Prediction Model for Microwave Radio Unit Testing

Abstract

During radio unit testing, the radio unit is tested in different test flows in order to ensure its conformance. But testing radio unit is time consuming and increases the unit’s production time. Prediction models are developed for the test flows to predict test points continuously by using the inputs from the previous test flows. Machine learning models presented in this thesis helps in the prediction of test point values before placing the radio unit in all testing stations. The developed prediction model helps in finding the measured value and their dependencies for failure. The model presented in this thesis helps to predict the future test point value and reduces the unit time to market. In addition, to obtain a more intuitive insight Graphical User Interface is built up to access the prediction model.