System and method for throughput-based optimization for target wake time interval adjustment

1. A method comprising:

obtaining wireless traffic statistics indicating current conditions and predicted future trends for at least one of throughput, latency, and device power consumption;

generating penalty functions that reflect the wireless traffic statistics, each of the penalty functions comprising a first term associated with a TWT introduced latency and a second term associated with a wasted power ratio;

determining a target wakeup time (TWT) interval based on the penalty functions; and

adapting the TWT interval based on a change in the wireless traffic statistics.

2. The method of claim 1 , wherein obtaining the wireless traffic statistics indicating the predicted future trends comprises:

predicting future traffic statistics by applying a time series prediction model to observed traffic statistics of one or more observation windows.

3. The method of claim 1 , wherein the penalty functions comprise at least one of:

a first penalty function based on an anchor latency or an anchor wasted power ratio,

a second penalty function based on upper and lower bounds of the latency and the wasted power ratio, and

a third penalty function based on a best minimum latency and a best minimum wasted power ratio.

4. The method of claim 1 , wherein determining the TWT interval based on the penalty functions comprises:

optimizing an objective function that includes the penalty functions, the objective function representing a trade-off between TWT introduced latency and TWT introduced power consumption.

5. The method of claim 1 , wherein adapting the TWT interval comprises:

upon a determination that zero packets are transmitted or received during multiple observation windows, increasing the TWT interval; and

upon a determination that (i) duty cycles during the multiple observation windows are higher than a pre-defined duty cycle threshold and (ii) the TWT interval is larger than a pre-defined interval threshold, decreasing the TWT interval to a pre-defined interval value.

6. The method of claim 1 , wherein the wireless traffic statistics correspond to a wireless device that is in a TWT early termination mode.

7. A device comprising:

a memory configured to store instructions; and

a processor operably connected to the memory, the processor configured when executing the instructions to:

obtain wireless traffic statistics indicating current conditions and predicted future trends for at least one of throughput, latency, and device power consumption;

generate penalty functions that reflect the wireless traffic statistics, each of the penalty functions comprising a first term associated with a TWT introduced latency and a second term associated with a wasted power ratio;

determine a target wakeup time (TWT) interval based on the penalty functions; and

adapt the TWT interval based on a change in the wireless traffic statistics.

8. The device of claim 7 , wherein to obtain the wireless traffic statistics indicating the predicted future trends, the processor is configured to:

predict future traffic statistics by applying a time series prediction model to observed traffic statistics of one or more observation windows.

9. The device of claim 7 , wherein the penalty functions comprise at least one of:

a first penalty function based on an anchor latency or an anchor wasted power ratio,

a second penalty function based on upper and lower bounds of the latency and the wasted power ratio, and

a third penalty function based on a best minimum latency and a best minimum wasted power ratio.

10. The device of claim 7 , wherein to determine the TWT interval based on the penalty functions, the processor is configured to:

optimize an objective function that includes the penalty functions, the objective function representing a trade-off between TWT introduced latency and TWT introduced power consumption.

11. The device of claim 7 , wherein to adapt the TWT interval, the processor is configured to:

upon a determination that zero packets are transmitted or received during multiple observation windows, increase the TWT interval; and

upon a determination that (i) duty cycles during the multiple observation windows are higher than a pre-defined duty cycle threshold and (ii) the TWT interval is larger than a pre-defined interval threshold, decrease the TWT interval to a pre-defined interval value.

12. The device of claim 7 , wherein the wireless traffic statistics correspond to a wireless device that is in a TWT early termination mode.

13. A non-transitory computer readable medium comprising a plurality of instructions that, when executed by at least one processor, is configured to cause the at least one processor to:

obtain wireless traffic statistics indicating current conditions and predicted future trends for at least one of throughput, latency, and device power consumption;

generate penalty functions that reflect the wireless traffic statistics, each of the penalty functions comprising a first term associated with a TWT introduced latency and a second term associated with a wasted power ratio;

determine a target wakeup time (TWT) interval based on the penalty functions; and

adapt the TWT interval based on a change in the wireless traffic statistics.

14. The non-transitory computer readable medium of claim 13 , wherein to obtain the wireless traffic statistics indicating the predicted future trends, the plurality of instructions is configured to cause the at least one processor to:

predict future traffic statistics by applying a time series prediction model to observed traffic statistics of one or more observation windows.

15. The non-transitory computer readable medium of claim 13 , wherein the penalty functions comprise at least one of:

a first penalty function based on an anchor latency or an anchor wasted power ratio,

a second penalty function based on upper and lower bounds of the latency and the wasted power ratio, and

a third penalty function based on a best minimum latency and a best minimum wasted power ratio.

16. The non-transitory computer readable medium of claim 13 , wherein to determine the TWT interval based on the penalty functions, the plurality of instructions is configured to cause the at least one processor to:

optimize an objective function that includes the penalty functions, the objective function representing a trade-off between TWT introduced latency and TWT introduced power consumption.

17. The non-transitory computer readable medium of claim 13 , wherein to adapt the TWT interval, the plurality of instructions is configured to cause the at least one processor to:

upon a determination that zero packets are transmitted or received during multiple observation windows, increase the TWT interval; and

upon a determination that (i) duty cycles during the multiple observation windows are higher than a pre-defined duty cycle threshold and (ii) the TWT interval is larger than a pre-defined interval threshold, decrease the TWT interval to a pre-defined interval value.

18. The method of claim 1 , wherein the wireless traffic statistics comprise at least one of a total packet size and a data rate.

19. The device of claim 7 , wherein the wireless traffic statistics comprise at least one of a total packet size and a data rate.

20. The non-transitory computer readable medium of claim 13 , wherein the wireless traffic statistics comprise at least one of a total packet size and a data rate.