14:32:11.04
LiDAR confidence decreased
Return quality fell below 0.45 on consecutive sweeps.
AI incident investigation for autonomous systems.
Pharion reconstructs failures across ROS, telemetry, sensors, and robot state — identifying what happened, why it happened, and how to prevent it.
Built for robotics engineering teams operating autonomous fleets.
INC-042AMR-17WH-EAST / aisle B414:32:18.02 UTCReconstructing
Investigation of incident 042: LiDAR degradation leads to localization instability, planner hold, and a protective stop on AMR-17.
- .04
- .21
- .47
- .63
- .71
- .89
Localization instability caused by degraded LiDAR observations. Evidence traces from return density through pose covariance to the protective stop.
When autonomous systems fail, engineers become detectives.
A single incident can span bags, telemetry, cameras, LiDAR, IMU, localization, planner state, controller output, and logs. The data exists. Reconstructing causality still happens by hand.
ROS / MCAP
bags · topics · tf
Telemetry
rates · estop · power
Cameras
streams · frames
LiDAR
returns · intensity
IMU
accel · gyro
Localization
pose · covariance
Planner
costmap · trajectory
Controller
cmd_vel · safety
System logs
nodes · diagnostics
An engineer opens each source in a different tool, aligns clocks by hand, and guesses which events belong to the same failure.
Reached hours later — if the trail can be reconstructed at all.
The data exists. The intelligence to connect it doesn't.
Observability tells you what happened.
Pharion investigates why.
Visualization and monitoring remain essential. Engineers still have to stitch timelines, compare robot state, and argue about causality. Pharion sits above that stack — it does not replace it.
- 01InfrastructureRobots, fleets, and the ROS / ROS 2 runtime already producing state.
- 02DataMCAP, bags, telemetry, LiDAR, cameras, IMU, GPS, logs, diagnostics.
- 03ObservabilityFoxglove, RViz, Grafana, and logs show what the system did. Necessary. Not an investigation.
04PHARIONInvestigation intelligence: correlate robot state, telemetry, logs, and sensor behavior into a causal chain.
Timeline
Reconstructed event clock
Evidence
Observable system behavior
Root cause
Why the failure occurred
Recommendation
What to change next
From fragmented streams to a defensible investigation.
01
Ingest
Connect the data autonomous systems already produce: ROS / ROS 2, MCAP, telemetry, cameras, LiDAR, IMU, GPS, diagnostics, and logs.
02
Reconstruct
Rebuild a single incident timeline across sensors, localization, planner state, controller state, and software events.
03
Investigate
Correlate anomalies and causal relationships — not just correlated charts, but ordered evidence.
04
Explain
Produce an evidence-backed root-cause analysis and recommended corrective actions the team can act on.
An investigation, not a dashboard.
Incident #042 — AMR-17 stopped in aisle B4. The interface below is the kind of workspace Pharion is built to be: dense, ordered, and accountable to the data.
INC-042AMR-17Fleet: WH-EAST14:32:18.02 UTCprotective_stop
Robot state
LiDAR conf.
0.41
↓ 0.37
Pose cov.
0.18
↑ 0.12
Loc. quality
LOW
warn
Planner
HOLD
changed
Causal reconstruction
LiDAR degradation → localization confidence ↓
Pose covariance ↑ → planner receives unstable pose
Controller enters safety state → robot stops
Recommendation
Inspect LiDAR occlusion and return-quality thresholds on AMR-17. Raise localization covariance gate before planner execution. Replay incident against last known-good map at 14:31:40.
Don't just get an answer. Get the evidence.
Every conclusion is traceable to observable system behavior. If it cannot be shown in the data, it does not ship as a finding.
Likely root cause
Localization instability caused by degraded LiDAR observations.
14:32:11.21
Localization covariance increased
AMCL covariance crossed the operational bound.
14:32:11.47
Pose divergence detected
Estimated pose diverged from wheel odometry by 0.42 m.
14:32:11.63
Planner state changed
Planner froze the trajectory after receiving unstable pose.
14:32:11.71
Safety controller triggered
Protective stop issued; drive torque commanded to zero.
The story of a failure, reconstructed.
Pharion aligns events from independent subsystems onto one clock so the sequence is visible — not inferred from memory.
- LiDARConfidence ↓ 0.78 → 0.41
- LocalizationCovariance ↑ 0.06 → 0.18
- State est.Pose divergence 0.42 m
- PlannerState HOLD (was TRACK)
- ControllerSafety trigger: protective_stop
- BaseRobot stopped · v = 0.00 m/s
Autonomous systems where failure investigation matters.
Pharion starts with industrial autonomy. The same investigation problem appears anywhere a machine acts without a human in the loop.
Initial wedge
Industrial robotics
Warehouse AMRs, autonomous forklifts, and manufacturing cells — where unexplained stops halt operations.
Horizon
Drones & defense
Autonomous aerial systems in dynamic environments, where after-action reconstruction cannot wait on tribal knowledge.
Horizon
Space robotics
Satellites, spacecraft, and remote platforms where replay is expensive and the system cannot be walked up to.
Horizon
Advanced robotics
Humanoids, field robots, inspection, and agricultural systems — high-DOF platforms with dense multimodal state.
The more autonomous systems scale, the more expensive unexplained failures become.
01
Autonomous systems are scaling
Robots are leaving controlled prototypes and entering live operations — warehouses, yards, fields, and airspace.
02
Systems are becoming more complex
Perception, planning, control, and learned models now share a vehicle. Failures cross subsystem boundaries.
03
Reliability is a competitive advantage
As fleets grow, unexplained downtime compounds. Teams that can investigate faster ship safer systems.
The reliability intelligence layer for autonomous systems.
Today
Engineers investigate failures manually — across bags, logs, and tribal knowledge.
Tomorrow
Autonomous systems continuously understand their own failures, with evidence attached.
Eventually
Every robot has a memory of what happened, why it happened, and what changed.
Built by engineers who have lived the investigation.
Pharion is being built by people with experience across robotics, autonomy, AI, and the software systems that keep fleets running. Founder details will land here.
- Robotics
- Autonomy
- AI systems
- Engineering infrastructure
When your autonomous system fails, know why.
See how Pharion turns fragmented system data into an evidence-backed investigation.