Secure consensus averaging for secure information fusion in sensor networks

In this work, we have examined the problem of distributed consensus averaging over senor networks from a novel point of view considering the need for security. We have proposed a method for incorporating privacy into the scalable average consensus mechanisms. Our proposed method, Random Projections Method (RPM), is lightweight and transparent since it is not based on cryptography and does not require any change in the fusion system. RPM is based on introducing a simple, yet effective pre-fusion algorithm. We mathematically derived the correctness of RPM and analyzed its effect on convergence of the system through simulation. Robustness of RPM against honest-but-curious adversaries is analyzed and it is shown that the proposed method has maximum robustness saving that the victim has at least one non-colluding neighbor.