Date: 2026-03-03
Author: GitHub Copilot (session with S. Bolling)
Status: Active Development
Branch: main (ogc-csapi-explorer)
Commits: 4b0c986, 23e62fb, b5aac7c, 178dd69, e4da800, 5816d41

The initial data simulator run was a firehose: it dumped hundreds of observations per node in rapid succession, all at once. The server accumulated them, and when the map loaded it pulled back ~500 per datastream. The result was 516 bearing lines radiating from all 3 nodes simultaneously — impossible to read, no sense of time progression, just a starburst of chaos.

Additionally, Live Mode didn't work at all initially because isLocationRelatedDatastream() only matched GPS/location keywords. LOB, Track Update, Scene Summary, and Classification datastreams were all filtered out, so the observation layer iterator had nothing to render.

• Simulator publishes 4 observation types per detecting node every 5 seconds with geometrically correct bearings computed from node→UAV positions
• Live Mode fetches only the last 10 observations per datastream and applies recency coloring
• Detection rings are scaled to 3km outer radius — overlapping coverage zones are clearly visible
• Mobile has full parity with desktop Live Mode toggle

The sidebar currently shows ~1,456 Lines of Bearing accumulated from the initial chaos run. In normal (non-live) mode, the map pulls up to 500 per datastream and renders them all, producing a dense mess. Decision: keep the data on the server for now; lean on Live Mode to hide the residue. Purging will be revisited later.

Based on a PowerPoint mockup created by the operator (orange annotation lines overlaid on the live map screenshot), the ideal visualization is:

1. Each LOB originates at a sensor node and extends outward along the computed bearing toward the UAV
2. Multiple LOBs from the same node fan slightly as the UAV moves — angular spread shows target motion
3. Where LOBs from 2–3 nodes cross = triangulation — the intersection zone is the estimated target position
4. The lines are clean and readable — approximately 3–5 per node visible at any time, not hundreds

This is the "money shot" for the demo: a human operator glances at the map and immediately sees where the drone is based on converging bearing lines.

~1,400+ stale LOBs from the chaos run still exist on the server. In non-live mode, these all render. Mitigation (current): Live Mode hides them by fetching only the last 10. Future option: purge via DELETE or add a server-side time filter.

Bearing lines extend a fixed distance from the sensor node. If too short, they look like stubby whiskers and don't show convergence. If too long, they clutter. The length needs to be tuned so that lines from adjacent nodes clearly intersect in the detection overlap zone.

The PowerPoint naturally shows where lines cross, but we don't currently compute or highlight the intersection point. A future enhancement could plot the estimated target position (the "fix") as a distinct marker where 2+ LOBs converge — this is the classic DF (direction-finding) workflow.

Currently set to 10 per datastream. With 3 LOB datastreams (one per node), that's up to 30 bearing lines at once. The PowerPoint mockup shows ~5–7 total, which feels right. Options:

• Drop to 3–5 per datastream for a cleaner look
• Keep 10 and rely on recency fading (old lines go near-invisible)

When Live Mode is off, what should the map show? Full history is overwhelming. Options for future work:

• Default "last hour" or "last 50" time filter
• Forensic/replay mode with a time slider
• Simply guide users to enable Live Mode for operational view

Three acoustic sensor nodes (AZ-MA-1, AZ-MA-2, AZ-MA-3) are deployed around a SNET in southern Arizona. A UAS enters the area. As it passes through the 3km detection envelopes, each node publishes lines of bearing in real-time via the OGC Connected Systems API. An operator watching the map sees converging LOBs and can visually triangulate the drone's position and track its movement across the sensor network.

• Oracle Server: https://os4csapi-osh.duckdns.org/sensorhub/api
• Deployed App: https://ogc-csapi-explorer.pages.dev/map
• Simulator Script: scripts/simulate_scenario.py
• Detection Range Config: Lines 159–195 of demo/src/pages/MapViewPage.vue
• Live Mode Logic: Lines 141–145 (state), 1917–1923 (refresh), 1647–1721 (render)
• Recency Styling: getRecencyBearingStyle() at line ~232
• Datastream IDs:
  • AZ-MA-1: LOB=0420, Track=042g, Scene=0440, Classification=0430
  • AZ-MA-2: LOB=0460, Track=0480, Scene=046g, Classification=0450
  • AZ-MA-3: LOB=049g, Track=04bg, Scene=04a0, Classification=048g
• Node Positions:
  • AZ-MA-1: 31.6490196, -110.2758537
  • AZ-MA-2: 31.6569236, -110.2659979
  • AZ-MA-3: 31.6637961, -110.2515496