Our lab member Vispi Karkaria presented a pioneering digital twin technology at the 2024 REMADE Circular Economy Technology Summit & Conference, hosted in Washington, D.C., from April 10-11. The technology, designed for predictive maintenance, specifically targets tire health monitoring to improve automotive safety and efficiency.
Our team’s predictive digital twin concept functions through three crucial steps. Initially, we tackle data complexity by employing reduction techniques to concisely represent physical tires using historical usage and telematics data. This foundation allows us to train a transformer-based model offline, predicting future tire health while quantifying uncertainties to provide confident predictions. In the second step, we integrate real-time data updates into the digital twin, using a hybrid modeling approach to maintain accuracy over the tire’s lifecycle. Finally, we implement a Tire State Decision Algorithm, which strategically uses the model’s forecasts to determine the most effective times for tire maintenance or replacement. This method not only predicts the state of tire health but also continually refines its representation and maintenance decisions, optimizing operations and prolonging tire life.