Enhancing IEEE 802.11ah for the Internet of things

Abstract

IEEE 802.11ah is a new wireless networking protocol published in 2017 to support the Internet of Things (IoT). The biggest challenge in IEEE 802.11ah is to deal with channel contention among thousands of low-power sensors. In this thesis, we focus on enhancing the performance of IEEE 802.11ah and five original contributions are made: an analytical model for studying RAW (restricted access window) slot assignments, a distributed AID (associate ID) shuffle mechanism for balancing traffic among RAW slots, a timer backoff mechanism for retrieving simultaneously triggered alarm-reports, an adaptive polling scheme to handle alarm-reports with large size, and an interleaved TWT (target wake time) grouping mechanism for scheduling periodic sensor traffic.

To alleviate the channel contention among event-driven sensors, restricted access window (RAW) , or a time interval reserved for a specified group of sensors/stations (STAs), is introduced in IEEE 802.11ah. A RAW can be divided into slots to further reduce frame collisions. To assign STAs into slots/subgroups, round-robin assignment and random assignment can be used. To gain insight into their designs, an analytical model is first constructed to study and compare their performances. We show that round-robin assignment outperforms random assignment in most cases. But when the traffic is heterogeneous, round-robin assignment suffers from the uneven traffic loads among different RAW slots, causing transmission unfairness and hurting the system throughput. To address this issue, a distributed AID shuffle mechanism is proposed to enhance the performance of round-robin assignment.

A special class of event-driven traffic is triggered by an alarm event such as a fire or an earthquake. Alarm-triggered traffic is time-critical, but large amount of concurrent transmissions from many sensors will cause massive frame collisions and excessive delay. To efficiently retrieve alarm-reports, a timer backoff mechanism called limited local extension (LLE) is designed to enhance the distributed coordination function (DCF). While LLE is efficient with small sized alarm-reports, its performance deteriorates with frame size. To address this issue, an adaptive polling scheme is then proposed such that an AP detects an alarm-reporting event based on channel status, and then uses RAWs to orderly retrieve alarm-reports.

In addition to event-driven traffic, many sensors will report their status periodically. In 802.11ah, periodic traffic can be scheduled using target wake time (TWT) mechanism: a sensor only needs to wake up in evenly spaced service periods (SPs) for frame exchange. Since IEEE 802.11ah also supports human-operated devices (e.g. mobile-phones), AP needs to indicate periodic RAWs (PRAWs) to prevent them from accessing SPs reserved by sensors. This incurs significant indication overhead. To minimize the overhead, a side-by-side TWT grouping mechanism can be used to pack SPs from different sensors back-to-back to share a single PRAW. But the delay performance of human-operated devices suffer. To address this issue, a novel interleaved grouping is proposed and studied in this thesis.