Dead reckoning is a necessary part of degraded navigation. The problem begins when an estimated track starts to inherit the visual authority of a trusted live fix.
That can happen quickly on screen. The position source weakens or drops, yet the route continues in a tidy line that feels more certain than it should.
What the scene shows
In the anomaly demo, a stable transit begins under live GNSS. Then the fix degrades and disappears, while the route continues as an estimated dashed track.
The visual is useful because it shows two different states clearly:
- trusted live navigation
- continued movement based on estimation rather than fresh position truth
That distinction matters operationally.
Why this situation is risky
An estimated track can still be useful, but it should not be mistaken for a fully trusted current fix.
If the presentation is too clean, the bridge may keep treating the picture as normal for longer than it should. The risk is not just data loss. The risk is overconfidence in inferred continuation.
This becomes especially important when the vessel is operating where margin is tight and decisions need to be made quickly.
Visual continuity can hide trust loss
One of the hardest parts of degraded navigation is that the screen may still appear calm and orderly. The route continues. Movement still seems explainable. The display still gives the impression of continuity.
But continuity after fix loss is not the same thing as renewed certainty.
A good monitoring system should help make that distinction clear, so crews understand when they are working with estimation rather than trusted live position.
The practical takeaway
Dead reckoning is valuable, but it should remain visibly separate from trusted live navigation. When estimated motion starts to look too neat, the display can unintentionally create more confidence than the underlying situation supports.
GeoWatch is built to make those changes in trust state easier to see, while keeping the event context and operator response tied to the same onboard workflow.