Abstract
Neural Model of Radiolocation Signal Patterns ExtractorSimulation researches have been conducted of programmable neural extractor model which realizes object detection and determines its position on the background of noise. For the simulation of radiolocation signal the model of signal has been used which is a sum of two independent processes, namely signal reflected from an object, and noise. This model takes into account impulse sounding, radiation characteristics of rotating aerial, fluctuation of echoing area of an object, and Doppler shift of carrier frequency. For comparative researches, models of superficial and superficial-amplitude extractors have been adopted. As a result of experience gathered during project work, the best network structure for the tasks given to neural extractor appeared to be a three-layer network trained by means of back propagation, having two hidden layers with tangensoidal activation function and a moveable threshold.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
