Back to results
US 12,176,182 B2Granted
Automated operational control of micro-tooling devices
AssigneeCarl Zeiss SMT GmbH
InventorsRamani Pichumani, Anthony Morin, David Dryden, Heyjin Chris Park
Life of this patent
Estimating the term…
Abstract
A micro-tooling device, such as, for example, a scanning electron microscope or a focused-ion beam microscope, provides images. A first machine-learning algorithm and a second machine-learning algorithm are sequentially coupled. The first machine-learning algorithm determines a progress along a predefined workflow based on feature recognition in images associated with the workflow. The second machine-learning algorithm predicts settings of operational parameters of the micro-tooling device in accordance with the progress along the predefined workflow.
Claims20
A method, comprising:
- obtaining a time series of images while using one or more first settings of operational parameters of a micro-tooling device;
- providing each image of the time series of the images to one or more first algorithms;
- obtaining, from the one or more first algorithms, a time series of one or more properties of a predefined feature included in the images;
- providing at least some of the images of the time series of images to a third algorithm;
- obtaining, from the third algorithm, at least one parameter selected from the group consisting an orientation of the predefined feature in the at least some of the images and a localization of the predefined feature in the at least some of the images;
- providing to a second algorithm:
- i) the time series of the one or more properties of the predefined feature included in the images;
- and ii) the at least one parameter;
- obtaining, from the second algorithm, a prediction for a second setting of the operational parameters of the micro-tooling device;
- and
- controlling the operation of the micro-tooling device in accordance with the second setting of the operational parameters,
- wherein the method further comprises:
- receiving, from a human-machine-interface, a third setting of the operational parameters of the micro-tooling device;
- comparing the second and third settings of the operational parameters of the micro-tooling device;
- and
- training a machine-learning kernel of the second algorithm based on the comparison of the second and third settings of the operational parameters of the micro-tooling device.
13 more elements
The method of 2, further comprising:
- monitoring the state of the predefined feature;
- and
- at least partially based on monitoring the state of the predefined feature, selectively triggering control of the operation of the micro-tooling device in accordance with the second setting.
A method, comprising:
- obtaining a time series of images while using one or more first settings of operational parameters of a micro-tooling device;
- providing each image of the time series of the images to one or more first algorithms;
- obtaining, from the one or more first algorithms, a time series of one or more properties of a predefined feature included in the images;
- providing at least some of the images of the time series of images to a third algorithm;
- obtaining, from the third algorithm, at least one parameter selected from the group consisting an orientation of the predefined feature in the at least some of the images and a localization of the predefined feature in the at least some of the images;
- providing to a second algorithm:
- i) the time series of the one or more properties of the predefined feature included in the images;
- and ii) the at least one parameter;
- obtaining, from the second algorithm, a prediction for a second setting of the operational parameters of the micro-tooling device;
- and
- controlling the operation of the micro-tooling device in accordance with the second setting of the operational parameters,
- wherein the method further comprises:
- receiving, from a human-machine-interface, a user input of the predefined feature in at least one image of the time series of images;
- and
- training a machine-learning kernel of the one or more first algorithms based on the user input.
12 more elements
A method, comprising:
- obtaining a time series of images while using one or more first settings of operational parameters of a micro-tooling device;
- providing each image of the time series of the images to one or more first algorithms;
- obtaining, from the one or more first algorithms, a time series of one or more properties of a predefined feature included in the images;
- providing at least some of the images of the time series of images to a third algorithm;
- obtaining, from the third algorithm, at least one parameter selected from the group consisting an orientation of the predefined feature in the at least some of the images and a localization of the predefined feature in the at least some of the images;
- providing to a second algorithm:
- i) the time series of the one or more properties of the predefined feature included in the images;
- and ii) the at least one parameter;
- obtaining, from the second algorithm, a prediction for a second setting of the operational parameters of the micro-tooling device;
- and
- controlling the operation of the micro-tooling device in accordance with the second setting of the operational parameters,
- wherein the predefined feature in the images of the time series comprises a trimer beamlet of a gas-field ion source of the micro-tooling device.
9 more elements
Dependencies are inferred from the claim text.