ADR-0003: How integration beginsLink to the section ADR-0003: How integration begins
On this page
Status: Accepted 2026-10-14 — sponsor confirmed the default ("keep all") Date: 2026-10-14 Profiles affected: all
ContextLink to the section Context
Adoption can start from the drone side (manufacturers implement UNIVIP in firmware), from the middle (gateways translate existing APIs), or both. Manufacturer firmware roadmaps are slow and commercially guarded; gateways can be built by anyone with API access.
Options consideredLink to the section Options considered
- SDK/controller/gateway first, vendor neutral. Fastest path to a demonstrator; proves value to VMS vendors; does not depend on manufacturer cooperation. Risk: gateways can only expose what the underlying API exposes.
- Directly with drone manufacturers. Highest fidelity; strongest endorsement. Risk: slow, and manufacturers may prefer their own ecosystems.
- Both simultaneously. Maximum coverage; splits a small team's effort.
DecisionLink to the section Decision
Option 1 as the critical path, with manufacturer engagement in parallel but never as a dependency. The reference implementation is a gateway with pluggable adapters (MAVLink first, then DJI Cloud API and/or Skydio). The specification treats native, controller-based and gateway-based modes as indistinguishable to the client, so a manufacturer who later implements UNIVIP natively replaces the gateway without any VMS change.
ConsequencesLink to the section Consequences
- The capability document must express partial support honestly: a gateway reports exactly the subset the underlying API allows.
- Mapping tables (ONVIF, MAVLink, MISB) are normative annexes so gateway authors converge.
- The roadmap targets a MAVLink/PX4-SITL adapter first because it needs no vendor agreement.
Open questionsLink to the section Open questions
- Which second adapter (DJI or Skydio) to build first; decided by partner availability.