Hi everyone,

I’ve been thinking about a security architecture problem related to autonomous systems operating in adversarial environments.

Many autonomous platforms rely heavily on sensor inputs such as GPS, cameras, radar, or other detection systems. If those inputs become degraded or manipulated (for example through interference or spoofing), the system may still attempt to operate autonomously even though the underlying data is unreliable.

One idea I’ve been exploring is computing a continuous authority value (A ∈ [0,1]) based on several runtime factors such as:

• operator trust level
• mission context confidence
• environmental threat conditions
• sensor trust

The authority value determines which operational tier the system is allowed to execute.

The goal is to prevent situations where autonomy escalates even though sensor trust has degraded or the system is operating in a hostile environment.

From a security engineering perspective I’m curious:

1. Are there established design patterns for gating system authority based on sensor trust?
2. How do security architectures typically handle degraded or adversarial sensor inputs in autonomous systems?
3. Would this type of authority gating normally live inside the control system, or as a separate security layer?

I’d be interested in hearing how people working in security architecture approach this type of problem.