Security & Integrity at Unearthed

How we use AI reasoning loops and GNSS interference detection to keep tracking data accurate—and keep client data private.

Modern tracking systems depend on GNSS. That makes them vulnerable: spoofing can deceive receivers into reporting false position/time, and jamming can disrupt reception and cause data gaps or loss. Unearthed is designed to protect both integrity (is the location real?) and confidentiality (who can see it, and how is it protected?).

The Threat Model

Understanding the risks to GNSS-based tracking:

Jamming: interference that overpowers signals and can cause partial/complete loss.

Spoofing: convincing fake signals that can produce incorrect data for longer periods, sometimes without obvious loss of signal power.

Our Approach: Layered Defenses

Research surveys group GNSS spoofing/jamming detection into multiple categories—signal quality monitoring, power-based checks (e.g., C/N0, AGC), correlation-peak monitoring, timing/angle methods, NMEA message analysis, pseudorange integrity checks, and machine/deep learning. Unearthed combines these ideas into a practical, production-minded pipeline:

Layer 1 — Real-time GNSS Health Checks

We continuously evaluate receiver/measurement signals commonly used to detect interference:

Signal power / C/N0 trends: useful, but not sufficient alone for sophisticated spoofing

AGC behavior: simple to integrate; stronger when combined with other signals

Correlation-peak monitoring: targets abnormal correlation behavior

Layer 2 — Message & Measurement Consistency

We cross-check internal GNSS consistency using approaches such as:

NMEA sentence integrity/consistency analysis

Pseudorange integrity-style checks (a major class of methods used to spot spoofed positioning)

Layer 3 — Machine Learning Classification

Reviews emphasize ML/DL as a major trend for detecting and classifying spoofing/jamming, often using public datasets (e.g., TEXBAT, OAKBAT). We use ML where it adds signal—primarily for pattern recognition across multiple indicators rather than single-metric alerting.

Layer 4 — Network-Aware Monitoring and Escalation

Some approaches detect spoofing across an area using network monitoring concepts (detecting patterns that indicate a common spoofer). Unearthed is designed to support fleet-wide correlation—so one suspicious event can harden the system for all assets in the same region/time window.

Where AI Reasoning Fits (The "Why" Behind Alerts)

We build autonomous detection as a reasoning + acting loop: the system forms a hypothesis, runs verification steps, and logs what it observed and why it concluded an incident occurred. This mirrors the core idea behind ReAct-style systems: interleaving reasoning with actions to improve robustness and allow inspection/correction by humans.

Example "Reason → Act → Observe" Flow (Simplified)

  1. Observe: sudden location jump + unusual GNSS health metrics

  2. Reason: "Could be spoofing, jamming, multipath, or device reset"

  3. Act: run additional checks (signal-quality trends, message consistency, route plausibility)

  4. Observe: results returned with confidence indicators

  5. Decide: mark event type + severity; attach an audit trail; notify only if thresholds are met

  6. Escalate: optional human review with a clear, editable incident narrative (controllable behavior)

What Clients Get (Product-Level Outcomes)

Integrity Alerts (Spoofing/Jamming Suspicion): with an evidence summary, not a mystery warning

Confidence Scoring: "how sure are we?" based on multiple indicators and consistency checks

Incident Timeline: what changed, what checks ran, and what the system concluded (auditability)

Privacy Controls: role-based access, retention windows, and export controls for sensitive tracking data

Privacy & Confidentiality (Non-Negotiables)

Data minimization: collect what's needed for operations—not everything that's possible

Encryption: protect data in transit and at rest

Tenant isolation: reduce blast radius and prevent cross-client exposure

Client control: retention and access policies that match your environment

Ready to Secure Your Fleet?

Get in touch with our team to learn how Unearthed can protect your tracking operations with enterprise-grade security and AI-powered integrity monitoring.

Contact Us