Application of hybrid edge-cloud algorithms to detect clutch abused usage on commercial vehicles

Abstract

With the widespread adoption of Telematics and advancements in edge computing, more and more hybrid edge-cloud solutions could be used to monitor vehicle operating conditions effectively. Clutch, as a critical power component of vehicle, transfers work from engine to wheels and is often damaged by driver’s inappropriate operations. In this study, a hybrid edge-cloud solution is proposed to monitor the usage of clutch. The edge algorithms are deployed in Engine Control Module (ECM) or T-Box, processing high-frequency and high-dimensional data, for example, engine speed and torque, then sending out Clutch Abused Index (CAI), a quantified value to indicate the energy consumption of clutch. Results show that CAI could be used as an accurate indicator to reflect the driver’s operation. Then, cloud algorithms are developed to monitor and detect the inappropriate operations. A varying threshold, interquartile range (IQR) 95th value of 16-trip moving window, is used to detect inappropriate operations with one trip latency. The standard deviation value of 16-trip moving window was used to detect the abnormal operation period, with about 3 days latency for abnormal and 1 day latency for inappropriate, respectively. The novel approach proposed in this study could be used as health monitoring of critical components that lack direct measurement sensors or that with high data transmission cost, contributing to vehicle maintenance strategies and reliability, such as extending lifespan and predicting remaining useful life.