Exploring Predictive Maintenance and Signal Processing Techniques for Automotive Health Monitoring

Predictive maintenance is one of the most crucial technologies in today's world, having applications in various domains like automotive, aerospace, manufacturing, and so forth. It is a data-oriented approach that helps predict if any machine or associated component with the machine needs maintenance or not. It also helps predict the remaining lifespan of the same. Various machine learning-based algorithms and models are being developed in this domain. S ome of them include Deep-learning-based approaches like Autoencoders, ANN s, CNN s, and so forth, and Signal Processing models using factors like Mean Squared Error for superimposition of primary signals with appropriate opposite phase signals. This article presents a systematic study of various research works done in the fields specific to the Automotive sector. This study focuses on the problems solved by the authors through their proposed solutions, some of the results and insights, and a brief description of their proposed solution. Approximately 30 research papers were included selectively in this study. This study concludes by suggesting a future direction in the predictive maintenance field based on periodic monitoring, in contrast with the trend of continuous monitoring for early diagnosis.