piotr/groundstation
1
0
Fork 0
Code Issues 9 Pull Requests 1 Actions Packages Projects Releases Wiki Activity

EPIC: Resilient flight telemetry and table-driven adaptive WFB link #20

New Issue
Open
opened 2026-11-03 22:39:16 +01:00 by hermes · 0 comments
hermes commented 2026-11-03 22:39:16 +01:00
Collaborator
Copy Link
Copy Source

Outcome

Deliver a field-capable Ground Station with independent OSD, receiver-link metrics, resilient FC flight telemetry, and Ground-Station-owned adaptive WFB control. Prioritize this work over kiosk stick navigation and touchscreen/multi-display UI.

Confirmed telemetry risk

The concern is valid. ExpressLRS is asymmetric: many small receivers transmit telemetry at the RF chip limit (about 13 dBm / 20 mW) while the handset uplink may use hundreds of mW or more. Official ELRS guidance recommends receiver PA hardware for long telemetry range. Therefore telemetry can be lost before RC control, especially on tiny-whoop/AIO receivers without a PA.

Telemetry ratio changes capacity/interval, not the receiver's fundamental link budget. Lower ELRS air rate, appropriate telemetry ratio, external antenna/diversity and a PA receiver can help, but cannot be assumed for every model.

Chosen alternate path

Use a separate, robust, low-rate WFB downlink stream for FC telemetry:

FC MSP → persistent VTX msposd/parser → versioned telemetry packet
→ dedicated WFB stream with robust MCS/FEC
→ Ground Station FlightTelemetry snapshot → OSD/logging

Do not embed telemetry in RTP/video. A separate stream can survive after high-bitrate video becomes unusable and remains independent of decoder/window state. It still shares the 5.8 GHz RF path, so it is not absolute redundancy; a future physically independent 868/915 MHz telemetry radio remains an option for missions requiring RF-path independence.

Execution order

  1. Current: #3 — OSD state model and transparent renderer.
  2. Next: #6 — read-only WFB receiver metrics for link quality.
  3. Alternate FC telemetry transport:
  4. Table-driven adaptive control:
  5. Optional handset bridge: piotr/edgetx-gx12#4

Explicitly deferred

  • #4 kiosk navigation using radio sticks — skip for now.
  • #5 multi-display/touchscreen UI — architecture remains valid, implementation follows adaptive/telemetry core.
  • touchscreen configuration UX — use config/profiles for initial table work.
  • analog receiver profiles — separate later milestone.

Alternative options considered

  1. Improve native ELRS telemetry: choose RX with PA/LNA, external antenna/diversity, lower air rate and suitable telemetry ratio. Best zero-software option but aircraft-specific and cannot be retrofitted onto every AIO whoop cheaply.
  2. WFB dedicated telemetry stream (chosen): no extra RF hardware, encrypted/authenticated, independent MCS/FEC and stream ID, likely survives beyond usable video because payload is tiny. Not independent of the 5.8 GHz RF path.
  3. Dedicated second RF telemetry link: 868/915 MHz ELRS AirPort, SiK/LoRa or similar. Strongest path diversity but adds radio, antenna, weight, power and spectrum planning. Reserve for future long-range aircraft, not Pavo/tiny-whoop MVP.
  4. Telemetry in video frames/OSD: rejected as primary machine-readable transport; it disappears with video/decoder and is difficult to validate/log.

Epic acceptance

  • OSD separates requested/applied/observed/stale states;
  • FC battery/GPS/arm data remains available on GS after native ELRS telemetry loss while WFB telemetry is alive;
  • receiver metrics drive only GS policy;
  • VTX executes acknowledged table levels and never chooses adaptation;
  • adaptive starts log-only, then follows field-reviewed thresholds;
  • video latency/render path never waits for telemetry or policy control.
## Outcome Deliver a field-capable Ground Station with independent OSD, receiver-link metrics, resilient FC flight telemetry, and Ground-Station-owned adaptive WFB control. Prioritize this work over kiosk stick navigation and touchscreen/multi-display UI. ## Confirmed telemetry risk The concern is valid. ExpressLRS is asymmetric: many small receivers transmit telemetry at the RF chip limit (about 13 dBm / 20 mW) while the handset uplink may use hundreds of mW or more. Official ELRS guidance recommends receiver PA hardware for long telemetry range. Therefore telemetry can be lost before RC control, especially on tiny-whoop/AIO receivers without a PA. Telemetry ratio changes capacity/interval, not the receiver's fundamental link budget. Lower ELRS air rate, appropriate telemetry ratio, external antenna/diversity and a PA receiver can help, but cannot be assumed for every model. ## Chosen alternate path Use a separate, robust, low-rate **WFB downlink stream** for FC telemetry: ```text FC MSP → persistent VTX msposd/parser → versioned telemetry packet → dedicated WFB stream with robust MCS/FEC → Ground Station FlightTelemetry snapshot → OSD/logging ``` Do not embed telemetry in RTP/video. A separate stream can survive after high-bitrate video becomes unusable and remains independent of decoder/window state. It still shares the 5.8 GHz RF path, so it is not absolute redundancy; a future physically independent 868/915 MHz telemetry radio remains an option for missions requiring RF-path independence. ## Execution order 1. **Current:** #3 — OSD state model and transparent renderer. 2. **Next:** #6 — read-only WFB receiver metrics for link quality. 3. **Alternate FC telemetry transport:** - VTX publisher: https://git.piksar.eu/piotr/openipc-builder/issues/4 - GS receiver/OSD merge: https://git.piksar.eu/piotr/groundstation/issues/19 4. **Table-driven adaptive control:** - #16 policy table generator/model; - #17 reliable table provisioning/identity; - EdgeTX runtime level: https://git.piksar.eu/piotr/edgetx-gx12/issues/3 - VTX executor: https://git.piksar.eu/piotr/openipc-builder/issues/3 - #18 adaptive level controller, first log-only then live. 5. **Optional handset bridge:** https://git.piksar.eu/piotr/edgetx-gx12/issues/4 ## Explicitly deferred - #4 kiosk navigation using radio sticks — skip for now. - #5 multi-display/touchscreen UI — architecture remains valid, implementation follows adaptive/telemetry core. - touchscreen configuration UX — use config/profiles for initial table work. - analog receiver profiles — separate later milestone. ## Alternative options considered 1. **Improve native ELRS telemetry:** choose RX with PA/LNA, external antenna/diversity, lower air rate and suitable telemetry ratio. Best zero-software option but aircraft-specific and cannot be retrofitted onto every AIO whoop cheaply. 2. **WFB dedicated telemetry stream (chosen):** no extra RF hardware, encrypted/authenticated, independent MCS/FEC and stream ID, likely survives beyond usable video because payload is tiny. Not independent of the 5.8 GHz RF path. 3. **Dedicated second RF telemetry link:** 868/915 MHz ELRS AirPort, SiK/LoRa or similar. Strongest path diversity but adds radio, antenna, weight, power and spectrum planning. Reserve for future long-range aircraft, not Pavo/tiny-whoop MVP. 4. **Telemetry in video frames/OSD:** rejected as primary machine-readable transport; it disappears with video/decoder and is difficult to validate/log. ## Epic acceptance - OSD separates requested/applied/observed/stale states; - FC battery/GPS/arm data remains available on GS after native ELRS telemetry loss while WFB telemetry is alive; - receiver metrics drive only GS policy; - VTX executes acknowledged table levels and never chooses adaptation; - adaptive starts log-only, then follows field-reviewed thresholds; - video latency/render path never waits for telemetry or policy control.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
No items
No labels
Milestone
No items
No Milestone
Projects
Clear projects
No projects
Assignees
Clear assignees
grzesiu hermes piotr (Piotr Sarna)
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: piotr/groundstation#20