Decision trees based multi-factor remote technical diagnostics of car retarders

• Main Authors: Panasov, Viktor L., Tseligorov, Nikolay A., Nechitaylo, Nikolay M.
• Format: Conference Proceeding
• Language: English
• Published: SPIE 27.04.2023
• ISBN: 9781510664845; 151066484X
• ISSN: 0277-786X
• DOI: 10.1117/12.2680707

Modern hump car retarder control systems and their technical diagnostic methods are reviewed. Disadvantages of existing analytical methods of technical diagnostics are revealed. Factors, which help to increase the accuracy of technical diagnostics of car retarders, available in existing hump automated control complexes, are shown. The paper suggests to build a new intelligent technical diagnostic system to improve the accuracy of the current ones. The suggested system will use decision trees, created by means of Data Mining approach under conditions of small training set amount. This approach is based on the ideas of intelligent system self-organization. The increase in accuracy of the decision rules is shown on the verification database examples.