Deep-Space Downlink Triage (2D): AI Data Prioritization Simulator

A spacecraft orbits a planet, gathering far more science observations than its scheduled Deep Space Network contact window can carry home. This simulator gives every observation a simulated novelty score — the kind an onboard anomaly detector produces — and lets an AI scheduler greedily fill each limited-bandwidth pass with the highest-value packets that clear an anomaly threshold, exactly the triage problem behind NASA's onboard science autonomy (Perseverance's AEGIS targeting, and the data-prioritization pipelines proposed for future observatories). Switch to FIFO mode to see a naive first-in-first-out downlink lose most of that value and overflow its buffer once observations arrive faster than they can be sent, while sliders and live telemetry track the budget, generation rate, and how much scientific value actually reaches Earth — with a live density-ranked knapsack chart showing exactly which packets the scheduler is choosing and why.