Uncertainty- and Age-Aware UWB Fleet Scheduling with an Admitted-Cycle Energy Proxy
Artykuł w czasopiśmie
MNiSW
140
Lista 2024
| Status: | |
| Autorzy: | Wójcik Dariusz, Styła Michał, Kulisz Monika, Skowron Łukasz, Kłosowski Grzegorz, Rymarczyk Tomasz, Adamkiewicz Przemysław, Golec Ewa, Gauda Konrad |
| Dyscypliny: | |
| Aby zobaczyć szczegóły należy się zalogować. | |
| Rok wydania: | 2026 |
| Wersja dokumentu: | Elektroniczna |
| Język: | angielski |
| Numer czasopisma: | 21 |
| Wolumen/Tom: | 19 |
| Numer artykułu: | 5090 |
| Strony: | 1 - 23 |
| Impact Factor: | 3,9 |
| Scopus® Cytowania: | 0 |
| Bazy: | Scopus |
| Efekt badań statutowych | NIE |
| Materiał konferencyjny: | NIE |
| Publikacja OA: | TAK |
| Licencja: | |
| Sposób udostępnienia: | Witryna wydawcy |
| Wersja tekstu: | Ostateczna wersja opublikowana |
| Czas opublikowania: | W momencie opublikowania |
| Data opublikowania w OA: | 26 października 2026 |
| Abstrakty: | angielski |
| Industrial ultra-wideband (UWB) tags can share a limited ranging-update budget through gateway-side admission control. We evaluate an artificial intelligence (AI)-assisted priority rule that combines predictive uncertainty, nonlinear age of information (AoI), and a cold-start indicator, using dimensionally explicit coefficients. The dataset contains 166,142 cleaned range events from seven tags and eight anchors. Forecasting is evaluated on sessions held out in chronological order within each device. Replay draws 16 paired synthetic fleets of 100 virtual tags from 598 overlapping 80 s windows of three test sessions of one physical tag. Two state-update assumptions are compared. Under causal state updates using only recorded observations, the proposed priority rule does not outperform periodic scheduling at any evaluated budget from 5% to 40%; at a 20% evaluation budget, its mean absolute raw-range reconstruction error is 10.23% higher. At the same budget, retrospective replay with linearly interpolated state updates yields lower error (0.361 m versus 0.366 m) and an estimated matched-quality reduction of 2.49% in admitted evaluation cycles (95% paired-replay interval: 1.58–3.46%). This reduction is conditional on the interpolated retrospective replay assumptions, which use subsequent observations to reconstruct controller state non-causally. Admitted-cycle count serves only as a proxy for relative tag energy under an equal per-cycle cost assumption; electrical energy consumption was not directly measured. The replay evidence is limited to three physical sessions from a single device, so these findings do not establish generalizability to independent devices, environments, or simultaneous physical fleets. |
