Real-time Neural Vision For Vehicle Navigation And Safety

Abstract

This paper presents an investigation of the application of image analysis, processing and transmission techniques to vehicle navigation and safety. Orthogonalization neural network architectures are used for complex tasks of real time image processing for purposes of navigation and safety, in situations of vehicle maneuvering in unknown and hazardous environments. Hardware-implemented neural networks are used for image processing and compression/decompression on both ends of a communication link between a Traffic Information Center and vehicle.