Building upon real-world automotive diagnostics and thermal monitoring benchmarks, this project leverages multi-variable sensor streams to estimate critical internal temperatures of a Permanent Magnet Synchronous Motor (PMSM). Instead of relying strictly on independent cross-sectional regression, participants will navigate sequential 2 Hz data streams, address the open challenge of missing values from cropped data segments, and generalize to entirely unseen driving cycles. By designing unified multi-output systems or ensembled sub-models under rigid offline constraints, this challenge highlights how advanced feature engineering and temporal dynamics can reliably infer economically unmeasurable physical parameters for motor safety and control optimization.
| Participant | ||
|---|---|---|
| 🥇 1 | Từ Đình Nguyên | 19.1769 |