ProxControl / Deep Tech
Turning noisy movement into useful industrial measurement.
Industrial spray-paint velocity and painting-state detection for layer-thickness analytics.
- Our contribution
- Signal processing & engineering validation
- Delivery
- Two-phase algorithm development
The challenge
Measuring a handheld spray gun’s velocity from inertial sensor data introduces drift. Vibration, pauses, and direction changes make it harder to distinguish real movement from accumulated error. That undermines the layer-thickness analytics the measurement needs to support.
How we solved it
Understand the physical process
We treated the painting motion as part of the engineering problem. Detecting stationary moments and understanding repeated sweeps gave the correction logic useful information about what the sensor should be measuring.
Correct drift, then validate the signal
The work combined zero-velocity updates with filtering and motion constraints. The two-phase approach separated the correction problem from refinement, with validation used to assess whether the output tracked the real movement.
The outcome
Industrial velocity drift: solvedRalphNex solved the industrial velocity-drift challenge and supported painting-state detection for layer-thickness analytics. The value was in making the measurement useful for the application, not adding unnecessary algorithmic complexity.
