Artificial Intelligence Based Predictive Maintenance for Two-Wheeler Automobiles

A wide variety of applications have been made possible by recent advancements in maintenance modeling that have been driven by data-based methodologies like machine learning (ML). Minimizing maintenance costs while guaranteeing functional safety over a product's lifetime has emerged as a formidable obstacle for the automobile industry. In order to do this, predictive maintenance (PdM) is an essential strategy. The vast amounts of operational data produced by current cars make ML a prime contender for PdM. Despite the abundance of literature on both PdM and ML in automotive systems, no comprehensive study has yet been published on ML-based PdM. There is a growing demand for this kind of study due to the rising quantity of articles in this sector. As a result, we evaluate the articles from both an application and ML standpoint after surveying and classifying them. After that, we will go over some open difficulties and potential areas for further study. We draw the following conclusions: (a) more research would be conducted with publicly available data; (b) most papers use supervised methods, which need labelled data; (c) combining data from multiple sources can improve accuracy; and (d) deep learning methods will continue to grow in popularity, but only if large amounts of labelled data are made available and efficient methods developed.