Adaptive power save with reinforcement learning

An apparatus for wireless communications, comprising:

• means for communicating, while in a first power mode associated with an active communication state, with an access point (AP), wherein data packets can be exchanged with the AP while in the first power mode;
• means for switching to a second power mode associated with a sleep state based at least in part on a first inactivity timeout (ITO) associated with the active communication state, wherein data packets cannot be exchanged with the AP while in the second power mode;
• means for selecting a second ITO that is shorter in time than the first ITO based at least in part on a traffic indicator metric associated with an observation window associated with a plurality of speculative wakeups from the sleep state, the traffic indicator metric being associated with a quantity of traffic indicator map messages received from the AP during the observation window;
• and
• means for performing, based at least in part on the traffic indicator metric being less than a traffic indicator map message threshold, an additional speculative wakeup and switching to the first power mode, wherein a duration of the additional speculative wakeup is based at least in part on the second ITO.

An apparatus for wireless communications, comprising a processor, memory in electronic communication with the processor, and instructions stored in the memory and operable, when executed by the processor, to cause the apparatus to:

• communicate, while in a first power mode associated with an active communication state, with an access point (AP), wherein data packets can be exchanged with the AP while in the first power mode;
• switch to a second power mode associated with a sleep state based at least in part on a first inactivity timeout (ITO) associated with the active communication state, wherein data packets cannot be exchanged with the AP while in the second power mode;
• select a second ITO that is shorter in time than the first ITO based at least in part on a traffic indicator metric associated with an observation window associated with a plurality of speculative wakeups from the sleep state, the traffic indicator metric being associated with a quantity of traffic indicator map messages received from the AP during the observation window;
• and
• perform, based at least in part on the traffic indicator metric being less than a traffic indicator map message threshold, an additional speculative wakeup and switch to the first power mode, wherein a duration of the additional speculative wakeup is based at least in part on the second ITO.

A non-transitory computer-readable medium storing code for wireless communications, the code comprising instructions executable by a processor to:

• communicate, while in a first power mode associated with an active communication state, with an access point (AP), wherein data packets can be exchanged with the AP while in the first power mode;
• switch to a second power mode associated with a sleep state based at least in part on a first inactivity timeout (ITO) associated with the active communication state, wherein data packets cannot be exchanged with the AP while in the second power mode;
• select a second ITO that is shorter in time than the first ITO based at least in part on a traffic indicator metric associated with an observation window associated with a plurality of speculative wakeups from the sleep state, the traffic indicator metric being associated with a quantity of traffic indicator map messages received from the AP during the observation window;
• and
• perform, based at least in part on the traffic indicator metric being less than a traffic indicator map message threshold, an additional speculative wakeup and switch to the first power mode, wherein a duration of the additional speculative wakeup is based at least in part on the second ITO.