Predicting Energy Consumption for Heavy-Duty Vehicles via Time Series Embeddings (in collaboration with Volvo)
| Title | Predicting Energy Consumption for Heavy-Duty Vehicles via Time Series Embeddings (in collaboration with Volvo) |
|---|---|
| Summary | Develop deep learning based methods for time series forecasting; explore self-supervised learning methods for multi-variate time series embeddings |
| Keywords | |
| TimeFrame | 2024 Fall |
| References | Nalmpantis, C., & Vrakas, D. (2019, May). Signal2vec: Time series embedding representation. In International conference on engineering applications of neural networks (pp. 80-90). Cham: Springer International Publishing.
Foumani, N. M., Tan, C. W., Webb, G. I., Rezatofighi, H., & Salehi, M. (2024). Series2vec: similarity-based self-supervised representation learning for time series classification. Data Mining and Knowledge Discovery, 1-25. Lee, S., Park, T., & Lee, K. (2023). Learning to embed time series patches independently. arXiv preprint arXiv:2312.16427. https://github.com/seunghan96/pits Luo, D., & Wang, X. (2024). Moderntcn: A modern pure convolution structure for general time series analysis. In The Twelfth International Conference on Learning Representations. https://github.com/luodhhh/ModernTCN?tab=readme-ov-file Fraikin, A., Bennetot, A., & Allassonnière, S. (2023). T-Rep: Representation Learning for Time Series using Time-Embeddings. arXiv preprint arXiv:2310.04486. Wen, Q., Zhou, T., Zhang, C., Chen, W., Ma, Z., Yan, J., & Sun, L. (2022). Transformers in time series: A survey. arXiv preprint arXiv:2202.07125. Ahmed, S., Nielsen, I. E., Tripathi, A., Siddiqui, S., Ramachandran, R. P., & Rasool, G. (2023). Transformers in time-series analysis: A tutorial. Circuits, Systems, and Signal Processing, 42(12), 7433-7466. |
| Prerequisites | |
| Author | |
| Supervisor | Yuantao Fan & TBD |
| Level | Master |
| Status | Open |
Accurate energy consumption prediction is crucial for optimizing the operation of electric commercial heavy-duty vehicles, particularly for efficient route planning, refining charging strategies, and ensuring optimal truck configuration for specific tasks. In this thesis project, you will explore and develop time-series embedding methods, using self-supervised learning on data collected from heavy-duty vehicles, for downstream tasks, e.g., forecasting energy consumption and clustering vehicle operations. The proposed approach will be evaluated, validated, and compared with State-of-the-art time-series embedding methods on relevant downstream tasks. One important aspect is to learn useful embeddings for time-series analysis that can capture key characteristics of the transportation tasks, via self-supervised learning with augmented auxiliary tasks.
You will work closely with the Advanced Analytics Team (Volvo Group Technology) and have the opportunity to collaborate with other domain experts as well as various stakeholders in different tech streams.
Please contact Yuantao for more details.
