0007-Uncrew Industrial Operations Platform

Andi Lamprecht ·2026-05-11· 16 min read· Draft

Decision Log

Problem Statement

Uncrew today is a mission planner and ground control station (GCS) that activates once a flight is requested with a start point, waypoints and objectives. There is no upstream orchestration layer for industrial customers to schedule inspections, trigger on-demand flights via API, manage recurring flight programs or monitor dock infrastructure. Additionally, mission planning lacks camera/gimbal awareness for inspection use cases and there is no pipeline for routing surveillance data to customer cloud environments.

Importance of Solving the Problem Now

DroneUp has a near-term opportunity to license Uncrew as a standalone product to industrial drone operators who need to scale from pilot programs to 100+ sites operating 24/7. The first such customer (Horizon Aerobotics) projects 100 sites in 2026 for a single railroad customer (Norfolk Southern), with a combined demand profile of 100,000-200,000 flight hours annually across all their customers. Horizon has stated that software is the least mature part of their portfolio and they are ready to stop investing in internal flight management software in favor of a third-party solution.

Without an orchestration and scheduling layer, Uncrew cannot serve this market. Missions must be manually created one at a time, there is no API for external system integration, no dock management capability and no way to deliver inspection data to customers. These gaps make Uncrew unsuitable for industrial operations at scale.

Why Should the Business Pursue This?

• New revenue stream: Variable cost licensing model (per flight hour) with high-volume industrial customers
• Platform expansion: Moves Uncrew from a single-use-case tool (delivery) to a multi-vertical platform (inspection, ISR, infrastructure monitoring)
• Market timing: Horizon alone projects 100 NS sites in 2026; CSX, BNSF and Union Pacific are forming a consortium that would adopt the same solution
• Competitive positioning: Industrial operators need BVLOS-capable, scalable fleet management software; few competitors offer this as a licensable product

Concept Overview

Build a platform-level orchestration and scheduling capability on top of Uncrew, along with dock management, camera-aware mission planning and a surveillance data passthrough pipeline. All capabilities are generic and multi-tenant.

Capabilities Summary

Target Audience

Customer Operations Manager

Works for the industrial customer (e.g. railroad company). Schedules recurring inspections, triggers on-demand flights, reviews mission reports and usage data. Interacts via the Uncrew UI and/or API.

Customer Integration System

Automated system (e.g. rail yard management software) that triggers flights via API when an event occurs (train departing, asset needs inspection). No UI interaction; pure API consumer.

RPIC (Remote Pilot in Command)

Pilot operating from a Remote Operations Center. Monitors scheduled missions, intervenes when needed. Uses the Uncrew GCS. Benefits from camera/gimbal status during inspection flights.

Fleet / Operations Manager

Manages dock fleet health, mission schedules across multiple sites and overall operational readiness. Uses the Uncrew UI.

Product Requirements

Epic 1: Customer Scheduling and Orchestration

US-1.1: Recurring inspection schedules As a Customer Operations Manager, I want to create a recurring inspection schedule (e.g. “inspect tracks 1-40 every 6 hours”) so that routine missions execute automatically without manual intervention.

• Schedule supports cron-like recurrence (hourly, daily, weekly, custom intervals)
• Each scheduled instance generates an Uncrew mission with pre-defined waypoints and objectives
• Schedule can be paused, modified or canceled without affecting in-flight missions
• Missed schedules (e.g. due to weather hold) are flagged in the UI with reason

US-1.2: On-demand mission API As a Customer Integration System, I want to request an on-demand mission via REST API with a mission template ID, priority and optional parameters so that external systems can trigger flights programmatically.

• API accepts mission template reference, priority (routine/urgent) and parameter overrides (e.g. specific track number)
• Response includes mission ID, estimated launch time and status polling endpoint
• Urgent requests are queued ahead of routine scheduled missions
• API is authenticated via OAuth2/API key, scoped to the customer’s tenant
• API contract is generic with no customer-specific fields or logic

US-1.3: Mission templates As a Customer Operations Manager, I want to define mission templates (reusable flight plans with objectives) so that recurring or on-demand missions do not require re-planning each time.

• Template includes: launch/land location (dock), waypoints, altitude, speed, camera objectives, expected duration
• Templates are versioned; editing creates a new version, existing schedules reference a specific version
• Templates can be cloned and modified

US-1.4: Operations calendar view As a Fleet/Ops Manager, I want to see a calendar/timeline view of all scheduled missions across all sites so that I can manage operational capacity.

• View shows scheduled, in-progress, completed and failed missions on a timeline
• Filterable by site, drone, status and date range
• Conflicts (overlapping missions on the same dock/drone) are visually flagged

US-1.5: Mission failure notifications As a Customer Operations Manager, I want to receive notifications when a scheduled mission cannot execute (weather, drone unavailable, dock offline) so that I can take corrective action.

• Notifications via webhook (for system integrations) and in-app (for UI users)
• Notification includes reason, affected schedule and suggested next window

Epic 2: Drone-in-a-Box / Dock Management

US-2.1: Real-time dock status As a Fleet/Ops Manager, I want to see real-time dock status (online/offline, battery level, weather conditions, drone readiness) so that I know which sites are operationally ready.

• Dashboard shows all docks with status indicators (green/yellow/red)
• Status includes: connectivity, battery charge level, HVAC status, last successful launch, drone present/absent
• Status refreshes at 30-second intervals or better

US-2.2: Dock readiness gating As the orchestration system, I need to query dock readiness before dispatching a mission so that missions are only sent to operational docks.

• Dock exposes a readiness API (or Uncrew polls dock status)
• Orchestration layer checks dock readiness before mission dispatch
• If dock is not ready, mission is held with reason logged

US-2.3: Dock health history As a Fleet/Ops Manager, I want to see dock health history and maintenance alerts so that I can proactively manage the fleet.

• Historical uptime/downtime per dock
• Alerts for anomalies (repeated failed launches, battery degradation trend, connectivity drops)

US-2.4: Standardized dock adapter interface As the Uncrew platform, I need a standardized dock communication protocol so that we can integrate with any manufacturer’s dock hardware.

• Dock integration is abstracted behind an adapter interface
• First implementation supports Horizon’s dock API/protocol
• Adding a new dock manufacturer requires only a new adapter, no core platform changes

Epic 3: Camera/Gimbal-Aware Mission Planning

US-3.1: Inspection objectives in mission planning As a Customer Operations Manager, I want to define inspection objectives (what to look at, from what angle, at what altitude) and have the mission planner optimize the route accordingly.

• Objectives specify: target GPS coordinates or linear asset (track segment), required camera angle (nadir, oblique, side-facing), minimum resolution/altitude
• Route optimizer generates waypoints that satisfy all objectives with minimal flight time
• Preview shows planned camera footprint overlaid on map

US-3.2: Gimbal monitoring during flight As an RPIC, I want to see the planned gimbal angles and camera targets during mission monitoring so that I can verify the drone is capturing the right data.

• GCS displays planned vs. actual gimbal position during flight
• Deviations from planned camera angles trigger alerts

US-3.3: Aircraft camera profiles As the mission planner, I need to know the drone’s camera/gimbal specs (FOV, tilt range, sensor resolution) to compute optimal waypoints.

• Camera/gimbal specs are part of the aircraft profile in fleet configuration
• Different aircraft types can have different camera profiles
• Mission planner uses camera profile to calculate ground sample distance and coverage

Epic 4: Surveillance Data Passthrough Pipeline

US-4.1: Data delivery to customer cloud As a Customer Operations Manager, I want inspection data (inference results, detection thumbnails, metadata) delivered to my cloud storage automatically after each mission.

• Data is routed to a customer-configured cloud destination (S3 bucket, GCS bucket or Azure Blob; at least one supported at MVP)
• Uncrew does not persistently store the customer’s inspection data; passthrough only
• Delivery confirmation is logged per mission

US-4.2: Data delivery webhook As a Customer Integration System, I want to receive a webhook notification when mission data is available so that my downstream processing pipeline can start.

• Webhook fires with mission ID, data location URI and summary metadata (flight duration, detection count, timestamp range)
• Webhook is configurable per customer tenant

US-4.3: Data delivery status As a Fleet/Ops Manager, I want to see data delivery status per mission (pending, delivered, failed) so that I can troubleshoot pipeline issues.

• Mission detail view shows data delivery status
• Failed deliveries can be manually retried

Epic 5: Usage Metering Dashboard

US-5.1: Customer-facing usage metrics As a Customer Operations Manager, I want to see my usage metrics (flight hours, mission count, data volume delivered) in the Uncrew UI so that I can track consumption against my contract.

• Dashboard shows current billing period totals and historical trends
• Filterable by site, date range and mission type
• Data sourced from existing metering backend (no new data capture needed)

US-5.2: Usage export As a Fleet/Ops Manager, I want to export usage reports (CSV/PDF) for internal billing and contract reconciliation.

• Export includes per-site, per-day breakdown of flight hours and mission counts
• Available via UI download and scheduled email delivery

Constraints

Explicitly Out of Scope

• M:N remote ops and pilot failover – separate PRD exists
• 44807 exemption support for HX1 – separate engagement (~$200k, 19 weeks)
• Part 91/107 regulatory pathway work
• Horizon’s ML inference models – they own their edge processing
• Dock hardware design – manufacturer builds the dock; we integrate
• Raw video streaming – only inference results and thumbnails pass through
• Customer-specific API contracts – all APIs are generic and multi-tenant
• Mobile application – UI is web-based (Apollo frontend) only
• Hubops delivery workflows – food delivery, Walmart, etc. are not affected

Technical Constraints

• New orchestration service must be built in Go (per org convention)
• All customer-facing APIs authenticated via existing Frontegg tenant auth (OAuth2/JWT)
• API keys available for system-to-system integration
• Dock communication encrypted in transit; dock adapter must validate dock identity
• Surveillance data encrypted in transit (TLS); no persistent storage in Uncrew
• Tenant isolation enforced at API, database and data pipeline layers

Operational Environment

Architecture Overview

Customer System --> [Orchestration API] --> [Mission Templates] --> [Uncrew Missions Service]
                         |                        |                         |
                    [Scheduler]              [Camera Planner]          [Mavlink Shim]
                         |                        |                         |
                    [Dock Manager] <---- [Dock Adapter] <---- [Physical Dock]
                         |                                          |
                    [Usage Metering]                          [Drone + Camera]
                                                                    |
                                                          [Data Passthrough] --> [Customer Cloud]

Data Flow:

1. Customer (UI or API) creates a schedule or triggers an on-demand mission referencing a mission template
2. Orchestration service resolves the template, checks dock readiness and queues the mission
3. Camera-aware planner optimizes the route based on inspection objectives and aircraft camera profile
4. Mission is dispatched to Uncrew Missions Service, then to Mavlink Shim, then to the drone
5. During/after flight, inference data flows from drone through Uncrew data pipeline to customer’s cloud bucket
6. Usage metrics are recorded and surfaced in the dashboard

Affected Repos

New Data Entities

Modified Entities

• mission (missions-service): add optional camera_objectives field and template_ref
• aircraft_profile (fleet config): add camera/gimbal spec (FOV, tilt range, sensor resolution)

Regulatory and Compliance

• No DO-178C/DO-278A impact – all capabilities are cloud-based software, not onboard safety-critical code
• No direct FAA filing required – mission planning rules (altitude limits, geofencing, flight time) are configuration driven by the customer’s operational approvals
• Data handling: Surveillance data passthrough architecture must ensure Uncrew does not retain customer inspection data beyond the transit window. Customer data retention policies apply at their cloud destination. Data in transit must be encrypted (TLS). If any transient caching is needed, define a maximum TTL of 24 hours with automatic purge.
• Multi-tenant isolation: Scheduling, mission templates, dock assignments and usage data must be strictly tenant-isolated. No cross-customer data leakage.
• Audit logging: All API-triggered missions and schedule changes must be logged with actor, timestamp and action for operational accountability.

Risks

Timeline

Phase 0: Discovery (2-3 weeks)

• Investigate Horizon’s dock communication protocol and document the interface
• Audit existing metering data completeness
• Prototype camera-objective route planning with a simple inspection scenario
• Define OpenAPI spec for orchestration API (contract-first design)
• Validate HX1 MAVLink gimbal command compatibility

Phase 1: MVP

• Orchestration service: mission templates, on-demand API-triggered missions, basic scheduling (fixed interval)
• Dock management: status polling, readiness checks (single adapter for Horizon dock)
• Apollo UI: mission template builder, dock status dashboard, basic schedule creation
• Camera planning: camera objectives in mission spec, simple waypoint ordering based on target locations
• Data passthrough: inference results to single cloud destination (GCS bucket)
• Metering dashboard: flight hours and mission counts (read from existing data)

Phase 1.1: Fast Follow

• Cron-like advanced scheduling (complex recurrence, blackout windows)
• Webhook notifications (mission status, data delivery, schedule failures)
• Usage export (CSV/PDF)
• Multi-cloud data destinations (S3, Azure Blob)
• Dock health history and maintenance alerts

Phase 2: Full Vision

• Camera-aware route optimization (minimize flight time while covering all objectives)
• Multi-dock site orchestration (select best dock for a mission based on proximity and readiness)
• Scheduled email reports
• Second dock manufacturer adapter (proving the abstraction)
• Calendar/timeline view with conflict detection

Dependencies

Success Criteria

1. A prospective industrial customer can sign an MOU based on demonstrated scheduling, dock management and inspection planning capabilities
2. First customer flies an inspection mission end-to-end on Uncrew (from scheduled request through mission execution to data delivery to their cloud)
3. Platform supports 100 or more sites with recurring scheduled missions and on-demand API-triggered flights
4. Aircraft-to-operator ratio of 5:1 or better supported by the orchestration layer (missions auto-dispatched, pilot assignment handled by existing M:N system)
5. Customer-facing usage dashboard surfaces flight hours, mission counts and billing-relevant metrics without manual data pulls

Estimation Input

prd_sizing_input:
  feature: "Uncrew Industrial Operations Platform"
  scope_clarity: "B"
  key_terms:
    - "orchestration"
    - "mission template"
    - "scheduling"
    - "dock management"
    - "camera gimbal mission planning"
    - "surveillance data passthrough"
    - "usage metering dashboard"
  risk_flags:
    - "new_service"
    - "external_integration"
    - "multi_repo"
    - "api_design"
    - "infrastructure"
  affected_repos:
    - "NEW-orchestration-service"
    - "uncrew-missions-service"
    - "uncrew-apollo-frontend"
    - "uncrew-api-private"
    - "uncrew-uav-camera-capture"
    - "uncrew-mavlink-shim"
    - "uncrew-pathfinding"
  domains:
    - "backend-go"
    - "frontend-react"
    - "firmware-cpp"
    - "python"
    - "protobuf"
  regulatory: false
  discovery_needed: true

Appendix A: Customer Context – Horizon Aerobotics