Improving Reliability of 2 Wheelers Using Predictive Diagnostics

Abstract

<div class="section abstract"><div class="htmlview paragraph">The On-Board Diagnostics (OBD) system can detect problems with the vehicle’s engine, transmission, and emissions control systems to generate error codes that can pinpoint the source of the problem. However, there are several wear and tear parts (air filter, oil filter, batteries, engine oil, belt/chain, clutch, gear tooth) that are not diagnosed but replaced often or periodically in motorcycles/ power sports applications. Traditionally there is a lack of availability of in-field and on-board assistive tools to diagnose vehicle health for 2wheelers. An alert system that informs the riders about health and remaining useful life of their motorcycle can help schedule part replacements, ensuring they are always trip-ready and have a stress-free ownership and service experience. This information can also aid in the correct assessment during warranty claims. With the increase of onboard sensors on vehicles, there has been a notable increase in the availability of condition-monitoring data such as vibration, temperature, pressure, voltage, and other electrical and mechanical parameters. The connectivity device on the motorcycle can transmit this onboard real time data to the cloud for analysis to derive the information of useful life of these components. This paper presents an edge-plus-cloud architecture with part of the algorithm in the Engine Control Unit (ECU) and final processing done on the cloud. Various sensor signals and other vehicle operating parameters are collected and processed using a combination of Machine learning, Fast Fourier Transform, Regression models and other data analytical algorithms. Based on the analysis, information transmitted back from cloud/ Edge device to Vehicle Instrument cluster/ Mobile App/ Web UI to inform rider before the failure has occurred, along with real time data of the remaining useful life of these components.</div></div>