Smart Environmental Monitor Based on Neural Networks and Multi-Specked Pattern Recognition

Accelerating industrialization has caused an increasing demand for environmental monitoring. Because of this, physical Optics Corporation (POC) proposes to develop a novel real-time smart environmental monitor (SEM), utilizing a novel holographic optical neural network (HONN) recently developed at POC for pattern recognition applications. These new architectures combine an optical implementation of the so called N4 neural network architecture, high efficiency compact holographic memory matrices, and new fast electronic algorithms for feature extraction and statistical data analysis. The proposed system will be extremely fast and highly rugged. It will be capable of detecting various contaminants simultaneously with extremely high resolution and sensitivity. POC's HONN can also be made compatible with imaging and mass spectromenters. POC's HONN system can be further expanded in Phase II to contain 64,000 fully interconnected neurons, with over 4x10 9 interconnections. POC's HONN will be capble of "learning" over 20,000 features of different spectral patterns and perform real-time spectral analysis and classifications. The success of this project will result in the most advanced level of a field deployable multi-spectral detection technology.