Root causes analysis and fault prediction in intelligent transportation systems: coupling unsupervised and supervised learning techniques

John William Mbuli (1) , Maroua Nouiri (1) , Damien Trentesaux (1) , Damien Baert (2)

1 LAMIH - Laboratoire d'Automatique, de Mécanique et d'Informatique industrielles et Humaines - UMR 8201
2 Bombardier

Digitization, new technologies and big data have the potential role to change the way to model, manage and organize industrial manufacturing, production and services. With emerging technologies and current trends like the fourth industrial revolution (industry 4.0) and the artificial intelligence (AI), the transportation sector is rapidly changing. This paper describes a method for root cause analysis and faults prediction in intelligent transportation systems (ITSs). The method is based on the coupling of unsupervised and supervised machine learning techniques. To assess its performance, the proposed method was tested on a Train Door System at Bombardier Transport (BT) as a case study. The proposed method improves the system's reliability and helps the maintenance supervisor in adjusting and setting new rules in maintenance decisions

Mots clés: "Maintenance engineering"; "Transportation"; "Clustering algorithms"; "Big Data"; "Supervised learning"; "Fault detection"; "Machine learning algorithms"

Domaines: Automatique / Robotique

John William Mbuli, Maroua Nouiri, Damien Trentesaux, Damien Baert. Root causes analysis and fault prediction in intelligent transportation systems: coupling unsupervised and supervised learning techniques. 2019 International Conference on Control, Automation and Diagnosis (ICCAD), Jul 2019, Grenoble, France. ⟨10.1109/ICCAD46983.2019.9037877⟩. ⟨hal-03468706⟩