Biologically-Based Signal Processing Chips with Emphasis on Telecommunication Defect Tracking and Reliability Estimation

This paper provides observations and motivations to mimic biological information processing. Alternative bio-inspired systems definitions, basics, approaches, algorithms, and chip implementations will be illustrated to offer a base of choice for bio-based Intelligent Information Processing (IIP) systems. Hybrid biological and bio-based IIP are briefly presented. Two specific applications follow with embedded bio-based systems: Bio-chemical sensing and detection E-nose; and Track improvements In the reliability of the software used in telecommunication network deployments. The biologically-based processing discoveries gleaned from observing the spikes in the brain activity of monkeys, introduced the concept of plasticity in synapses used in our embedded Spiking Neural Network (SNN) system for the E-Nose The mathematical construct of a defect tracking classifier is nonlinear, and the event to be recognized involves a sequentially varying or non-stationary phenomenon for telecommunication defect tracking and reliability estimation. Thus, Adaptive Recurrent Dynamic Neural Network (ARDNN) system using wavelet function as the basis improved the failure event estimation of software defect tracking in telecommunications and reduced the error from 88% to L25-8%.