Controller truth
PLC and OPC data can be safety-relevant, but only if the tags, states, and timing are interpreted carefully.
Verification Method
AviRob starts where industrial reality actually starts: mixed source families, ugly raw signals, and requirements that already exist before verification begins. The hard part is not just ingestion. It is inspecting raw source truth, assigning meaning carefully, and keeping that meaning governed as the site evolves.
Installed and on-prem deployment patterns are supported for industrial environments where source data needs to stay local.
Industrial edge truth is rarely clean. AviRob starts from the sources teams actually have, not the sources they wish they had.
PLC and OPC data can be safety-relevant, but only if the tags, states, and timing are interpreted carefully.
VDA/FMS and orchestration streams add mission and coordination context, but they do not replace direct proof.
Folders, mixed exports, MCAP, rosbag, CSV, and text logs are often the fastest realistic entry point.
Aliases, partial IDs, code tables, and register values often need review before anyone should treat them as canonical.
Some inputs explain or correlate behavior, but they are not strong enough on their own to justify a verdict.
Derived or synthetic views can accelerate review, but AviRob keeps them explicit so reviewers know what they depend on.
AviRob keeps direct, diagnostic, derived, and synthetic evidence distinct so weaker lanes are not silently upgraded into verdict truth. If a required dependency is missing, the result stays NOT_VERIFIABLE.
Use the strongest available source when a requirement needs direct proof from the safety-relevant layer.
Diagnostics can explain what happened, but they are not silently upgraded into direct proof.
Derived views stay tied to their source dependencies so reviewers can see what the claim rests on.
Synthetic or simulated views can accelerate workflow, but they stay explicit and are not mistaken for direct proof.
FAT/SAT is an entry point. The maintain-phase wedge is what makes the lane durable.
Rerun checks after robot updates, site changes, fleet growth, or changed operating assumptions.
See when an old proof package no longer matches the site reality you are now operating.
Find repeated issues, hotspots, and drift without dropping back to raw-log archaeology every time.
Bring one representative source bundle, one trust boundary, and one deployment question you want answered without hand-waving.