Predictive Modelling of Two wheeler Durability Testing Based on Customer Usage Profiles

Proving ground testing is a critical and resource-intensive aspect of vehicle durability qualification. The development of test schedules is traditionally based on customer usage patterns, which vary significantly across different markets, often requiring multiple and costly data acquisition efforts. This study presents a machine learning approach to streamline this process. By collecting comprehensive data on diverse road and load conditions, we trained an AI model to predict proving ground (PG) schedules for new customer usage profiles without needing new data acquisition and simulations. The trained model, a Support Vector Machine (SVM), successfully predicted test schedules with a maximum error of 7.0% when validated against unseen data. The primary advantage of this method is the rapid and cost-effective generation of proving ground test schedules for new markets that share similar road types but have different usage distributions, eliminating repetitive data collection and lengthy simulations. Keywords: Two-wheeler, Durability, ground testing, Artificial Intelligence (AI), Machine Learning (ML), Vehicle RLDA, Track Optimization, RapidMiner, Deep Learning, Support Vector Machine (SVM)