How GUARA handles operator data

GUARA is designed to estimate operational human state while minimizing unnecessary exposure of personal information. Its core detection can run locally on the device and does not require identifying the operator.

Local processing
No identity recognition required
No continuous cloud video
Minimized telemetry

01What GUARA analyzes

GUARA looks at visual signals associated with attention and alertness while a person operates a vehicle or a safety-critical task. It interprets how these signals behave and change over time:

Eye openness and blinking
Gaze direction
Head orientation
Attention-related visual patterns
Temporal changes associated with fatigue, drowsiness or distraction

The goal is to estimate an operational state, such as whether attention appears to be degrading, not to determine who the person is.

02What GUARA does not need

Facial recognition to determine identityThe system evaluates operational-state signals. It does not need to recognize or match a face to know who the operator is.

Permanent Internet connectivityCore detection can run on the device itself, so it keeps working in areas with limited or no connectivity.

Continuous cloud video transmissionThe architecture does not require streaming video from the cabin or workstation to a remote server.

03What can remain local

Camera images can be processed on the device to extract the signals described above. Depending on the deployment configuration, visual input can be analyzed locally and then discarded, without building a continuous video record.

This keeps the most sensitive raw material, the image of the person, as close as possible to where it is captured and minimizes how much of it ever leaves the device.

04What may be transmitted

If a deployment is configured for it, GUARA may synchronize operational events (for example, a detected drowsiness episode) and minimized technical metrics (for example, aggregated detection statistics or system health data). These are used to support monitoring, safety follow-up and system performance.

Even without video, telemetry may still be personal data if it can be linked to a specific person, vehicle or shift. It should be handled with the same care as any other operational record about people.

05Privacy modes

Deployments can be configured to match the needs and requirements of each organization. Typical profiles include:

MODE 01Local Only

Detection and alerts run on the device. Events and metrics are not synchronized to external systems.

MODE 02Operational Telemetry

Operational events and minimized technical metrics are synchronized for fleet or safety monitoring. No continuous video is uploaded.

MODE 03Authorized Research

For validation or model improvement under explicit authorization, agreed scope and defined retention, data collection can be expanded within those limits.

06Designed for operational safety

GUARA is intended to monitor risk-related operational states such as fatigue, drowsiness or distraction, so that risk can be detected and addressed in time. It is not intended to measure personal productivity or to serve as a general disciplinary surveillance tool.

FAQFrequently asked questions

Does GUARA identify me?

No. GUARA does not require facial recognition to evaluate operational state. It analyzes signals such as eye openness, gaze and head orientation, not who you are.

Does GUARA need Internet?

No. Core detection can run locally on the device. Connectivity is only needed if a deployment is configured to synchronize events or metrics.

Is GUARA recording me continuously?

Not necessarily. The architecture can process images locally without storing continuous video.

Can video be stored?

Only if a specific deployment is configured to do so, for example under an authorized research setup. In that case, the scope, access and retention should be defined in advance by the organization.

What data can be transmitted?

Operational events and minimized technical metrics, when the configuration requires it. Continuous cloud video is not part of the standard operating model.

Can settings differ between deployments?

Yes. Each organization can define what stays local, what is synchronized and for how long it is kept, according to its own privacy, safety and legal requirements.

Data retention, telemetry and access rules should be defined for each deployment according to the organization’s privacy, safety and legal requirements.

Questions about privacy in a deployment?

We can walk your operations, HR or compliance team through how GUARA would be configured in your environment.

guaraplatform@guaraplatform.com