DeepTrace

System-level diagnostics for unmanned systems

DeepTrace is ANOX's diagnostic intelligence system.

It builds a system model from existing telemetry and logs across subsystems and time, traces causal relationships, and produces clear root-cause insights with recommended actions.

Built for unmanned systems (UAVs, ROVs) across remote, autonomous, and hybrid operations.

cloud_done

Delivered as a service. No hardware. No heavy integration.

build_circle[WHAT DEEPTRACE DOES]

Core Diagnostics

  • check_circleAnalyzes data across hardware, control, navigation, and mission phases
  • check_circleDetects hard failures and performance degradation
  • check_circleIdentifies inefficient configurations and interaction effects
  • check_circleExplains issues in system context, not as isolated signals
  • check_circleProvides concrete recommendations for corrective action

settings_suggest[HOW IT WORKS]

upload_file

1. Provide Your Data

Upload post-mission logs or connect existing telemetry sources. Works with ArduPilot, PX4, and custom systems.

psychology

2. System-Level Analysis

DeepTrace evaluates interactions across subsystems, missions, and operational modes.

summarize

3. Diagnostic Output

Clear explanations, supporting evidence, and recommended actions within minutes.

No manual labeling. No brittle thresholds. No custom rules per vehicle.

fact_check[REAL ISSUES DEEPTRACE HAS DIAGNOSED]

error

Example 1: Unexplained Loss of Control

Hard FailureArduCopter 4.6
infoSymptoms

Loss of control after 6 minutes of flight. Motor C1 saturated to ~1900 µs while other motors remained normal. Crash disarm with 153° angle error. No obvious hardware damage, power sag, or vibration issues.

searchRoot Cause Found

Localized thrust deficit on C1 arm. Controller saturated trying to compensate for weak/failed motor, ESC, or prop. Likely intermittent connection or thermal-induced ESC desync that manifested after components warmed up.

check_circleResolution

Component swap testing isolated faulty ESC. Replaced ESC, switched to DShot protocol, and updated firmware.

Result: Stable flights with balanced motor outputs. 90% diagnostic confidence.

warning

Example 2: Intermittent Instability with Large Props

ConfigurationArduCopter 4.6
infoSymptoms

Late-flight thrust deficit with RCOU saturation on one channel. System using PWM ESC signaling with 11-inch props.

searchRoot Cause Found

PWM signaling with large props increases ESC desync risk under high load or yaw maneuvers. Conservative ESC timing combined with electrical/mechanical stress from 11" props can cause transient desyncs.

check_circleResolution

Switched to DShot protocol and updated ESC timing/demag settings for motor KV and prop size.

Result: Eliminated desync events during aggressive maneuvers. 65% diagnostic confidence.

verified

Example 3: Ruling Out False Leads

ValidationArduCopter 4.6
infoSymptoms

Crash event could have been caused by vibration, power brownout, compass interference, or EKF failure.

searchAnalysis

DeepTrace systematically ruled out alternative causes: vibrations within acceptable limits (no clipping), battery voltage/current normal (no brownout), EKF yaw and magnetometer innovations nominal until tumble began.

check_circleResolution

Confirmed propulsion as root cause, preventing wasted time on compass recalibration, vibration damping, or power system upgrades.

Result: Focused repair effort, faster return to operations. 80-85% confidence on excluded causes.

support[SUPPORTED PLATFORMS]

check_circle

Available Now

  • arrow_rightArduPilot
update

Expanding

  • arrow_rightPX4
  • arrow_rightAdditional autopilot platforms
  • arrow_rightCustom & proprietary systems

Custom integrations available for enterprise customers.

layers[PRODUCT SCOPE & EVOLUTION]

One Product. Three Layers.

DeepTrace is designed around three layers of machine understanding.

We sell the first layer today. We are actively expanding into the second. We architect the system for the third.

check_circle

1. Diagnostics (Available Today)

This is the core of the current product.

DeepTrace performs automated diagnostics on autonomous and remotely operated systems, including:

  • circleRoot cause analysis of failures and abnormal behavior
  • circleParameter anomaly detection
  • circleEvidence-backed explanations
  • circleActionable recommendations for correction

This replaces manual telemetry review and ad-hoc troubleshooting with a repeatable, system-level diagnostic process.

trending_up

2. Operational Intelligence (Early / Expanding)

Beyond failures, many problems show up as inefficiency, instability, or performance regression.

DeepTrace is expanding into operational intelligence, including:

  • circleEfficiency analysis (battery usage, control effort, tuning quality)
  • circleDetection of performance degradation across missions
  • circleAnalysis of interdependencies between components and parameters
  • circleExplanations such as "this parameter change caused that downstream behavior"

This allows teams to understand not just what broke, but what slowly got worse.

auto_awesome

3. Autonomy & Immunity (Long-Term Direction)

The long-term direction is autonomous system health.

This includes:

  • circleEarly prevention of known failure patterns
  • circleSelf-correction based on learned behavior
  • circleClosed-loop learning across missions and fleets

This layer is not sold as autonomy today — it is the direction the system is being built toward.

Ready to try DeepTrace?

arrow_forwardTry DeepTrace