| Wireless sensor networks have been widespread applied in the field of militaryand civilian. However, different from traditional networks, wireless sensor networksare usually deployed in dangerous and hostile environment, the structure of sensor isso simple with low cost and energy limited that can be easily captured andcompromised by various attacks, interfering with user decisions, and even makes thenetworks become unreliable. So how to ensure the collected data safe, reliable andcomplete at the low cost of energy consumption have become the issues that theresearchers need to be resolved. This paper mainly opens out the research on how toensure data collected by sensor networks safely and reliable based on Cloud modelwatermarking technology, the main works of this paper are as follows:Firstly, this paper studies the differential watermarking based on cloud model. Wefirst get the cloud drop watermarking tag by cooperative modulating the collectedsensor data, then can get the binary watermark after processing by hash function, lastusing re-round differential to embed n watermark into n sensor data. The duplicatedata packets can be eliminated by transit node through timestamp and the index vectorformed by timestamp and the identity ID of node. Theoretical analysis and experimentshow that the proposed algorithm can meet the blind extraction and detection ofdistributed wireless sensor networks, has a certain vulnerability to resist thecompromised node and tampering sensor data. The algorithm also has certainrobustness by analyzing the overall characteristics of the cloud watermark model.Secondly, the lossy of in-network data processing usually presents in sensornetwork data transmission. It will cut down the number of data amount but make theresult of data and watermarking in defective condition. In order to resist in-networklossy data processing, the data aggregation with Cloud model watermarkingtechnology is explored in this paper. We embedded the sort of cloud dropwatermarking tag in the q bit LSB, that method come from the characteristic of imagewatermarking can be embedded by some sub-block. The q bit LSB is chosen by thenode’s NID. The simulation results and analysis show that this method can make fulluse fragile of Cloud drop Drop(x,y) and get perfect effect of data be distort or not.Attacker can’t get the complete Cloud model watermarking information from theCloud drops tag so that it verifies the robustness of the proposed method. Thirdly, a simulator system for Cloud model watermarking in sensor networksbased on OMNeT++is designed and developed. Simulation results show that the twoproposed watermarking schemes can ensure data security authentication under lowpower consumption. The first scheme makes full use of the synergy characteristics ofsensor data to generates the embed watermark, has a good effect to ensure dataintegrity validation and resist various kinds of attacks. The second scheme views onecloud as an image combining the idea of image watermark to embedded in a singlesensor data in sub-cluster, can resist the watermark incomplete caused by dataprocessing in networks effectively, especially has a high sensitivity to tempering datawhile processing the feature subset selection. |
