21
Wireless Pers Communication (2014) 74:803–821 Springer Science + Business Media New York 2013 Authors: Gyanendra Prasad Joshi Seung Yeob Nam SungWon Kim Presented By: Iffat Anjum Date : 22/03/2014

Decentralized Predictive MAC Protocol for Ad Hoc Cognitive Radio Networks

Embed Size (px)

DESCRIPTION

Introduction Related Work System Model Protocol Description Control Operation Incumbents’ Reclaiming Resolution and Spectrum Handoff Channel Access Delay in CN Window Conclusion

Citation preview

Page 1: Decentralized Predictive MAC Protocol for Ad Hoc Cognitive Radio Networks

Wireless Pers Communication (2014) 74:803–821

Springer Science + Business Media New York 2013

Authors:Gyanendra Prasad JoshiSeung Yeob NamSungWon Kim

Presented By:Iffat Anjum

Date : 22/03/2014

Page 2: Decentralized Predictive MAC Protocol for Ad Hoc Cognitive Radio Networks

2

Page 3: Decentralized Predictive MAC Protocol for Ad Hoc Cognitive Radio Networks

• Cognitive radio (CR) is a form of wireless communication in which a transceiver can intelligently detect which communication channels are in use and which are not, and instantly move into vacant channels while avoiding occupied ones.

• Research Challenges:

▫ Spectrum scarcity.

▫ Expensive spectrum license.

▫ Determining spectrum opportunity.

▫ Ensuring incumbent licensees’ right.

▫ Computing complexity and flexibility.

▫ Traditional hidden node and exposed node problem.

▫ Multichannel hidden node problem.

3

Page 4: Decentralized Predictive MAC Protocol for Ad Hoc Cognitive Radio Networks

• PMAC is a Synchronization based multi-transceiver and multichannel MAC protocol.

• Historical prediction method used for channel selection algorithm.

• Distributed sensing for spectrum opportunity.

• Cooperative sensing.

• Utilization of Channel Negotiation(CN) window.

• Decentralized, flexible and low complexity.

• Spectrum hand-off method.

4

Page 5: Decentralized Predictive MAC Protocol for Ad Hoc Cognitive Radio Networks

• IEEE 802.22:

▫ Standardized a MAC layer based on CR[Cognitive Radio] for reusing the spectrum allocated to the TV broadcast service.

▫ But, the architecture is centralized.

• DCSS (Distributed coordinated spectrum sharing MAC protocol for cognitive radio)[Nan, H., Hyon,T.,&Yoo; IEEE symposium on new Frontiers in

DySpan 2007]

▫ RTS-CTS based data communication.

▫ Multiple antenna-based protocol.

▫ Negotiates for a data channel per data packet.

▫ Relies fully on a dedicated common CCC[Common Control Channel].

▫ All nodes need to contend for access to the control channel and data channels remain underutilized.

5

Page 6: Decentralized Predictive MAC Protocol for Ad Hoc Cognitive Radio Networks

• MMAC-CR (A distributed multichannel MAC protocol for multi-hop cognitive radio networks)[Timmers,M., Pollin, S., Dejonghe, A.,Van der

Perre, L.,&Catthoor, F. (2010); IEEE Transactions on Vehicular Technology, 59(1), 446–459.]

▫ Time is divided into an ad hoc traffic indication message (ATIM) window and data window.

▫ One handshake on the CCC is needed per connection during the BI.

▫ Have to wait until the end of ATIM window to send and receive data packets.

▫ Multiple channel switching overheads

▫ Multichannel hidden node problems.

6

Page 7: Decentralized Predictive MAC Protocol for Ad Hoc Cognitive Radio Networks

• OC-MAC (Cross-layer based opportunistic MAC protocols for QoS provisioning over cognitive radio wireless networks.)[Su, H., & Zhang, X. (2008); . IEEE Journal on Selected Areas in Communications, 26(1), 118–129 ]

▫ Two transceivers, one is for a dedicated control channel and the other is for selected data channel.

▫ P-persistent carrier sense multiple access is used in the negotiation phase for negotiation.

▫ But, it cannot guarantee fairness to the Su’s[Secondary Users].

7

Page 8: Decentralized Predictive MAC Protocol for Ad Hoc Cognitive Radio Networks

• {Ch|Chi , i = 1, 2, . . . N} non-overlapping incumbent license channels (LCs) conditionally and opportunistically accessible by the SUs.

• Dedicated CCC available with all the required qualities for reliable communication for all times.

• Each node is synchronized by periodic beacon transmission to learn the network-wide spectral opportunities.

▫ Secondary Transmission are pushed on those times.

8

Page 9: Decentralized Predictive MAC Protocol for Ad Hoc Cognitive Radio Networks

9

Figure: Structure of the P-MAC (a view from time-channel domain)

Page 10: Decentralized Predictive MAC Protocol for Ad Hoc Cognitive Radio Networks

• Each CR[cognitive radio] device is equipped with two half-duplex transceivers

▫ Control transceiver.

Wakes up before fast sensing ends, at middle of data window.

Enters into the doze state after the CN window.

▫ Data transceiver.

Capable of switching multiple types of frequencies.

Send data and receive ACK.

Responsible for fine sensing and fast sensing.

10

Page 11: Decentralized Predictive MAC Protocol for Ad Hoc Cognitive Radio Networks

• Incumbent Detection or Sensing:

▫ Fast sensing (time < 1ms/channel) gives three results:

channel is busy,

channel is idle and

Uncertain.

▫ Fine sensing

Provide sensing accuracy but at the cost of time. Fine sensing gives two results:

▫ channel is busy.

▫ channel is idle.

11

Page 12: Decentralized Predictive MAC Protocol for Ad Hoc Cognitive Radio Networks

• Distributed Spectrum Sensing:

▫ P-MAC employs distributed spectrum sensing (DSS) to prevent incumbent-SU collision.

▫ At the very beginning,

the nodes wait for a synchronization signal.

When a node receives information about the next BI, it selects a channel randomly for fast sensing (FS).

From the next BI, as a node receives information of the other channels from distributed sensing

it selects a channel with the least nodes performing FS.

▫ After sensing,

each SU updates the current channel status by its own sensing result &

overhearing the neighbors’ CNs on the CCC.

12

Page 13: Decentralized Predictive MAC Protocol for Ad Hoc Cognitive Radio Networks

13

Table : Channel status table (CST) CUL :channel usage list SNR :signal to noise ratio

The SUs maintain channel ranking according to the channel selection factor (CSF)

Page 14: Decentralized Predictive MAC Protocol for Ad Hoc Cognitive Radio Networks

14

• Nodes has a minimum back-off of tm,

tm = tf + tb

tf -> time for fast sensing,

tb -> random back-off time, 0 < tb ≤ CW.

• Each node maintains its own level for PCL [preferred channel list]

▫ Send it with the CN packet, or with the CN-ACK.

• The PCL is derived from the CSF[channel selection factor].

Control Operation

Ut (Chi )-> Utilization ofchannel i by the incumbent

SNRmin->threshold for the minimum SNR required

SNR(Chi ) ->SNR of channel i

Page 15: Decentralized Predictive MAC Protocol for Ad Hoc Cognitive Radio Networks

15

• P-MAC uses linear HPM[Historical Prediction Model] to predict the incumbent’s arrival.

• After receiving the CN message from any sender that is destined to the receiver,

▫ It selects the channel common to both and with the best PCL level.

▫ Sends CN-ACK including RCPCL[receiver’s common PCL] to the sender.

Control Operation

Ut−1 (Chi ) ->channel utilization in the last time slot γ ->smoothing factorˆU (Chi ) ->average utilization of the time slot of channel i in the pastη-> number of records kept so far

Page 16: Decentralized Predictive MAC Protocol for Ad Hoc Cognitive Radio Networks

16

• The sender sends a CN reservation packet (CN-RES) to confirm the channel reservation after receiving the RCPCL.

• Neighboring nodes update their channel status.

• The random back-off time (RBT) is calculated,

If the current packet has its first transmission, CW is set to W. CW is doubled after each collision with this packet, until it reaches Wm.

• The CN/CN-ACK packet contains network allocation vector (NAV) information to avoid the hidden terminal problem.

Control Operation

Page 17: Decentralized Predictive MAC Protocol for Ad Hoc Cognitive Radio Networks

17

• In PMAC, the nodes do not wait to send data until the start of the data window.

▫ Maximum achievement in

each BI with this strategy is,

• If there is no data in the transmission queue and no packets to receive,

▫ It begins fine sensing (FiS) after the CN window.

▫ The node enters the doze state after fine sensing.

Data Operation

Prs ->success probability of a CR node for channel access CNW ->size of CN window CN-> number of available data channels

Page 18: Decentralized Predictive MAC Protocol for Ad Hoc Cognitive Radio Networks

18

• When the incumbents reclaim the band, the affected communication links of the SUs will be lost.

• On the other hand, if the SUs can sense idle sub-bands, they can reconstruct the communication links to them.

• The channel access delay[Contention Delay] is the time a node spends to obtain access to the channel.

• The contention window size Wi in the back-off stage i:

Incumbents’ Reclaiming Resolution and Spectrum Handoff

Channel Access Delay in CN Window

Page 19: Decentralized Predictive MAC Protocol for Ad Hoc Cognitive Radio Networks

• X denote the MAC layer access delay.

• The expectation of X can be expressed as:

19

Channel Access Delay in CN Window

Page 20: Decentralized Predictive MAC Protocol for Ad Hoc Cognitive Radio Networks

• The simulation results show that the P-MAC protocol increases the good-put and decreases the delay.

• Although P-MAC uses two transceivers, it turns off the transceivers to conserve energy without decreasing SU performance.

• The analysis results show that the CN window becomes a bottleneck as the number of active nodes increases. More study will be needed to address the CCC bottleneck problem.

20

Page 21: Decentralized Predictive MAC Protocol for Ad Hoc Cognitive Radio Networks

21