Browsing by Issue Date, starting with "2025-05-28"
Now showing 1 - 1 of 1
Results Per Page
Sort Options
- Generation of Tire Monitoring Data: A Deep Learning ApproachPublication . MOURA, JOSÉ CARLOS CARDOSO DE SERRA E; Santos, Veríssimo Manuel Brandão LimaThe increasing complexity of automotive systems and the need to reduce development time and costs have motivated the adoption of synthetic data generation strategies as an efficient alternative to collecting and obtaining real-world data, especially for Deep Learning applications. This dissertation proposes an approach based on Long Short-Term Memory and Transformer Decoder models to generate synthetic data for sensors used in indirect tire pressure monitoring systems. Real-world data collected from road tests were used to train models, enabling them to capture the relationship between vehicle speed, steering wheel angle, and the sensors associated with the indirect tire pressure monitoring system. After preprocessing, normalization, and segmentation into temporal window sequences, the models were evaluated using reserved test data and performance metrics such as Mean Squared Error and Mean Absolute Percentage Error. The results demonstrate that both models are capable of generating synthetic data for most variables. This approach proves to be a promising tool for expanding the available dataset, thereby reducing the number of on-track tests required during the development and testing of indirect tire pressure monitoring systems.