METHOD FOR DETECTING FREEZING CONDITIONS FOR AIRCRAFT BY SUPERVISED AUTOMATIC LEARNING

• Main Authors: ALICE CALMELS, JEAN PASQUIE
• Format: Patent
• Language: Chinese; English
• Published: 30.07.2019
• Subjects: CALCULATING; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The present invention relates to a method for detecting icing conditions for an aircraft. The method comprises the steps of: measurement (320) of the parameters of embedded systems, with the exceptionof probes dedicated to the detection of ice, transformation (330) of the measurements of the symptomatic parameters to obtain P-tuples of values of explanatory variables which explain icing conditions, classification (370) of the measurements by classifiers on the basis of the P-tuples of values thus obtained, the classifiers having been previously trained in a supervised manner, each classifierproviding a prediction of membership in a class of icing conditions, the predictions of the various classifiers being consolidated (380) to provide a consolidated prediction (390) of icing conditions. 本发明涉及一种用于通过监督式自动学习来探测飞行器的冰冻状况的方法，该方法包括：测量(320)除了专用于探测冰的探头以外的嵌入式系统的参数；对所述征兆参数的测量结果进行转换(330)以获得解释结冰状况的解释变量的值的P元组；由分类器基于由此获得的值的P元组对所述测量结果进行分类(370)，所述分类器先前已经进行了监督式训练，每个分类器提供对结冰状况类别的隶属度的预测，这些不同分类器的预测被合并(380)以提供对结冰状况的合并预测(390)。