To Send or Not to Send - Learning MAC Contention

Amuru, SaiDhiraj; Xiao, Yuanzhang; Schaar, Mihaela van der; Buehrer, R. Michael

doi:10.1109/glocom.2015.7417224

Cited by 18 publications

(26 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…A joint scheme of adjusting node transmission power according to vehicle density and adjusting CW according to instantaneous collision rate is proposed in [ 10 ]. Amuru et al [ 11 ] pointed out that the exponential back-off mechanism in WAVE is sub-optimal in throughput performance, especially in an unknown dynamic network environment. They modeled the RTS-CTS handshake mechanism as a Markov decision process, and used a post-decision state (PDS)-based learning algorithm to select the backoff window value according to the system state.…”

Section: Related Workmentioning

confidence: 99%

“…References [ 12 , 13 ] proposed two methods for adjusting CW based on Q-Learning algorithm with different reward mechanisms. In [ 10 , 11 , 12 , 13 ], simulation experiments were performed on their proposed improvement schemes and the WAVE standard scheme or a fixed CW value scheme. The results show that their schemes have improved different communication performances.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Joint Optimization of Multi-Hop Broadcast Protocol and MAC Protocol in Vehicular Ad Hoc Networks

Pei

Wang

Lei

et al. 2021

Sensors

View full text Add to dashboard Cite

Beacon messages and emergency messages in vehicular ad hoc networks (VANETs) require a lower delay and higher reliability. The optimal MAC protocol can effectively reduce data collision in VANETs communication, thus minimizing delay and improving reliability. In this paper, we propose a Q-learning MAC protocol based on detecting the number of two-hop neighbors. The number of two-hop neighbors in highway scenarios is calculated with very little overhead using the beacon messages and neighbor locations to reduce the impact of hidden nodes. Vehicle nodes are regarded as agents, using Q-learning and beacon messages to train the near-optimal contention window value of the MAC layer under different vehicle densities to reduce the collision probability of beacon messages. Furthermore, based on the contention window value after training, a multi-hop broadcast protocol combined with contention window adjustment for emergency messages in highway scenarios is proposed to reduce forwarding delay and improve forwarding reliability. We use the trained contention window value and the state information of neighboring vehicles to assign an appropriate forwarding waiting time to the forwarding node. Simulation experiments are conducted to evaluate the proposed MAC protocol and multi-hop broadcast protocol and compare them with other related protocols. The results show that our proposed protocols outperform the other related protocols on several different evaluation metrics.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Joint Optimization of Multi-Hop Broadcast Protocol and MAC Protocol in Vehicular Ad Hoc Networks

Pei

Wang

Lei

et al. 2021

Sensors

View full text Add to dashboard Cite

show abstract

“…O uso de aprendizado de máquina em protocolos MAC tem aumentado recentemente. Um mecanismo para ajuste dinâmico de tempos de backoff no protocolo MAC CSMA/CA em redes IEEE802.11 com aprendizado por reforçoé apresentado em [Amuru et al 2015]. Em [Choe et al 2020] propõe-se um algoritmo adaptativo para um protocolo MAC, em que se usa Deep-Q Network (DQN) para melhorar o desempenho de broadcast em redes veiculares.…”

Section: Figura 1 Dinâmica Entre Agente E Ambienteunclassified

Um controlador para um protocolo MAC híbrido em redes sem-fio baseado em aprendizado por reforço

Sousa¹,

Sobral²

2020

Anais Da XVIII Escola Regional De Redes De Computadores (ERRC 2020)

View full text Add to dashboard Cite

Neste trabalho, propõe-se um mecanismo baseado em aprendizado por reforço para escolher o modo de operação de um protocolo MAC híbrido, dependendo da carga de tráfego existente na rede sem-fio. Protocolos como esse combinam MAC do tipo CSMA/CA e TDMA, buscando proporcionar maiores vazões (throughput) e menores latências, e têm sido usados em redes de acesso sem-fio com tecnologia IEEE 802.11. Este trabalho apresenta um modelo para um controlador para esse tipo de MAC, o qual usa experiência para aprender a escolher o melhor modo de operação dependendo da percepção sobre o estado da rede.

show abstract

“…There has been emerging work on employing Reinforcement Learning towards handling the channel access control problem in wireless networks. Amuru, et al in [16] formulate the problem of optimizing the IEEE 802.11 backof f mechanism as an MDP, and propose Reinforcement Learning algorithms as a solution. Liu, et al in [17] adopt Reinforcement Learning as an energy-efficient channel sharing technique for wireless sensor networks.…”

Section: Literature Reviewmentioning

confidence: 99%

A Q-Learning Approach With Collective Contention Estimation for Bandwidth-Efficient and Fair Access Control in IEEE 802.11p Vehicular Networks

Pressas

Sheng

Ali

2019

IEEE Trans. Veh. Technol.

View full text Add to dashboard Cite

Daxin (2019) A Q-Learning approach with collective contention estimation for bandwidth-efficient and fair access control in IEEE 802.11p vehicular networks. IEEE Transactions on Vehicular Technology.Abstract-Vehicular Ad hoc Networks (VANETs) are wireless networks formed of moving vehicle-stations, that enable safetyrelated packet exchanges among them. Their infrastructure-less, unbounded nature allows the formation of dense networks that present a channel sharing issue, which is harder to tackle than in conventional WLANs, due to fundamental differences of the protocol stack. Optimizing channel access strategies is important for the efficient usage of the available wireless bandwidth and the successful deployment of VANETs. We present a Q-Learning-based approach to wirelessly network a big number of vehicles and enable the efficient exchange of data packets among them. More specifically, this work focuses on a IEEE 802.11p-compatible contention-based Medium Access Control (MAC) protocol for efficiently sharing the wireless channel among multiple vehicular stations. The stations feature algorithms that "learn" how to act optimally in a network in order to maximise their achieved packet delivery and minimise bandwidth wastage. Additionally, via a Collective Contention Estimation (CCE) mechanism which we embed on the Q-Learning agent, faster convergence, higher throughput and short-term fairness are achieved.

show abstract

To Send or Not to Send - Learning MAC Contention

Cited by 18 publications

References 14 publications

Joint Optimization of Multi-Hop Broadcast Protocol and MAC Protocol in Vehicular Ad Hoc Networks

Joint Optimization of Multi-Hop Broadcast Protocol and MAC Protocol in Vehicular Ad Hoc Networks

Um controlador para um protocolo MAC híbrido em redes sem-fio baseado em aprendizado por reforço

A Q-Learning Approach With Collective Contention Estimation for Bandwidth-Efficient and Fair Access Control in IEEE 802.11p Vehicular Networks

Contact Info

Product

Resources

About