An introduction to assured positioning, navigation and timing (PNT), navigation integrity and why predictable behaviour matters more than a single position estimate.
Zantasy Defence Systems · 2026 · 6 MIN READ
Assured autonomy is the discipline of keeping an uncrewed system predictable, transparent and safe when the signals it depends on cannot be assumed. In a GNSS-degraded environment — where GNSS may be degraded, denied, unavailable or inconsistent with other navigation sources — that discipline starts with navigation.
The question for a mission system is not simply “where am I?” It is “how much should I trust that answer right now, and what should I do if the answer changes?” This explainer sets out the ideas behind assured positioning, navigation and timing (PNT) as Zantasy approaches them: integrity monitoring, confidence states and authorised behaviour.
Definitions
01
Position is not the same as confidence
A navigation solution that reports a single position estimate, without any measure of how that estimate was produced, leaves the rest of the autonomy stack guessing. Two identical position outputs can carry very different levels of trust depending on which sources contributed, how recently they agreed, and whether any of them is under stress.
Assured PNT treats integrity as a first-class output. Sources are monitored continuously, their health is assessed, and the navigation layer reports not only a best estimate but a confidence state that downstream functions can reason about.
Predictable behaviour matters more than a single position estimate.
02
When sources disagree
In contested environments, disagreement between sources is normal rather than exceptional. GNSS may drift while inertial and terrain-referenced inputs remain coherent, or a single source may begin to diverge from the others. The first task of a navigation-integrity layer is to detect that divergence early and say so plainly.
Zantasy describes these conditions as confidence states: agreement, when sources corroborate one another; watch, when divergence is detected and confidence is reduced; and conflict, when sources materially disagree and the system must stop trusting the contested input. Each state is a signal to the mission layer, not a silent failure.
03
Degrading into authorised behaviour
Detecting a problem is only useful if the system has somewhere safe to go. In an assured architecture, confidence states map to authorised behaviours defined before the mission: continue, hold, re-plan, return, or hand control back to the operator. The system degrades gracefully along a path the command chain has already approved.
Just as importantly, the system records what happened. Evidence of source behaviour, confidence transitions and the resulting actions gives programme teams something they can assess — and gives assurance activities a factual basis rather than a narrative one.
Conclusion
Assured autonomy in a GNSS-degraded environment is not a promise that signals will never fail. It is an engineering commitment that the system will recognise uncertainty, behave predictably when it does, and produce the evidence needed to trust it again. That is the layer Zantasy builds.