Research On Methods Of Secure Data Aggregation For The Sensing Layer Of Internet Of Things

Compared with traditional network, "Internet+" achieved great development from the angle of technology aggregation and network coverage object. It creates and promotes the rapid development of the Internet of things(IOT) technology and application. The sensing layer of IOT, which has a multitude of sensor nodes, shows some characteristics, such as collecting information redundancy, node energy sensitivity,network distribution openness, data requirements reliability and so on. Therefore,research on methods of secure data aggregation for the sensing layer of IOT has important theoretical significance and application prospects.According to the characteristics of IOT sensing layer, in order to guarantee the authenticity and reliability of the results from data aggregation of IOT sensing layer, a secure data aggregating model combining with data preprocessing mechanism and node creditability evaluation mechanism is proposed.Data preprocessing is the primary judgment and filter of fusion nodes in regard to sensory plenty of data by sensor nodes before secure data aggregation. Gross error criterion is adopted to identify and eliminate the abnormal data obviously deviated from normal data(or true value) by means of gross error theory. The abnormal data is caused by network failure or hostile attack. It can avoid subsequent node creditability calculation, data aggregating, communication between nodes, etc. In addition, it can prepare the conditions for saving node energy, eliminating the influence of abnormal data to data aggregating results and next efficient processing.Secure data aggregating phase: First of all, nodes creditability are calculated and updated by means of probability and statistics theory. At the same time, a penalty factor is introduced, which helps to achieve the effect of slow increase and quickly decreases and quickly find and identify malicious nodes when calculate and update nodes creditability. Then the data of nodes with high creditability is allowed to involve in data aggregation only. This can not only isolate the effects of malicious node data to aggregation results, but also reduce the amount of calculation data aggregation. Finally,Josang trust model is used to evaluate data aggregation results by means of nodes creditability. It can improve the reliability of data aggregation results further.The simulation experiment results show that the model not only helps to guaranteeauthenticity and reliability of data aggregation results from IOT sensing layer, but also can reduce the calculating overload of data aggregation and the demand for sensor node resources by means of data preprocessing method based on gross error theory. Estimate to data aggregation results can guarantee the security of aggregation data by means of nodes creditability.