Method for automatically adjusting manufacturing limits prescribed on an assembly line

A method comprising:

• for each assembly unit in a historical population of assembly units assembled on an assembly line during a first time period and tested according to a target test following a target assembly stage on the assembly line:
• accessing feature values, of a corpus of features, detected in the assembly unit in a sequence of assembly stages upstream of the target test;
• and
• accessing a failure status of the assembly unit at the target test;
• for each feature, in the corpus of features:
• deriving a correlation between feature values, of the feature, and failure status at the target test;
• deriving an effective limit of the feature based on scope of feature values, of the feature, in the historical population of assembly units;
• and
• calculating an action score for the feature based on:
• the correlation;
• and
• a width of the effective limit of the feature;
• selecting a particular feature exhibiting a greatest action score in the corpus of features;
• defining a preemptive test for the particular feature upstream of the target assembly stage during a second time period succeeding the first time period;
• and
• assigning a target limit, narrower than an effective limit of the particular feature, to the preemptive test.

A method comprising:

• for each assembly unit in a historical population of assembly units assembled on an assembly line during a first time period:
• accessing feature values, of a corpus of features, extracted from a set of images of the assembly unit, the set of images captured by a set of optical inspection stations arranged on the assembly line;
• and
• accessing a failure status of the assembly unit;
• for each feature, in the corpus of features:
• deriving a correlation between feature values, of the feature, and failure status within the historical population of assembly units;
• accessing a tolerance limit assigned to the feature;
• characterizing a sensitivity of a failure frequency within the historical population of assembly units based on feature values of assembly units falling outside of the tolerance limit assigned to the feature;
• calculating an action score for the feature based on:
• the correlation;
• and
• the sensitivity of the failure frequency;
• selecting a particular feature, from the corpus of features, based on a particular action score of the particular feature;
• generating a prompt to reduce failure frequency of assembly units assembled on the assembly line by increasing width of a particular tolerance limit of the particular feature during a second time period succeeding the first time period;
• and
• serving the prompt to a user affiliated with the assembly line.