Observability
Identify which object, tool, robot and process states are directly measured, inferred or left unknown.
System anatomy · 07
Sensing and control
Examine how vision, force, encoders, calibration and process feedback detect variation, choose action and close the quality loop.
Evidence statusEngineering decision framework · cell-specific validation requiredLast reviewed: 11 August 2026
07
FUURAA thesis
Nominal sensor accuracy matters less than calibrated, observable behaviour under the variations and failures production actually creates.
Engineering map
Turn the title into engineering objects that can be observed, measured and reviewed.
Each layer states the boundary to establish and the evidence needed for the next decision.
Calibration chain
Trace camera, tool, base, fixture, force and process frames with uncertainty, ownership and recheck triggers.
Control and fallback
Define confidence, limits, rejection, retry, safe stop and human escalation for ambiguous observations.
Verification questions
Write the questions first, then decide whether a demo, test, pilot or operating record can answer them.
Each question needs an object, conditions, denominator, threshold and accountable decision owner.
- 01
Which variation is represented in validation data?
- 02
How is calibration drift detected before bad product escapes?
- 03
What happens when sensors disagree or confidence is low?
- 04
Can logs reconstruct the observation and decision after an incident?
Evidence to preserve
Enable the next reader to reconstruct conditions, results, failures and the decision.
A conclusion alone loses reviewability; raw records, configuration and exclusions matter too.
- 01
Evidence package 1
Calibration hierarchy, uncertainty budget and recheck record
- 02
Evidence package 2
Confusion, miss, false-positive and drift results by condition
- 03
Evidence package 3
Versioned sensor, model, threshold and fallback configuration
Scope boundary
State what this evidence still cannot be generalised to.
- 01
A perception benchmark or calibration certificate does not prove robust closed-loop production under changing material and lighting.
- 02
This is an engineering reading and decision framework; it does not replace application-specific risk assessment, conformity work, procurement acceptance or professional advice.
Sources and evidence status
Read standards scope, measurement evidence and application conclusions separately.
Source dates and review status remain visible; external sources open in a new tab.
01
NIST · 2019-12-02
↗02Helping Robots Stay on Target
Explains why repeatability and accuracy are distinct and why preparation and maintenance affect both.
NIST · Reviewed 2026-08-11
↗Dexterous grasping, manipulation and contact-safety measurement
Develops metrics and test methods for grasping, manipulation, control and contact safety.