Research On Enhance The Security Of Information Hiding In WSN

Wireless sensor networks (WSNs) get more and more extensive application inthe real life, especially in the military, medical, public safety and other importantareas. However, sensor nodes are inherent resource constraints in the data-centricsensor networks. The transmission of data is vulnerable to the eavesdropping, forgery,tampering and so on. Especially in the military application, the tracking andmonitoring sensitive information should be effectively protected. Thus, how toimprove the security of information hiding is the focus of our research.Now, there are still a lot of problems for the information hiding in WSNs. Suchas embedding the sensitive information into the ordinary data regularly, then theattacker will easily detect the existence of hidden information; the key management ismissing in the data hiding algorithm, which can not achieve the key updates or otherprecautions; packet loss and error bits in wireless communications can seriouslyaffect the hidden information extraction.In this thesis, two aspects can improve the security of information hiding: first,data embedded rule will avoid the regular watermark embedding, and this rule isavailable to all existing watermarking algorithm; the second is based on the sessionkey management, compensating for the lack of key management of the data hidingalgorithm, which will update and protect the key.The main contributions as follows:1) We proposed Bloom Filter-based sensor network data hiding scheme. In thisscheme, sensitive information and ordinary data are a whole, and the hidden data israndomly and uniformly distributed in the data set, in order to avoid the regularembedding. The hidden data has a good invisibility. It is worth noting that the schemecan make use of any existing data hiding algorithms.This thesis builds a smallprototype system, and the algorithm is implemented in Telosb nodes. The propertiesof hash function can achieve semantic security. Through analyzing invisibility andtampering attacks, experimental results demonstrate that the proposed scheme canefficiently protect the sensitive data, while involving in little additional energyconsumption.2) Multi-class key management scheme for information hidden in WSNs ispresented. The scheme is based on the establishment of session key, achieving the keymanagement of the data hiding algorithm, which will further ensure the securitytransmission of sensitive information.Sink distributes keys of data hiding algorithmto the nodes that need to transmit sensitive information, by anomaly detection or periodically updating keys, to avoid the security information disclosing securityinformation for nodes or sink capture. Performance and security properties of theproposed scheme are studies both analytically and comparatively. Experimentalresults demonstrate that the complexity of the scheme is acceptable, and the schemecan effectively improve in the information security in WSNs compared to thetraditional key management.