Granted Patent B2

US 12,524,286 · App. 18/541,536 · Granted Jan 13, 2026

Inventors: Taylor Jensen (Chicago, IL); Fiona J. O'Laughlin (East Peoria, IL); Diana E. Huerta (Chicago, IL); Christopher M. Burkard (Dunlap, IL); Christopher Ha (Champaign, IL); David Jason McIntyre (Peoria, IL); Torsten Van Wassenhove (Edwards, IL)

Assignee: Caterpillar Inc.

G06F11/0703G06Q10/06316

Abstract

Claims (153)

1 . A non-transitory computer-readable storage medium storing instructions, which, when executed by at least one data processor of a system, cause the system to:

obtain an indication of multiple troubleshooting guides associated with a first issue experienced by a first machine, a first model associated with the first machine, a first type associated with the first machine, and a first business in which the first machine is configured to operate,

wherein a troubleshooting guide among the multiple troubleshooting guides includes multiple troubleshooting procedures to perform,

wherein a troubleshooting procedure among the multiple troubleshooting procedures includes a testing step associated with the first issue and a repair step associated with the first issue;

generate training data for an artificial intelligence using the multiple troubleshooting guides by:

extracting from the multiple troubleshooting guides multiple paths leading to the repair step,

wherein the multiple paths do not include multiple repair steps;

encoding the multiple paths into multiple unique identifiers;

train the artificial intelligence using the multiple unique identifiers to provide a most probable repair step that avoids multiple testing steps;

receive an input indicating a second model associated with a second machine, a second type associated with the second machine, a second business in which the second machine is configured to operate, and an indication of a second issue experienced by the second machine,

wherein the second machine is different from the first machine;

based on the input and the indication of the multiple troubleshooting guides, generate a distribution of probabilities associated with the multiple troubleshooting procedures;

based on the distribution of probabilities, provide, using the artificial intelligence, the most probable repair step that avoids multiple testing steps.

2 . The non-transitory computer-readable storage medium of claim 1 , comprising instructions to:

generate a second troubleshooting guide directed to the second machine by:

adding the indication of the second issue and most likely troubleshooting procedure to the second troubleshooting guide;

receiving a result of the most likely troubleshooting procedure to obtain a received result; and

performing below steps:

based on the received result, generate a second distribution of probabilities associated with the multiple troubleshooting procedures;

select a second most likely troubleshooting procedure,

wherein the second most likely troubleshooting procedure includes a repair step associated with the second issue or the testing step associated with the second issue;

add the second most likely troubleshooting procedure to the second troubleshooting guide;

receive a result of the second most likely troubleshooting procedure to obtain the received result;

until the received result indicates that the second issue is resolved.

3 . The non-transitory computer-readable storage medium of claim 1 , wherein the artificial intelligence includes a generative artificial intelligence including a Long Short-Term Memory neural network.

4 . The non-transitory computer-readable storage medium of claim 1 , comprising instructions to:

generate a second troubleshooting guide directed to the second machine by:

adding the indication of the second issue and most likely troubleshooting procedure to the second troubleshooting guide;

receiving a result of the most likely troubleshooting procedure to obtain a received result; and

performing below steps:

based on the received result, generate a second distribution of probabilities associated with the multiple troubleshooting procedures;

select a second most likely troubleshooting procedure,

wherein the second most likely troubleshooting procedure includes a repair step associated with the second issue or the testing step associated with the second issue;

add the second most likely troubleshooting procedure to the second troubleshooting guide;

receive a result of the second most likely troubleshooting procedure to obtain the received result;

until the received result indicates that the second issue is resolved;

obtain a probability threshold;

determine whether two or more probabilities associated with two or more troubleshooting procedures among the multiple troubleshooting procedures satisfy the probability threshold; and

upon determining that the two or more troubleshooting procedures satisfy the probability threshold, create a branch in the second troubleshooting guide by adding the two or more troubleshooting procedures to perform to the second troubleshooting guide

wherein the two or more troubleshooting procedures include the most likely troubleshooting procedure or the second most likely troubleshooting procedure.

5 . The non-transitory computer-readable storage medium of claim 1 , comprising instructions to:

obtain a probability threshold;

determine whether two or more probabilities associated with two or more troubleshooting procedures among the multiple troubleshooting procedures satisfy the probability threshold; and

upon determining that the two or more troubleshooting procedures satisfy the probability threshold, suggest the two or more troubleshooting procedures to perform to resolve the second issue experienced by the second machine.

6 . The non-transitory computer-readable storage medium of claim 1 , comprising instructions to:

receive a result of most likely troubleshooting procedure;

determine whether the result indicates that the second issue is resolved;

upon determining that the result does not indicate that the second issue is resolved, based on the result, generate a second distribution of probabilities associated with the multiple troubleshooting procedures;

select a second most likely troubleshooting procedure,

wherein the second most likely troubleshooting procedure includes a repair step associated with the second issue or the testing step associated with the second issue; and

suggest the second most likely troubleshooting procedure to perform.

7 . A method comprising:

obtaining an indication of multiple troubleshooting guides associated with a first issue experienced by a first device, and device information describing the first device,

wherein a troubleshooting guide among the multiple troubleshooting guides includes multiple troubleshooting procedures to perform,

wherein a troubleshooting procedure among the multiple troubleshooting procedures includes a testing step associated with the first issue and a repair step associated with the first issue;

generating training data for an artificial intelligence using the multiple troubleshooting guides by:

extracting from the multiple troubleshooting guides multiple paths leading to the repair step,

wherein the multiple paths do not include multiple repair steps;

encoding the multiple paths into multiple unique identifiers;

train the artificial intelligence using the multiple unique identifiers to provide a most probable repair step that avoids multiple testing steps;

receiving an input including device information describing a second device, and an indication of a second issue experienced by the second device,

wherein the second device is different from the first device;

based on the input and the indication of the multiple troubleshooting guides, generating a distribution of probabilities associated with the multiple troubleshooting procedures;

based on the distribution of probabilities, providing, using the artificial intelligence, the most probable repair step that avoids multiple testing steps .

8 . The method of claim 7 , comprising:

generating a second troubleshooting guide directed to the second device by:

adding, to the second troubleshooting guide, the indication of the second issue experienced by the second device and most likely troubleshooting procedure;

receiving a result of most likely troubleshooting procedure to obtain a received result; and

performing below steps:

based on the received result, generating a second distribution of probabilities associated with the multiple troubleshooting procedures;

selecting a second most likely troubleshooting procedure, wherein the second most likely troubleshooting procedure includes a repair step associated with the second issue or the testing step associated with the second issue;

adding the second most likely troubleshooting procedure to the second troubleshooting guide;

receiving a result of the second most likely troubleshooting procedure to obtain the received result;

until the received result indicates that the second issue is resolved.

9 . The method of claim 7 , comprising:

generating a second troubleshooting guide directed to the second device by:

adding, to the second troubleshooting guide, the indication of a second issue and most likely troubleshooting procedure to the second troubleshooting guide;

receiving a result of the most likely troubleshooting procedure to obtain a received result; and

performing below steps:

based on the received result, generating a second distribution of probabilities associated with the multiple troubleshooting procedures;

selecting a second most likely troubleshooting procedure,

wherein the second most likely troubleshooting procedure includes a repair step associated with the second issue or the testing step associated with the second issue;

adding the second most likely troubleshooting procedure to the second troubleshooting guide;

receiving a result of the second most likely troubleshooting procedure to obtain the received result;

until the received result indicates that the second issue is resolved;

obtaining a probability threshold;

determining whether two or more probabilities associated with two or more troubleshooting procedures among the multiple troubleshooting procedures satisfy the probability threshold; and

upon determining that the two or more troubleshooting procedures satisfy the probability threshold, creating a branch in the second troubleshooting guide by adding the two or more troubleshooting procedures to perform to the second troubleshooting guide,

wherein the two or more troubleshooting procedures include the most likely troubleshooting procedure or the second most likely troubleshooting procedure.

10 . The method of claim 7 , comprising:

obtaining a probability threshold;

determining whether two or more probabilities associated with two or more troubleshooting procedures among the multiple troubleshooting procedures satisfy the probability threshold; and

upon determining that the two or more troubleshooting procedures satisfy the probability threshold, suggesting the two or more troubleshooting procedures to perform as a resolution to the second issue experienced by the second device.

11 . The method of claim 7 , comprising:

receiving a result of most likely troubleshooting procedure;

determining whether the result indicates that the second issue is resolved;

upon determining that the result does not indicate that the second issue is resolved, based on the result, generating a second distribution of probabilities associated with the multiple troubleshooting procedures;

selecting a second most likely troubleshooting procedure,

wherein the second most likely troubleshooting procedure includes a repair step associated with the second issue or the testing step associated with the second issue; and

suggesting the second most likely troubleshooting procedure to perform.

12 . A system comprising:

at least one hardware processor; and

at least one non-transitory memory storing instructions, which, when executed by the at least one hardware processor, cause the system to:

obtain an indication of multiple troubleshooting guides associated with a first issue experienced by a first device, and device information describing the first device,

wherein a troubleshooting guide among the multiple troubleshooting guides includes multiple troubleshooting procedures to perform,

wherein a troubleshooting procedure among the multiple troubleshooting procedures includes a testing step associated with the first issue and a repair step associated with the first issue;

generate training data for an artificial intelligence using the multiple troubleshooting guides by:

extracting from the multiple troubleshooting guides multiple paths leading to the repair step,

wherein the multiple paths do not include multiple repair steps;

encoding the multiple paths into multiple unique identifiers;

train the artificial intelligence using the multiple unique identifiers to provide a most probable repair step that avoids multiple testing steps;

receive an input including device information describing a second device, and an indication of a second issue experienced by the second device,

wherein the second device is different from the first device;

based on the input and the indication of the multiple troubleshooting guides, generate a distribution of probabilities associated with the multiple troubleshooting procedures;

based on the distribution of probabilities, provide, using the artificial intelligence, the most probable repair step that avoids multiple testing steps.

13 . The system of claim 12 , comprising instructions to:

generate a second troubleshooting guide directed to the second device by:

adding, to the second troubleshooting guide, the indication of the second issue experienced by the second device and most likely troubleshooting procedure;

receiving a result of the most likely troubleshooting procedure to obtain a received result; and

performing below steps:

based on the received result, generate a second distribution of probabilities associated with the multiple troubleshooting procedures;

select a second most likely troubleshooting procedure,

wherein the second most likely troubleshooting procedure includes a repair step associated with the second issue or the testing step associated with the second issue;

add the second most likely troubleshooting procedure to the second troubleshooting guide;

receive a result of the second most likely troubleshooting procedure to obtain the received result;

until the received result indicates that the second issue is resolved.

14 . The system of claim 12 , wherein the artificial intelligence includes a generative artificial intelligence.

15 . The system of claim 12 , comprising instructions to:

generate a second troubleshooting guide directed to the second device by:

adding, to the second troubleshooting guide, the indication of a second issue and most likely troubleshooting procedure to the second troubleshooting guide;

receiving a result of the most likely troubleshooting procedure to obtain a received result; and

performing below steps:

based on the received result, generate a second distribution of probabilities associated with the multiple troubleshooting procedures;

select a second most likely troubleshooting procedure,

wherein the second most likely troubleshooting procedure includes a repair step associated with the second issue or the testing step associated with the second issue;

add the second most likely troubleshooting procedure to the second troubleshooting guide;

receive a result of the second most likely troubleshooting procedure to obtain the received result;

until the received result indicates that the second issue is resolved;

obtain a probability threshold;

determine whether two or more probabilities associated with two or more troubleshooting procedures among the multiple troubleshooting procedures satisfy the probability threshold; and

upon determining that the two or more troubleshooting procedures satisfy the probability threshold, create a branch in the second troubleshooting guide by adding the two or more troubleshooting procedures to perform to the second troubleshooting guide,

wherein the two or more troubleshooting procedures include the most likely troubleshooting procedure or the second most likely troubleshooting procedure.

16 . The system of claim 12 , comprising instructions to:

obtain a probability threshold;

determine whether two or more probabilities associated with two or more troubleshooting procedures among the multiple troubleshooting procedures satisfy the probability threshold; and

upon determining that the two or more troubleshooting procedures satisfy the probability threshold, suggest the two or more troubleshooting procedures to perform to resolve the second issue experienced by the second device.

17 . The system of claim 12 , comprising instructions to:

receive a result of most likely troubleshooting procedure;

determine whether the result indicates that the second issue is resolved;

upon determining that the result does not indicate that the second issue is resolved, based on the result, generate a second distribution of probabilities associated with the multiple troubleshooting procedures;

select a second most likely troubleshooting procedure,

wherein the second most likely troubleshooting procedure includes a repair step associated with the second issue or the testing step associated with the second issue; and

suggest the second most likely troubleshooting procedure to perform.

Assignments (1)

Continuity (1)

References Cited (29)