Systems, methods, and media for artificial intelligence feedback control in manufacturing

Abstract

Additive manufacturing systems using artificial intelligence can identify an anomaly in a printed layer of an object from a generated topographical image of the printed layer. The additive manufacturing systems can also use artificial intelligence to determine a correlation between the identified anomaly and one or more print parameters, and adaptively adjust one or more print parameters. The additive manufacturing systems can also use artificial intelligence to optimize one or more printing parameters to achieve desired mechanical, optical and/or electrical properties.

What is claimed: 
     
         1 . A manufacturing system, comprising:
 an image sensor configured to capture a first image of a specimen; and   at least one hardware processor configured to:
 receive the first image from the image sensor; 
 obtain one or more desired properties for the specimen; 
 generate a second image of the specimen based on the first image; 
 identify an anomaly on the specimen from the second image using a first artificial intelligence algorithm that is configured to detect anomalies on the specimen; 
 determine a correlation between the identified anomaly and one of a plurality of processing parameters used to process the specimen using a second artificial intelligence algorithm that is configured to determine correlations between identified anomalies and the one of the plurality of processing parameters; 
 adjust a value for the one of the plurality of processing parameters to be used by the manufacturing system for processing the specimen; and 
 cause the manufacturing system to continue processing the specimen using the value for the one of the plurality of processing parameters to substantially achieve the one or more desired properties. 
   
     
     
         2 . The manufacturing system of  claim 1 , wherein generating the second image of the specimen based on the first image comprises:
 applying one of a shape-from-focus algorithm, a shape-from-shading focus algorithm, a photometric stereo algorithm, and a Fourier ptychography modulation algorithm to the first image.   
     
     
         3 . The manufacturing system of  claim 1 , identifying the anomaly on the specimen from the second image using the first artificial intelligence algorithm comprises:
 comparing the second image with a production design for the specimen.   
     
     
         4 . The manufacturing system of  claim 1 , wherein the at least one hardware processor is further configured to:
 identify a processing parameter that affects at least one of a mechanical property, an optical property, and an electrical property of the specimen;   measure the at least one of the mechanical property, the optical property, and the electrical property after the specimen undergoes a first process step;   determine an anomaly rate of the specimen; and   determine an impact of the anomaly rate and an impact of the processing parameter on at least one of the mechanical property, the optical property, and the electrical property of the specimen.   
     
     
         5 . The manufacturing system of  claim 1 , wherein the at least one hardware processor is further configured to:
 identify a processing parameter that affects at least one of a mechanical property, an optical property, and an electrical property of the specimen;   measure the at least one of the mechanical property, the optical property, and the electrical property after the specimen undergoes a first process step;   determine an anomaly pattern of the specimen; and   determine an impact of the anomaly pattern and an impact of the processing parameter on at least one of the mechanical property, the optical property, and the electrical property of the specimen.   
     
     
         6 . The manufacturing system of  claim 1 , wherein the at least one hardware processor is further configured to:
 train the first artificial intelligence algorithm to identify anomalies on training images of training specimens.   
     
     
         7 . The manufacturing system of  claim 1 , wherein the at least one hardware processor is further configured to:
 train the second artificial intelligence algorithm to identify correlations between training anomalies and training process parameters.   
     
     
         8 . A computer-implemented method, comprising:
 receiving, by a computing system, a first image of a specimen from an image sensor of a manufacturing system;   obtaining, by the computing system, one or more desired properties for the specimen;   generating, by the computing system, a second image of the specimen based on the first image;   identifying, by the computing system, an anomaly on the specimen from the second image using a first artificial intelligence algorithm that is configured to detect anomalies on the specimen;   determining, by the computing system, a correlation between the identified anomaly and one of a plurality of processing parameters used to process the specimen using a second artificial intelligence algorithm that is configured to determine correlations between identified anomalies and the one of the plurality of processing parameters;   adjusting, by the computing system, a value for the one of the plurality of processing parameters to be used by the manufacturing system for processing the specimen; and   causing, by the computing system, the manufacturing system to continue processing the specimen using the value for the one of the plurality of processing parameters to substantially achieve the one or more desired properties.   
     
     
         9 . The computer-implemented method of  claim 8 , wherein generating, by the computing system, the second image of the specimen based on the first image comprises:
 applying one of a shape-from-focus algorithm, a shape-from-shading focus algorithm, a photometric stereo algorithm, and a Fourier ptychography modulation algorithm to the first image.   
     
     
         10 . The computer-implemented method of  claim 8 , identifying, by the computing system, the anomaly on the specimen from the second image using the first artificial intelligence algorithm comprises:
 comparing the second image with a production design for the specimen.   
     
     
         11 . The computer-implemented method of  claim 8 , further comprising:
 identifying, by the computing system, a processing parameter that affects at least one of a mechanical property, an optical property, and an electrical property of the specimen;   measuring, by the computing system, the at least one of the mechanical property, the optical property, and the electrical property after the specimen undergoes a first process step;   determining, by the computing system, an anomaly rate of the specimen; and   determining, by the computing system, an impact of the anomaly rate and an impact of the processing parameter on at least one of the mechanical property, the optical property, and the electrical property of the specimen.   
     
     
         12 . The computer-implemented method of  claim 8 , further comprising:
 identifying, by the computing system, a processing parameter that affects at least one of a mechanical property, an optical property, and an electrical property of the specimen;   measuring, by the computing system, the at least one of the mechanical property, the optical property, and the electrical property after the specimen undergoes a first process step;   determining, by the computing system, an anomaly pattern of the specimen; and   determining, by the computing system, an impact of the anomaly pattern and an impact of the processing parameter on at least one of the mechanical property, the optical property, and the electrical property of the specimen.   
     
     
         13 . The computer-implemented method of  claim 8 , further comprising:
 training, by the computing system, the first artificial intelligence algorithm to identify anomalies on training images of training specimens.   
     
     
         14 . The computer-implemented method of  claim 8 , further comprising:
 training, by the computing system, the second artificial intelligence algorithm to identify correlations between training anomalies and training process parameters.   
     
     
         15 . A non-transitory computer readable medium containing one or more sequences of instructions, which, when executed by a processor, cause a computing system to perform operations comprising:
 receiving, by the computing system, a first image of a specimen from an image sensor of a manufacturing system;   obtaining, by the computing system, one or more desired properties for the specimen;   generating, by the computing system, a second image of the specimen based on the first image;   identifying, by the computing system, an anomaly on the specimen from the second image using a first artificial intelligence algorithm that is configured to detect anomalies on the specimen;   determining, by the computing system, a correlation between the identified anomaly and one of a plurality of processing parameters used to process the specimen using a second artificial intelligence algorithm that is configured to determine correlations between identified anomalies and the one of the plurality of processing parameters;   adjusting, by the computing system, a value for the one of the plurality of processing parameters to be used by the manufacturing system for processing the specimen; and   causing, by the computing system, the manufacturing system to continue processing the specimen using the value for the one of the plurality of processing parameters to substantially achieve the one or more desired properties.   
     
     
         16 . The non-transitory computer readable medium of  claim 15 , wherein generating, by the computing system, the second image of the specimen based on the first image comprises:
 applying one of a shape-from-focus algorithm, a shape-from-shading focus algorithm, a photometric stereo algorithm, and a Fourier ptychography modulation algorithm to the first image.   
     
     
         17 . The non-transitory computer readable medium of  claim 15 , identifying, by the computing system, the anomaly on the specimen from the second image using the first artificial intelligence algorithm comprises:
 comparing the second image with a production design for the specimen.   
     
     
         18 . The non-transitory computer readable medium of  claim 15 , further comprising:
 identifying, by the computing system, a processing parameter that affects at least one of a mechanical property, an optical property, and an electrical property of the specimen;   measuring, by the computing system, the at least one of the mechanical property, the optical property, and the electrical property after the specimen undergoes a first process step;   determining, by the computing system, an anomaly rate of the specimen; and   determining, by the computing system, an impact of the anomaly rate and an impact of the processing parameter on at least one of the mechanical property, the optical property, and the electrical property of the specimen.   
     
     
         19 . The non-transitory computer readable medium of  claim 15 , further comprising:
 identifying, by the computing system, a processing parameter that affects at least one of a mechanical property, an optical property, and an electrical property of the specimen;   measuring, by the computing system, the at least one of the mechanical property, the optical property, and the electrical property after the specimen undergoes a first process step;   determining, by the computing system, an anomaly pattern of the specimen; and   determining, by the computing system, an impact of the anomaly pattern and an impact of the processing parameter on at least one of the mechanical property, the optical property, and the electrical property of the specimen.   
     
     
         20 . The non-transitory computer readable medium of  claim 15 , further comprising:
 training, by the computing system, the first artificial intelligence algorithm to identify anomalies on training images of training specimens; and   training, by the computing system, the second artificial intelligence algorithm to identify correlations between training anomalies and training process parameters.