Viterbi soft-decision decoding method for convolutional codes based on classa ocean environment noise model

Abstract

The invention relates to a Viterbi soft-decision decoding method of a convolutional code based on a Class A marine environment noise model. In this method, the data preprocessing operation is carried out before the traditional Viterbi soft-decision decoding of the convolutional code based on the Gaussian model, so the additional calculation amount is small. In this method, a simplified Class A noise model is firstly derived, and then a Viterbi soft-decision branch metric and decoding method for convolutional codes in additive Class A noise channels are derived. The method provided by the invention can improve the reliability of the Viterbi soft decision decoding method in the shallow sea channel with strong pulse characteristics, and further improve the communication quality of the underwater acoustic communication system.

Description

technical field [0001] The invention belongs to the field of signal processing, in particular to a convolutional code Viterbi soft-decision decoding method based on a Class A marine environment noise model. Background technique [0002] As an important part of the underwater acoustic communication system, the signal receiver's performance is directly related to the success or failure of the system communication. In almost all systems, there are both desired signals and undesired interfering signals (namely noise signals) in the observation data. In other words, a model that reflects the real statistical characteristics of the noise data will help the design of signal receivers. Due to the central limit theorem, the Gaussian noise model can bring linear processing, and the Gaussian model can be described by only two parameters of mean and variance, most traditional signal receivers assume that the noise is a Gaussian model. Although the Gaussian assumption of noise simplifie...

AI Technical Summary

Problems solved by technology

Although the Gaussian assumption of noise simplifies the design of system signal receivers, a random variable subject to a Gaussian distribution shifts its mean exponentially by the square of the deviation, so the probability of random sampling far from the mean is very small, for example, The probability of a Gaussian random sampling with an offset of ±10δ from the mean is 1.5×10 -23 , which means that the Gaussian distribution is not suitable for describing the distribution that may produce a large number of data mutations

In practice, due to the strong impulsive noise such as industrial noise and biological noise in the shallow sea environment, the traditional classic Gaussian model can no longer describe this strong impulsive shallow sea environmental noise, which will directly lead to the signal reception based on the Gaussian model. The use performance of the machine in the non-Gaussian model is greatly reduced, and even the effective processing of weak signals cannot be achieved, which will seriously hinder the receiving end from making correct decisions

Images