An aerospace and automotive concept defining a technical component, process, or performance measure used in vehicle design, production, or operation. It applies when relevant engineering prerequisites are satisfied and produces defined effects on safety, efficiency, reliability, or manufacturability. It does not ensure outcomes without validated design assumptions and appropriate testing and controls. It materially affects lifecycle performance and cost by influencing design tradeoffs, verification effort, and operational robustness. The concept is generally stable, though methods and standards evolve as technology advances over time.

Principle

Combine sensing, state estimation, decision logic, and task execution in an onboard software/hardware stack; use models, planners, and policies to map perceived state into commands while accounting for constraints, uncertainties, and verification limits to maintain safety and mission goals.

Demonstration

A planetary lander executes hazard detection, trajectory retargeting, and powered descent autonomously during entry because the communication delay prevents ground-in-the-loop control; onboard autonomy identifies safe touchdown zones and aborts or adjusts maneuvers without operator input.

Misapplication

Deploying autonomy with insufficiently validated perception models or inadequate fail-safe behaviors, leading the system to make unsafe decisions in novel environments or to reject valid operator overrides in marginal scenarios.

Consequence

When correctly designed and validated, onboard autonomy reduces latency, enables operations in communication-poor environments, increases mission resilience to anomalies, and scales complex tasks without continuous human supervision.

Reversal

Invert by adopting a ground-in-the-loop paradigm where all critical decisions are deferred to human operators; this increases operator burden, requires continuous communications, and reduces responsiveness to fast or unforeseen events.

Boundary

Refers to decision-making capabilities executed on the platform; excludes delegated supervisory policies or batch-planned operations executed solely on the ground. Practical limits include onboard compute, power, certifiability, explainability, and the requirement for safe fallback to minimal trusted behaviors.

Semantic Tension

Tension between 'autonomy' and 'automation': automation performs predefined sequences under known conditions, while autonomy implies adaptive decision-making under uncertainty; the boundary blurs when systems use advanced automation with limited adaptivity.

Synthesis

Onboard autonomy is an integrated set of onboard perception, planning, and control functions that enables a vehicle to act and react to mission demands and anomalies without immediate human commands, trading some predictability for responsiveness and operational independence.