# From UAV Research to Production Systems

_Published 2026-08-25._

Research and production do not use the same vocabulary all the time, but they keep asking related questions.

The published research covers UAV telemetry security, simulation modelling, vulnerabilities in information-control channels, and machine-learning-based cyberattack detection. The public field records turn those questions into architecture: what crosses the boundary, what gets validated, and what the operator is allowed to believe.

## The path from question to system

The [research page](/research) names the two 2026 publications and their domains. The [UAV field record](/work/uav-resilient-field-systems) shows telemetry passing through validation before persistence while control remains separate. The inference records show a related discipline for model outputs: detections become events with operational context.

A paper is not a deployment report, and a public case study is not a claim that every research result has become a product feature. The useful relationship is methodological.

Simulation can make an attack path inspectable before a field system is exposed to it. A validation rule can turn a concern about replay or message integrity into a decision at the ingest boundary. A detector can be evaluated as part of an event path rather than as a score in isolation.

Research gives the problem a sharper shape. Production gives the shape consequences.