase only, SLM in ATR (Automatic Target Recognition) correlation functions was able to recognize nine out of ten targets whereas the conventional electronic approach only recognized five out of ten targets. Based on this demonstration, it is predicted that the ID channelizer capability will meet the operational requirements of many military users. ➡️ An experimental commercial optical neural network using a ferroelectric SLM and a LCLV CRT as an algorithm testbed shows promise as a programmable associative memory. The feasibility for 2D image processing at 10 to the 10th power (TeraHertz) ops/s was demonstrated. The above highlights of HSOP (High Speed Optical Processing) set the stage for the acceptance of this new emerging technology by the developers of military hardware, but the transition to actual deployed systems must be formulated according to a specific plan. https://apps.dtic.mil/sti/tr/pdf/ADA445369.pdf #Optical_Computing #Neural_Networks #Holographic_Memory #Lithium_Niobate #Analog_Computing #DTICs #SLM