Spike Modeling Of Hebbian Learning For Sensor Deployment

Abstract

In the next generation wireless technologies like 5G/6G and beyond, there is a scope for intelligent and smart way of information transfer and services to the society, which is scalable. With such potential communication capacity, many smart activities are initiated to tackle more intrinsic and extrinsic services in the areas like smart city, agriculture, health, industry, automation etc. In all of these areas to become a smart environment, there is a requirement of embedded sensors, Internet of Things (IoT), edge computing, smart sensors to sense and tag the phenomenon of interest with digital decision support systems. Here, the work proposes a novel technique to identify the dynamics of the phenomenon using the gradient of different parameters. Spike based hebbian learning model is introduced with mathematical analysis to hand hold the dynamics of the phenomenon using the adaptable patterns for energy efficient sensor usage. Implementation results are compared with the analytics in a given simulation environment. System is trained and tested for stochastic gradient analysis to model the error.