Novel EMI Assessment Method Based on Statistical Detectors to Protect Sensitive Digital Radio Receivers

This paper presents a methodology to assess the degradation performance caused by EMI to digital radio receivers. The procedure developed employs statistical information of the interference and the communication signal to create estimators capable of classifying the disturbances according to metrics like the BER. The estimator is computed using machine learning techniques to relate the changes introduced to the APD diagram with the BER. The methodology has been validated through a laboratory experiment, where a narrowband QPSK system is interfered by additive white gaussian noise with different duty cycle and power level parameters. The validation results show that the estimator operates correctly and overcomes the problematics of relaying on the peak or weighting detectors with fixed resolution bandwidth.