Belief-Aware Scheduling for Predictive Wildfire Hazard Mapping under Sparse-Window Telemetry
This work provides a principled approach to scheduling and belief maintenance for edge-based predictive mapping under communication constraints, relevant to wildfire monitoring and similar sparse-telemetry applications.
The paper addresses the problem of predicting wildfire hazard maps under limited down-link telemetry by proposing a belief-aware scheduling framework. Results show that a lightweight 40k-parameter encoder outperforms the FAIR activity-paced reference by ~28% on default and ~11% on structured landscapes, while a deeper Transformer does not improve performance.
An edge node monitoring a wildfire observes more than a duty-limited or windowed down-link can carry. The receiver must predict the H-step-ahead hazard map from whatever the link delivers. We argue the operative design problem is not which neural architecture to use but how to derive a structured belief sufficient for the receiver's prediction task and maintain it through a scheduler that anticipates future transmission opportunities. We formalize this as a partially observed sequential allocation problem with three coupled per-region action axes (sensing, representation, transmission), and derive each component of the structured belief from the H-step forward operator's input requirements. Identifying these mechanisms requires independent control over the window period P, per-window capacity C, predictive horizon H, and fuel composition, which is not separable in real-landscape data; we therefore evaluate on a physics-calibrated synthetic environment. Three empirical observations support the principle: the gap between a non-myopic activity-paced reference and uniform pacing is unimodal in window-period sparsity, peaking at intermediate spacing; ablating the structured belief, the dominant operative component flips between a default landscape (temporal staleness) and a structured landscape (static-risk prior), while the per-cell intensity belief is redundant in both; and a 40 k-parameter lightweight cross-region attention encoder exceeds the FAIR activity-paced reference by ~28% on the default landscape and ~11% on the structured landscape. A deeper Transformer encoder does not improve over the lightweight encoder in mean predictive loss and exhibits higher training-seed variance. Within this task class and regime, a modest architectural inductive bias suffices when the belief and the scheduling problem are correctly posed.