Fault correction based on meaning or intent of 5G/6G messages
View Patent ↗A receiver may use a trained AI model to recover a faulted 5G/6G message by interpreting the meaning or intent of the message by correlating the message content with one of the “expected” message types. For example, the AI model may consider changes to the message, for consistency with an associated error-detection code, thereby producing a series of candidate messages. The AI model can then determine a likelihood that each of the candidate messages is correct, in the context of the receiver (such as an action or condition of the receiver, or a planned activity of the receiver) or is commonly received in that context. For example, the AI model can be trained to recognize the expected messages or message types, and thereby indicate which candidate message has the highest likelihood of being correct. The AI model may also consider waveform parameters to identify likely faults.
1. A method for a wireless receiver to demodulate a message, the method comprising:
a) receiving a received message and a received error-detection code;
b) determining, according to a disagreement between the received message and the received error-detection code, that the received message or the received error-detection code, or both, is or are corrupted;
c) providing, as input to an artificial intelligence (“AI”) model, the received message and the received error-detection code;
d) determining, as output from the AI model, a plurality of candidate versions, each version comprising:
i) a candidate message;
ii) a candidate error-detection code that is consistent with the candidate message; and
iii) a likelihood value indicating an estimated likelihood that the candidate message is correct;
e) wherein, for each version:
i) when the candidate message differs from the received message, the candidate error-detection code equals the received error-detection code; and
ii) when the candidate error-detection code differs from the received error-detection code, the candidate message equals the received message.
2. The method of claim 1 , wherein the message is transmitted according to 5G or 6G technologies.
3. The method of claim 1 , further comprising providing, as further input to the AI model, a plurality of expected messages.
4. The method of claim 3 , wherein the expected messages comprise messages previously received by the receiver.
5. The method of claim 3 , wherein the expected messages comprise AI-generated messages generated by the AI model or a different AI model.
6. The method of claim 3 , wherein, for each candidate version, the likelihood value is based at least in part on a comparison between the received message and the expected messages.
7. The method of claim 3 , wherein, for each candidate version, the likelihood value is higher when the candidate message is an expected message, and the likelihood value is lower when the candidate message is not an expected message.
8. The method of claim 1 , wherein, for each candidate version, the likelihood value is based at least in part on a meaning or intent of the received message, as determined by the AI model.
9. Non-transitory computer-readable media in a wireless receiver, the non-transitory computer-readable media containing instructions that, when executed by a computing environment, cause a method to be performed, the method comprising:
a) using an artificial intelligence (“AI”) model which has been trained to determine a likelihood value of a message comprising symbols, according to data about the wireless receiver and the message symbols;
b) receiving a received message and a received error-detection code, and determining, according to a disagreement between the received message and the received error-detection code, that the received message or the received error-detection code, or both, is or are faulted;
c) providing, as input to the AI model, the received message and the received error-detection code; and
d) determining, as output from the AI model, a plurality of candidate versions, each candidate version comprising a candidate message, a candidate error-detection code consistent with the candidate message, and a likelihood value indicating a likelihood that the candidate message is correct.
10. The non-transitory computer-readable media of claim 9 , the method further comprising providing, as further input to the AI model, data about a current activity of the wireless receiver, wherein the likelihood value of each candidate version is based at least in part on a determination, by the AI model, whether the candidate message is related to the current activity or is a common response to the current activity.
11. The non-transitory computer-readable media of claim 10 , wherein for each candidate version, the likelihood value is higher when the AI model determines that the candidate message is related to, or a common response to, the current activity, and is lower when the AI model determines that the candidate message is neither related to, nor a common response to, the current activity.
12. The non-transitory computer-readable media of claim 9 , the method further comprising providing, as further input to the AI model, data about prior messages received by the receiver, wherein:
a) the likelihood value of each candidate version is higher when the candidate message is included in the prior messages or is closely related to one of the prior messages, as determined by the AI model; and
b) the likelihood value of each candidate version is lower when the candidate message is not included in the prior messages and is not closely related to any of the prior messages, as determined by the AI model.
13. The non-transitory computer-readable media of claim 9 , wherein the AI model is configured to determine the likelihood value of each candidate message, based at least in part on one or more example messages provided to the AI model during training.
14. The non-transitory computer-readable media of claim 9 , the method further comprising providing, as further input to the AI model, an indication of which symbols of the received message are predetermined by a format rule, wherein each likelihood value is lower when at least one of the symbols violates the format rule, and is higher when all of the symbols of the candidate message obey the format rule.
15. A method for a wireless receiver to correct a message fault, the method comprising:
a) receiving a message comprising an error-detection code, and determining, according to a disagreement between the message and the error-detection code, that the message or the error-detection code, or both, is or are corrupted;
b) providing, as input to an artificial intelligence (“AI”) model, the message and the error-detection code;
c) determining, as output from the AI model, a plurality of candidate messages, and a likelihood value associated with each of the candidate messages, the likelihood value indicating a likelihood that the candidate message is correct.
16. The method of claim 15 , further comprising providing, as further input to the AI model, data about a context of the message, and determining, as further output of the AI model, a meaning or intent of the message, wherein the context comprises a current or planned action related to the message or to the wireless receiver.
17. The method of claim 16 , wherein the current or planned action comprises an action commonly associated with the message, as determined by the AI model.
18. The method of claim 16 , wherein the current or planned action is consistent with a meaning or intent of the message, as determined by the AI model.
19. The method of claim 16 , wherein for each candidate message, the likelihood value is higher when the candidate message is commonly received in the context of the message, as determined by the AI model, and is lower when the candidate message is rarely or never received in the context of the message, as determined by the AI model.
20. The method of claim 15 , further comprising determining, as further output from the AI model, that none of the likelihood values is greater than a threshold value, and responsively transmitting a request for a retransmission of the message.