Fleet Architecture & Technical Specs

Aeion compiles real flight missions and streams them to physical drones via Aeion Bridge.

Mission Configuration:

A mission is defined by a drone ID, a package ID, a destination (lat/lng/altitude), and an optional ordered list of intermediate waypoints (lat/lng/altitude).

Dispatch Pipeline:

1. Mission Compilation: Aeion generates the flight mission as both a readable summary AND a real serialized MAVLink v2 binary

1. Bridge Streaming: the mission is serialized to a genuine MAVLink v2 binary (real MISSION_COUNT + MISSION_ITEM_INT frames with a valid CRC) and uplinked via Bridge over 900MHz/2.4GHz radio to a connected Ground Control Station

1. RTK GPS Telemetry: Real-time kinematic GPS returns high-accuracy position data

1. Flight Status Tracking: Mission ID, estimated flight time, status updates

MAVLink Mission Protocol:

• Commands: NAV_WAYPOINT, NAV_TAKEOFF, NAV_LAND, DO_SET_SERVO
• Waypoint queue: List of GPS coordinates with altitude
• VTOL support: Vertical takeoff and landing for urban delivery
• Payload drop: Servo command at destination

Use Cases:

• Last-mile delivery in rural areas
• Medical supply delivery
• Food delivery in campus environments
• Emergency response supplies

Aeion creates virtual boundaries and monitors entry/exit/dwell events against them.

Geofence Types:

• Circle: a center point (lat/lng) plus a radius in meters
• Polygon: an ordered array of lat/lng coordinates describing the boundary

Geofence Settings:

Each geofence can be enabled or disabled independently, and can be configured to trigger on entry, exit, and/or dwell — with a configurable dwell time (seconds inside the boundary before the dwell action fires).

Actions on Boundary Crossing:

• Alert — internal fleet alert
• Notification — push notification
• Webhook — HTTP webhook to your systems
• Task update — update a linked task's status
• Custom — custom handler
• Cyber-physical toggle — toggle a physical device

Geofence Events:

Every boundary crossing produces an event carrying the geofence, the entity (vehicle or driver), the event type (entry / exit / dwell), the location, and — for dwell events — how long the entity stayed inside.

Polygon Point-in-Geofence: A ray-casting algorithm determines whether a GPS point falls inside a polygon boundary.

Aeion integrates with GPS trackers, OBD-II devices, dash cameras, and ELD compliance devices.

Device Types & Providers:

Supported device types: OBD-II, GPS, dashcam, ELD, and custom. Supported providers: Geotab, Samsara, Verizon, and custom integrations. Every registered device reports a status (active / inactive / error / offline), a last-seen timestamp, and firmware version where available.

Telemetry Data:

Each telemetry record carries a timestamp plus, where available: location (lat/lng, accuracy, altitude, heading, speed), engine diagnostics (RPM, speed, fuel level, engine temperature, battery voltage, odometer, engine hours), fault codes (code, description, severity), and driver-behavior flags (harsh braking, harsh acceleration, harsh cornering, speeding, idling).

ELD Compliance (HOS Logs):

Each duty-status entry logs the status (off-duty / sleeper-berth / driving / on-duty), start and end time, duration, location, odometer, and engine hours — the full Hours-of-Service audit trail your DOT auditor needs.

Aeion solves the Vehicle Routing Problem (VRP) against real-world constraints.

Route Stop:

Each stop carries a location, a type (pickup / delivery / service), an optional time window, an expected duration at the stop, a priority, and optional requirements — required driver skills, required vehicle type, and payload capacity needed.

Optimization Constraints:

Routes respect max distance, max duration, max stops per route, vehicle capacity, driver shift-end time, toll avoidance, and whether to prioritize time over distance minimization.

Optimization Algorithm (Nearest Neighbor + 2-opt):

1. Sort stops by priority and time windows
2. Start with nearest stop to depot
3. Greedily add nearest unvisited stop
4. Apply 2-opt local search to improve route
5. Calculate metrics: total distance, duration, cost

Optimized Route Output:

The result is an ordered stop list plus total distance (km), total duration (minutes), total cost (USD), an efficiency score (0-100), and estimated start/end timestamps.

Aeion uses AI to analyze vehicle data and predict maintenance needs before they become breakdowns.

AI Prediction Output:

Each prediction names the affected component (e.g. engine oil, brake pads, transmission), a predicted date, a confidence score (0-100), a severity (critical / high / medium / low), an estimated repair cost, a plain-language reason, and the specific telemetry indicators (value vs. threshold) that triggered it.

AI Pipeline:

1. Gather: maintenance history, usage data, current diagnostics
2. Build a structured request with the vehicle's data
3. Send to your configured AI provider for analysis
4. Parse the structured response (confidence, severity, indicators)
5. Return the resulting maintenance predictions

Maintenance Schedule:

Each scheduled item carries a type (preventive / predictive / reactive), the component, a scheduled date, estimated duration and cost, a priority, and a status (pending / scheduled / in-progress / completed / cancelled).