Virtual sensor prediction on CAN bus
0.119MSEUnified model + Variate embedding V, V'
Evaluation results
| Method | Links | |
|---|---|---|
| Unified model + Variate embedding V, V'Inference Time=58.23 ms, Memory=3.46 GB, Parameters=7.8 M2026.01 | 0.119 | |
| Unified model + Variate embedding V, V' + Signal relevance R' + N_train = 4 + Sparse input sets Zj + Sensor selectionInference Time=0.14 ms, Memory=0.0036 GB, Parameters=7.8 M2026.01 | 0.121 | |
| Unified model + Variate embedding V, V' + Signal relevance R' + N_train = 4Inference Time=58.41 ms, Memory=3.46 GB, Parameters=7.8 M2026.01 | 0.136 | |
| Individual models + Variate embedding V, V'Inference Time=58.05 ms, Memory=3.42 GB, Parameters=124.8 M2026.01 | 0.177 | |
| Unified modelInference Time=58.19 ms, Memory=3.46 GB, Parameters=7.0 M2026.01 | 0.727 |