Research On Reputation-based Energy Efficient Trust Scheme In Wireless Sensor Networks

Wireless sensor networks have undergone tremendous technologicaladvances over the last few years and are now applied widely in many fields ofmilitary and civilian technology. The core application of wireless sensor networks isto detect and report events. But the sensor nodes are often deployed in an unattended,hostile territories, then the adversary can easily capture a sensor node physically andextract all the secret information stored on board, including the importantcryptographic keys. The malicious nodes are often masqueraded as normal ones，hided within the network and undermine the network security. The traditionalcryptography can t insure the security of wireless sensor networks. So, in order toovercome this unique situation, a new secure scheme for wireless sensor networks isrequired. A new secure scheme called reputation-based trust scheme used as aneffective supplement of wireless sensor network security is introduced in this paper.According to the uncertainty of a trust scheme, a new trust scheme named asnTRUST which is based on Certainty Theory in uncertainty reasoning is proposed.On this basis, in order to solve the problem about high trust nodes oftenrecommending scandalously, and the energy efficiency of the node, a new method isconsidered. Finally, in the simulation environment OMNET++, the experimentresults show that these new secure schemes are feasible and effective.The principal works of this thesis are as followed:1．A new trust scheme named as nTRUST is proposed in wireless sensornetworks. The trust scheme collects the past behaviors of node, such as the originalreputation and then deduce possible future behavior of nodes, cooperative oruncooperative behaviors, with uncertainty. So, a new trust scheme nTRUST based onCertainty Theory is proposed to solve above problems. The mechanism uses theintegrated trust value as the node trust evaluation which is better versatile andfeasible.2．The existing trust schemes, such as RFSN, only accept the recommendationtrust values sent from the nodes whose trust values exceed the threshold set, but can’tcope with scandalous recommendation or libelous attacks from high trust nodes. In order to solve such problems, we further cluster and process the acceptedrecommendation trusts based on setting the trust threshold. At last, the larger datasets will be selected as the final recommendation trust sets. Using this new approach,a good experiment result is obtained.3． Due to the limited energy the sensor nodes carried, the new schemeconsidering node s trust will cause further energy consumption. Meanwhile, with thetrust scheme, when a node has a high trust value, it will be easier to be selected asthe service node, such as routing node. This will cause such these nodes run out ofenergy quickly and be not conducive to the network survival time. Therefore, mainlyfrom the view of energy efficiency and further consideration about residual energy ofthe nodes, two different trust schemes RGR and RFSN are improved our newschemes. The two new trust schemes called RGR-ECC and RFSN-ECC, areproposed. Better results are achieved through experiments. At the same time,combined with exploration on low energy cost in the classic trust scheme, nTRUSTis improved as nTRUST-ECC. The improved model can balance the node s energyconsumption and prolong the whole network lifetime.