Method for evaluation of unsupervised vector magnitude error signals with low signal to noise ratio

摘要

The present invention describes an unsupervised method for evaluation of quality of signals transmitted in digital communication systems by means of the measure of the magnitude of the error vector (EVM). The estimation of the magnitude of the error vector (EVM) is based on parts of the diagram of the constellation with a lower probability of errors of decision of the symbols.The choice of the areas of decision depends on the knowledge of the position of the centers of clusters of samples related to each symbol constellation used. This position can be known a priori or may be determined by the analysis of the received signal.Once defined the areas of decision, the partial measures of the magnitude of the error vector (EVM) are calculated independently in each zone. Finally, the partial measures of the magnitude of the error vector (EVM) are combined to obtain the magnitude of the error vector (EVM) total.The proposed method is suitable for application in situations in which the number of errors of symbols is large, such as signals with low signal to noise ratio, and can be used even into signals which do not employ brokers of error codes.