Research On WSNs Trust Model Based On Cloud Model

In recent years, with the progress of the network technology, Internet of Things(IoT) is rapidly popularized to be applied into different fields, and it is playing more and more important role in our life. However, since in most scenarios, there are so many people involving it, abundant individual information come with much risk. A huge security risk to individuals and businesses is brought when the information is leaked. Unlike the Internet, wireless sensor networks, perception layer of IoT, can't be protected by security schemes of Internet, which is why it need new better security scheme. But, the focuses of current schemes are more on outside attacks than inside attacks. In addition, energy consumption has been the core issue of sensor network, so how to solve these threats is the key.After analyzing exiting security solutions, in order to solve those problems above-mentioned, the communication trust and energy trust among nodes are researched in detail, specific research contents and innovations are as follows:Consider the problems of decision-making and recognizing malicious nodes in traditional node reputation security scheme based on light-weighted cloud model, this thesis proposes a Modified Light-weighted Cloud Model(MLCM), and a novel node reputation security scheme is given by then. Meanwhile, on the basis of node trust, determine whether node energy of next hop is reliable. Energy Prediction scheme based on Cloud Model(EPCM) is proposed.1. In order to resolve trust misjudgment when node launches single attack, before and after being attacked, direct and indirect trust values of nodes about communication state are treated comprehensively before its calculation with simplified backward cloud operator, on the basis of the traditional approach of wireless sensor network trust management on trust value. The experiment shows that the scheme can figure out matter of decision-making.2. Consider the problem of calculation on recommendation node trust in traditional reputation model; cloud membership function is leveraged to obtain recommendation trust value to improve accuracy of recognizing malicious nodes.3. Consider uncertainty of node energy in sensor networks, cloud model is introduced to divide concept of energy consumption into several cloud concepts. Energy history cloud and energy trend cloud are proposed, these two concepts are weighted to form energy prediction rule, which improve accuracy of node energy prediction in short term.