Real response, not guesswork
Roll, pitch, heave and acceleration are tied to loading, speed, heading and sea context.
Motion Digital Twin · Trained by real voyages
A vessel model
that learns at sea.
ShipMotionIQ turns real vessel response into a living Motion Digital Twin: reusable for simulation, API access, routing workflows, cargo limits and signed voyage evidence.
Measure
Real hull response, loading and sea context
Model
Any critical point in vessel coordinates
Share
Owner-controlled API, exports and evidence
What makes it different
Real response becomes a model advantage.
The product is built around one asset: a vessel-specific Motion Digital Twin that improves as measured voyages arrive and can be reused outside the dashboard.
Any point from one frame
One measured body frame lets the model resolve cargo, bridge, gangway, crane hook or moonpool points.
Sister-vessel learning
Similar hulls can shorten cold starts while each vessel keeps its own calibration and confidence.
Owner-controlled access
Versioned model outputs can be exposed to routing, cargo, offshore and engineering workflows.
Cloud-native vessel architecture
A shipboard node feeds an Azure learning loop.
The MVP Azure foundation is already in place. The product architecture is designed to collect maritime IoT telemetry, validate it in the cloud and turn measured voyages into versioned model updates.
01 / ONBOARD NODE
Motion IoT edge kit
A low-footprint node anchors the vessel body frame and keeps collecting when connectivity is limited.
Roll, pitch, heave and acceleration
GNSS, AIS/NMEA, speed and heading where available
Loading, draft, cargo and sea-state context
02 / AZURE INGESTION
IoT Hub + cloud state
Cloud-native services keep device identity, raw batches, operational state and observability separated.
Azure IoT Hub for secure device ingress
Blob Storage for raw and aggregate voyage data
Cosmos DB for twin state, coverage and lineage
Key Vault, App Insights and Log Analytics
03 / AI + MODEL LAYER
Physics plus measured learning
The designed model loop combines naval-architecture priors with measured-voyage correction, coverage and confidence.
RAO, tank-test or third-party model comparison
ML residual learning as real voyages arrive
Versioned outputs for portal, API and Passport
Modern foundation, honest roadmap.
The portal, IoT Hub, Cosmos DB, Storage, Key Vault and observability are provisioned for the MVP. Automated training and model registry are the next production layer as vessel data grows.
IoT
Ship telemetry
AI/ML
Model learning
API
Reusable outputs
Measurement-trained motion model
The model is the durable asset.
Each voyage adds coverage, confidence and lineage to the vessel model, so the asset improves without turning every trip into a one-off report.
Record real response
Voyages add measured roll, pitch, heave and acceleration with loading, speed, heading and sea context.
Compare with theory
Observed response is checked against RAO assumptions, tank-test curves, class files or third-party tools.
Learn across sisters
Similar hulls can share coverage while each vessel keeps its own calibration, confidence and limits.
Reuse the model
Planning, cargo limits, offshore workability and evidence can reuse one versioned twin.
Fleet learning
Sister vessels shorten the learning curve.
A new hull starts with physics and sister-transfer coverage, then becomes vessel-specific as measured voyages arrive. Owners keep the model lineage visible.
MV Atlasmeasured3850h94%
MV Meridiansister-transfer2100h81%
MV Borealmeasured1680h72%
Newbuild C-04interpolated0h48%
Portal in action
The MVP already looks like a product.
Real portal screens show the core loop: monitor vessels, inspect response surfaces and expose the model as a reusable asset.
Motion Monitoring
Live fleet state
The same vessel snapshot drives map status, limits and model confidence.
Model Studio
Response surfaces
Scenario views expose the measured model instead of a one-off report.
Fleet Models
Operational coverage
Owners see which vessels are measured, stale, limit-bound or model-ready.
Model Studio
Simulate the vessel model, not a generic hull.
Model Studio turns the measured digital twin into response surfaces: any heading, loading, sea state and virtual point can be compared against theory and exported.
Scenario compare Model API · CSV · JSON
ScenarioIllustrative interface
Absolute% limitUncertaintyObserved vs model
Illustrative response surfaceHeading x Tp
Peak wave period Tp (s)
16151413121110987654
0.46
0.59
1.14
1.82
2.50
3.01
3.25
3.17
2.81
2.26
1.67
1.19
0.92
0.54
0.83
1.38
2.06
2.74
3.25
3.49
3.41
3.05
2.50
1.91
1.43
1.16
0.99
1.28
1.83
2.51
3.19
3.70
3.94
3.86
3.50
2.95
2.36
1.88
1.61
1.64
1.93
2.47
3.16
3.83
4.35
4.59
4.51
4.14
3.60
3.01
2.53
2.26
2.33
2.62
3.16
3.85
4.52
5.04
5.28
5.20
4.83
4.29
3.70
3.22
2.95
2.79
3.08
3.62
4.31
4.98
5.49
5.73
5.65
5.29
4.75
4.16
3.68
3.41
2.79
3.08
3.62
4.31
4.98
5.49
5.73
5.65
5.29
4.75
4.16
3.68
3.41
2.33
2.62
3.16
3.85
4.52
5.04
5.28
5.20
4.83
4.29
3.70
3.22
2.95
1.64
1.93
2.47
3.16
3.83
4.35
4.59
4.51
4.14
3.60
3.01
2.53
2.26
0.99
1.28
1.83
2.51
3.19
3.70
3.94
3.86
3.50
2.95
2.36
1.88
1.61
0.54
0.83
1.38
2.06
2.74
3.25
3.49
3.41
3.05
2.50
1.91
1.43
1.16
0.46
0.59
1.14
1.82
2.50
3.01
3.25
3.17
2.81
2.26
1.67
1.19
0.92
0.46
0.49
1.03
1.72
2.39
2.91
3.15
3.07
2.70
2.16
1.57
1.09
0.82
0153045607590105120135150165180
Relative wave heading (deg): 0 following / 90 beam / 180 head
CompareTheory vs measured hull response
ViewsSurface, heading sweep and time-series
ExportCSV · JSON · API
Access layer
The model leaves the portal under owner control.
The portal is the operational interface. The deeper value is a versioned model that trusted teams can query, export and verify without taking ownership of raw data.
Versioned modelCoverage, confidence and lineage
Scoped accessAPI, CSV, JSON and reports
Reusable outputsRouting, engineering and evidence
Scoped query access
Access can be limited by vessel, model version, voyage context or vendor workflow.
Routing and fuel systems
External tools can request motion-aware response surfaces instead of static assumptions.
Engineering review
Measured-vs-theory outputs can be shared with naval architects, class or shipyards.
Evidence on demand
Selected voyages can become signed Motion Passport records with model lineage.
Signed model output
Motion Passport is generated from the model.
The Passport stays valuable because it carries the model version, measured coverage, loading context, active limits, observed voyage record and verification payload.
Pre-voyageFreeze assumptions
Loading condition, route, weather assumptions, learned model version and active limit versions.
During voyageTrack the observed record
Absolute motions, gaps, filtered intervals and anomalies stay tied to the same voyage context.
Post-voyageSign the output
The final record carries version lineage and a reproducible verification payload.
Motion Passport
Illustrative structure · voyage-specific output
SIGNED
HASH
HASH
Learned model version, loading condition and planned assumptions
Active vessel, cargo or offshore criteria with version trace
Absolute motions, data gaps, filtered intervals and events
Hash inputs, signature payload and third-party reproducibility
Absolute values remain primary
LifecyclePre-voyage · during voyage · post-voyage
VerificationVersioned payload, signature lineage and exportable evidence
Early Access
Start with one vessel.
Build the first measured Motion Digital Twin, review Model Studio and expose the first controlled API/export workflow.
90-day pilot · No drydock · Edge node + portal + first model export