Abstract: A neural network is used within a receiver to discriminate a large set of input waveforms without using a very large set of conventional matched filters. The neural network is trained under actual line conditions as opposed to the requirement for ideal signals when using matched filters. The finite waveforms are based on digital modulation principles. A best match is made between a received waveform from the noisy channel and that of previously trained waveforms in order to extract data. Neural network based matched filter allows data be discriminated separately for each sub-carrier channel in the receiver. The neural network system allows fast processing and is suitable for high-speed data communications systems.

Abstract: The proposed invention teaches basic principles of “orthonormal transform” to be used to convert a set of discrete samples into a set of coefficient real samples that is contained in a finite field. The number of real values in each coefficient samples is finite and coded for transmission using digital modulation. It also teaches that handling of multi-path fading of Doppler effects implies that the Bit Error Rate (BER) performance as a function of Bit Energy/Noise (Eb/N0) is close to the performance of Additive White Gaussian Noise (AWGN) channel. The effect of impairments is minimized and only the effect of thermal noise (AWGN) is maintained. The inventive apparatus is simple and maintains constant end-to-end response time, sustainable effective data rate and bounded error performance which is conducive to specify a Quality of Service (QoS) which is useful for service provisioning.