Solution · PredictiveIQ Edge

Deploy Physics AI on platform and at platform.

PredictiveIQ Edge turns live or offloaded vehicle data into health, remaining useful life and diagnostic evidence, then carries it through VHMS: live on the vehicle for crew awareness, and at the platform for maintainer workflows.

PredictiveIQ Edge compute for on platform and at platform Physics AI inference
Why condition

A calendar cannot see how the asset has actually been used.

Two nominally identical assets accumulate different degradation depending on load, environment, duty cycle and prior damage. PredictiveIQ Edge estimates current condition and remaining useful life from how each asset has actually operated, then delivers that judgment in two operating contexts: live on the vehicle for crews, and at the platform for maintainers. One product, one validated package, distinguished only by where the model runs and who needs the answer.

Calendar-based

  • Time
  • Mileage
  • Operating hours

Scheduled maintenance

Based on condition

  • Operating history
  • Current HI / CI
  • Degradation

Better informed maintenance window

Remaining useful life

From current condition to the intervention region.

Remaining useful life is tied to current condition and assumed future operation, not a universal countdown. Every estimate carries the assumptions behind it, so crews and maintainers know what the number depends on.

Health and remaining useful life trajectory from current condition to the intervention region
On platform deployment

Crew / Operator

Live vehicle data for continuous awareness.

On vehicle compute, a VHMS-linked prognostic module consumes live sensor streams so the crew sees health state, developing anomalies and remaining useful life with stated confidence, without waiting for connectivity or post-mission analysis.

  • VHMS-linked module on vehicle edge compute
  • Live sensor and telemetry streams
  • Continuous health and anomaly awareness
  • HI / CI and RUL with stated confidence
  • Built for crew and operator decisions
On platform · VHMS
Crew station with live vehicle health awareness through PredictiveIQ Edge
Live dataHI / CIRULCrew awareness
At platform deployment

Maintainer

Offloaded data for maintainers at the platform.

At the maintenance platform, that same module works from offloaded or available data to support PMCS, fault isolation and deeper diagnostic review. Condition and remaining life carry the same meaning the crew saw, so a maintenance window can be planned around evidence rather than an interval.

  • Same package, different data path
  • Offloaded logs and available mission data
  • HI / CI and RUL with stated confidence
  • Dynamic PMCS and troubleshooting support
  • Workflows for maintainers and depot teams
At platform · Maintainer
Maintainer reviewing offloaded vehicle health data at the platform
Offloaded dataPMCSTroubleshootingMaintainer
Diagnostics and anomaly detection

A system can be abnormal before a single sensor crosses a limit.

Developing problems often appear first as a change in the relationship between signals, operating conditions and expected behavior. Physics AI models what the system should be doing, compares it with what it is doing, and adds that context next to the alarms already in place. The same comparison runs in both profiles: crews see it on platform, maintainers work it at platform.

1 · Inputs

Operating conditions

+

Sensor / telemetry data

+

Engineering context

2 · Compare

Expected system behavior

vs

Observed system behavior

Residual / difference across signals and operating state

3 · Classify

Normal

No action required

Anomalous

Investigate further
4 · Isolate

Failing subsystem or component

·

Affected signals

·

Likely failure mode

5 · Communicate

Crew / operator on platform

·

Maintainer at platform

Diagnostic alert through VHMS with fault evidence and context

What runs

A validated Physics AI package near the asset.

PredictiveIQ Edge executes compact model packages produced by PredictiveIQ Model Factory on the compute already available around the vehicle or platform.

What goes in

Operational inputs

Sensor measurements, telemetry, operating conditions and derived signals.

What runs

A compact model

A Physics AI model optimized for the target hardware and runtime environment.

What comes out

Decision context

Health index, anomaly score, fault information, degradation estimate, remaining useful life and model confidence.

Where it runs

Near the asset

Vehicle compute, industrial PC, gateway or edge system at the site, plus hybrid exchange to enterprise systems when approved.

The same runtime carries whichever degradation the program cares about, because the physics and the health meaning travel inside the package from PredictiveIQ Model Factory.

Structural fatigueBearing degradationGearbox wearLubricant conditionBattery healthEngine and powertrain

Tell us where the model needs to run: on platform or at platform.

Contact us