Probability-based opportunity dynamic adaptation (PODA) of contention window for home M2M networks

Abstract

With the emergence of the Internet of Things (IoT), the growing use of autonomous sensing and actuating devices in areas such as smart grid, e-healthcare, home networking, and machine-tomachine (M2M) communication has become an important communication paradigm. Nonetheless, to fully exploit the applications facilitated by M2M communication, service requirements such as data throughput, scalability and reliability must be in place. This paper proposes a new backoff adaptation mechanism known as probability-based opportunity dynamic adaptation (PODA) for M2M communication using the IEEE 802.11ah protocol. The proposed PODA is an enhanced version of the binary exponential backoff (BEB) where a station estimates the number of contending stations in a distributed manner and adaptively tunes its minimum contention window (CW) prior to the contention process for better network throughput and packet delivery ratio. Owing to its great flexibility and ease of implementation, BEB has been extended to home M2M communication such as wireless sensor networks and smart grid technologies without relying on wide area communication. However, the current form of BEB has its shortcomings in the emerging M2M paradigm. The adaptation of CW in PODA is based on the optimal station's access opportunity to improve network performance instead of direct CW scaling. Using the proposed adaptation method, the network throughput can be improved by as much as 18 percent in the home M2M network studied, while enhancing network reliability and fairness.

Performance Improvement of Random Access by Prioritizing Collisions

“…Note that the ACPB requires no information on the number of stations in the network and operates in a fully distributed manner. So we compare the performance of the ACPB with three backoff schemes, the A-RAP + [21], the AMOCW [20], and the Probability-based Opportunity Dynamic Adaptation (PODA) [43], which operate completely distributed. In these schemes, each station estimates the number of stations in the network and adjusts the backoff parameters based on this.…”

Deep reinforcement learning-based contention window optimization for IEEE 802.11 networks

“…DDQN, however, diverges from DQN (6). By utilizing gradient descent, it computes a loss function based on the difference between predicted Q-values from the online network and target Q-values.…”

“…For instance, adaptive contention window control (ACWC) [5] utilizes a single backoff stage, while retaining the existing standard IEEE 802.11. Another approach, probability-based opportunity dynamic adaptation (PODA) [6] is the enhanced version of the binary exponential backoff (BEB) algorithm that adjusts the minimum CW before the station enters into the contention process. Additionally, exponential increase exponential decrease (EIED) [7] and linear increase linear decrease (LILD) [8] increase and decrease the CW size adjustment exponentially and linearly, respectively.…”

An Adaptive EDCA Selfishness-Aware Scheme for Dense WLANs in 5G Networks

“…It adjusts the CW size according to the collision probability and uses temporary back-off within the existing back-off counter. In [29], the authors improved the binary exponential backoff (BEB) based on the estimated number of the competing nodes which adapt the CWmin before the contention phase to enhance the overall network throughput and the packet delivery ratio. The authors of [30] improved the wireless full-duplex cognitive MAC protocol that effectively resolved the problem of reactivationfailure in multichannel non-time slotted cognitive radio networks.…”