Vehicle autonomy

Autonomy in defined increments, each independently useful and independently priced

Capability levels

From manned to conditional autonomy

Autonomy capability levels L0 through L4: manned, teleoperated, assisted, supervised, and conditional.

L0

Manned

Baseline vehicle. Kit installed, no driving automation.

Operator in vehicle

L1

Teleoperated

Remote driving over encrypted link. Continuous human control of steering, throttle, brake.

Operator in the loop

L2

Assisted

Path keeping, obstacle warning, speed governing, return-to-safe on link loss.

Operator in the loop

L3

Supervised

Waypoint navigation, follow-me, convoy keeping, obstacle avoidance within a defined operating domain.

Operator on the loop

L4

Conditional

Mission execution within a defined operational design domain. Degraded-GNSS navigation. Human on call.

Operator on call

Full unsupervised autonomy is outside our current scope. We will not represent a capability we cannot document a safety case for.

Autopilot platform architecture

A layered stack with an independent safety monitor and a human supervision path.

Autopilot platform architecture: sensing, perception, localization, world model, planning, and control, with an independent safety monitor and an operator supervision path.

Sensing

Lidar, radar, EO/IR, inertial, wheel odometry, GNSS where available.

Perception

Obstacle detection, terrain classification, traversability assessment.

Localization

GNSS-degraded and denied positioning. Visual and inertial dead-reckoning.

World model

Fused local map, tracked objects, confidence and staleness scoring.

Planning

Route, path, and behaviour planning within the operational design domain.

Control

Actuation to steering, throttle, braking, and transmission via the platform bus.

Safety monitor

Independent of the planning stack. Enforces geofence, speed, and standoff limits, and commands safe-stop on fault, link loss, or domain exit.

Operator supervision

Single console across ground and air assets. Take-over at any level, full telemetry logging, and replayable mission record for after-action review.

Every level of automation retains a human take-over path and a logged, replayable record.

How an autonomy program runs

Five phases. Fixed-price. Each gated on acceptance of the last.

0

Scoping study

Concept of operations, Interface Control Document, power and thermal budget, sensor architecture, safety framework, firm price for the next phase.

6–8 weeks

1

Demonstrator

Integrated mission kit on a donor platform. Teleoperated or optionally-manned. Field demonstration to end users.

90 days

2

Supervised autonomy

Waypoint, follow-me, convoy, obstacle avoidance. Operator on the loop. Degraded-GNSS navigation.

12–18 months

3

Production engineering

Line integration, qualification, configuration control at rate, technician training and certification.

9–12 months

4

Conditional autonomy

Operational design domain expansion, safety case, verification, fielding support.

Scoped after Phase 2

Development is customer-funded and milestone-billed. A113 retains its background and developed intellectual property and licenses it by field of use.

Next step

Start with a scoping study

Six to eight weeks, fixed price. You receive the concept of operations, the Interface Control Document, the electrical and thermal budget against your actual platform, a safety framework, and a firm price for the build.

If the study does not earn the next phase, you have learned that for the cost of a study.

Request a scoping study