29-08-2014, 11:17 AM
Clustering based performance improvement strategy for Mobile Ad-hoc Network
Clustering.pptx (Size: 172.72 KB / Downloads: 11)
Outline
Project Title
Project Objective
Our Approach
Project Plan
Project work done so far..
Next Phase to be implemented..
Comparison of Direct Trust & Indirect Trust Approach
Some Indirect Trust Approaches
References
Project Title
“Clustering based performance improvement strategy for MANET”
Project Objectives
To evaluate the trust using Indirect Trust Computation Approach based on the second hand information.
To validate & analyze the performance of this approach with the help of following experiments:
1.Validation against Deviation.
2.Validation against Attacks
a)Bad Mouthing (BM)
b)Ballot Stuffing (BS)
c)Random Opinion (RO)
From these experiments we will be able to conclude that ,the given approach could be one of the good option for evaluating trust in clustered MANET.
Work Done So Far
Dishonest Recommendation Algorithm implemented on Java Platform
Scenarios created & tested for random recommendations for multiple unknown nodes.
Validation against deviation (Mean Offset(MO)) implemented and tested for Ballot Stuffing(BS) attack.
Mean Offset = |Mean of Dishonest Rcc-Mean of honest Rcc|
Next Phase to be implemented..
Simulation of 3 Attacks :
Ballot Stuffing (BS)(Boost Trust Value)
Bad Mouthing (BM)(Degrade Trust Value)
Random Opinion (RO)(Opposite Recommendations based on T.V)
Mean Offset Deviation for BM & RO Attack.
Comparison with existing approaches (using Matthews correlation coefficient (MCC) )