Intelligent diagnostics
Constela Insight
A diagnostic system that watches a machine from the outside and reconstructs what happened when it fails.
Early developmentReconstructed sequence
- 0.20 sRobot reaches the pick positionMotion stops and the gripper closes on the part.
- 0.47 sVacuum reaches hold pressure late180 ms slower than the recorded baseline.
- 1.15 sPart shifts during transferThe camera and the vibration sensor both register the slip.
- 1.62 sPlacement check failsThe part lands outside its tolerance.
Likely contributor: slow vacuum acquisitionThree independent channels support it. The evidence is attached for an engineer to confirm.
The problem
When automated equipment fails, the evidence is scattered across controller logs, sensor states, camera footage and what the operator remembers. Engineers lose hours lining it all up before they can begin to ask why.
Our approach
Insight is an independent layer of cameras, microphones, vibration and current sensors running on one synchronized clock. It needs no changes to the machine’s control system. When something goes wrong, it gives you a single timeline of what physically happened before, during and after.
What we are building first
A portable prototype that records synchronized video and sensor data, flags repeatable anomalies and produces event timelines. Reliable evidence comes first; automated root-cause analysis follows once that evidence can be trusted.
Who it is for
- Operators of complex equipment
- Robotic system developers
- Equipment integrators
- Field engineering teams
- OEM test and validation teams