Transcript
Page 1: COGNITIVE NETWORK ACCESS USING FUZZY DECISION MAKING Nicola Baldo and Michele Zorzi Department of Information Engineering – University of Padova, Italy

COGNITIVE NETWORK ACCESS USING FUZZY DECISION MAKING

Nicola Baldo and Michele ZorziDepartment of Information Engineering –

University of Padova, Italy

Presented By: Andrew D’SouzaPetar Kramaric,Srdjan Lakovic

RYERSON UNIVERSITY

Page 2: COGNITIVE NETWORK ACCESS USING FUZZY DECISION MAKING Nicola Baldo and Michele Zorzi Department of Information Engineering – University of Padova, Italy

• To achieve maximum performance or throughput for connecting to a wireless network.

• To identify a solution which is able to work well and adapt to various scenarios

RYERSON UNIVERSITY

Topic Problem:

Page 3: COGNITIVE NETWORK ACCESS USING FUZZY DECISION MAKING Nicola Baldo and Michele Zorzi Department of Information Engineering – University of Padova, Italy

• Several schemes have been put into practice:– Highest RSSI Scheme– Linked Capacity Scheme– Network Capacity Scheme– Low-Delay Scheme

• Problem: these schemes consider specific wireless technologies (802.11).

• Problem: these schemes target scenarios in which the wireless link is the bottleneck.

RYERSON UNIVERSITY

Previous Implementations

Page 4: COGNITIVE NETWORK ACCESS USING FUZZY DECISION MAKING Nicola Baldo and Michele Zorzi Department of Information Engineering – University of Padova, Italy

• The approach proposed: cognitive network access using fuzzy decision making.

• Fuzzy arithmetic is used to evaluate the communication quality from each access point (AP).

• From this the most suitable access point is selected.

RYERSON UNIVERSITY

Proposed Implementation

Page 5: COGNITIVE NETWORK ACCESS USING FUZZY DECISION MAKING Nicola Baldo and Michele Zorzi Department of Information Engineering – University of Padova, Italy

• Concentrate specifically on solving communication performance issues.

• Specifically throughput, delay, and reliability.• The proposed scheme can adapt to various

technologies.• Cognitive because it makes use of Fuzzy Decision

Making.• The type of network model being used is a

cognitive network model.

RYERSON UNIVERSITY

Proposed Implementation [2]

Page 6: COGNITIVE NETWORK ACCESS USING FUZZY DECISION MAKING Nicola Baldo and Michele Zorzi Department of Information Engineering – University of Padova, Italy

• Different components of communication performance:– Radio link performance– Transport layer performance– Core network performance– User application requirements

• Using known eqn’s to find the above components, the paper produces the following formulas

RYERSON UNIVERSITY

Proposed Methodology

Page 7: COGNITIVE NETWORK ACCESS USING FUZZY DECISION MAKING Nicola Baldo and Michele Zorzi Department of Information Engineering – University of Padova, Italy

• The network layer end-to-end performance for each AP i is determined by (1):

• Then, transport-layer performance is derived (2):

RYERSON UNIVERSITY

Proposed Methodology [2]

Page 8: COGNITIVE NETWORK ACCESS USING FUZZY DECISION MAKING Nicola Baldo and Michele Zorzi Department of Information Engineering – University of Padova, Italy

• To obtain an overall measure of the fitness of AP i to meet the users needs:

• Derives to:

RYERSON UNIVERSITY

Proposed Methodology [3]

Page 9: COGNITIVE NETWORK ACCESS USING FUZZY DECISION MAKING Nicola Baldo and Michele Zorzi Department of Information Engineering – University of Padova, Italy

• Step 1:– Collect fuzzy performance metrics– Throughput, Delay and Reliability for radio link, core network, end-to-

end, transport and application requirements– Application requirements produced by the application– Radio Link metrics provided by the AP– Transport Layer Performance (end-to-end) collected in two ways:

• Direct measurement• Estimates calculated by the cognitive engine

– Core Network Performance measured by all peers

RYERSON UNIVERSITY

Algorithm

Page 10: COGNITIVE NETWORK ACCESS USING FUZZY DECISION MAKING Nicola Baldo and Michele Zorzi Department of Information Engineering – University of Padova, Italy

• Step 2:– Process the the metrics collected using proposed

formulas– The network layer performance for each AP is

determined by combining Radio Link and Core Network performance

– The transport Layer is derived by applying an extension principle

RYERSON UNIVERSITY

Algorithm [2]

Page 11: COGNITIVE NETWORK ACCESS USING FUZZY DECISION MAKING Nicola Baldo and Michele Zorzi Department of Information Engineering – University of Padova, Italy

• Step 3:– The fuzzy metrics calculated provide an estimate

of the communication performance– In this step we compare them with the estimates

of the application requirement– The degree of fitness for a particular AP is defined

RYERSON UNIVERSITY

Algorithm [3]

Page 12: COGNITIVE NETWORK ACCESS USING FUZZY DECISION MAKING Nicola Baldo and Michele Zorzi Department of Information Engineering – University of Padova, Italy

• Set two Access Points– Two mobile device (N95) acting as AP using 3G

connection• Java program:– Runs on the client and gathers data from our

cognitive network database– Process data using proposed formulas– Display the suitability of both nodes

RYERSON UNIVERSITY

Implementation

Page 13: COGNITIVE NETWORK ACCESS USING FUZZY DECISION MAKING Nicola Baldo and Michele Zorzi Department of Information Engineering – University of Padova, Italy

• How to deal with users that maliciously provide wrong information to influence other nodes decisions

• Identification of effective means and strategies to achieve information sharing in Cognitive Radio Networks

RYERSON UNIVERSITY

Future Work

Page 14: COGNITIVE NETWORK ACCESS USING FUZZY DECISION MAKING Nicola Baldo and Michele Zorzi Department of Information Engineering – University of Padova, Italy

RYERSON UNIVERSITY

LA

Page 15: COGNITIVE NETWORK ACCESS USING FUZZY DECISION MAKING Nicola Baldo and Michele Zorzi Department of Information Engineering – University of Padova, Italy

• Numerical results show that the proposed (cognitive network) scheme performs significantly better than state of the art solutions, in terms of both overall performance and fairness.

• This scheme is suitable for multi-technology scenarios, not just the 802.11 technologies that are in current use.

RYERSON UNIVERSITY

Proposed Conclusion

Page 16: COGNITIVE NETWORK ACCESS USING FUZZY DECISION MAKING Nicola Baldo and Michele Zorzi Department of Information Engineering – University of Padova, Italy

Results from Study

RYERSON UNIVERSITY

Page 17: COGNITIVE NETWORK ACCESS USING FUZZY DECISION MAKING Nicola Baldo and Michele Zorzi Department of Information Engineering – University of Padova, Italy

RYERSON UNIVERSITY

Questions?


Top Related