Cross-layer CS based secure compression for energy-efficient sensor data collection
Cryptographic techniques are used in wireless sensor networks to resist physically insecure environments. However the use of these techniques such as encryption and hashing largely increases the energy consumption of sensors, which aggravates the originally critical energy constraint problem of wireless sensor networks. In this paper, an energy-efficient and compressive sensing based secure compression scheme for correlated multimedia sensor data collection is presented. It incorporates effective cross-layer strategy that concerning PHY layer, MAC layer to realize energy-minimization. Secure compression combines encryption and integrity checking with compressive sensing, which is particularly well suited since it charges very low for computation and energy. Energy consumption minimization is formulated as a non-linear convex optimization problem and the optimum transmission structure is determined in terms of transmit power, source coding rate and slot allocation. Numeric analysis for energy minimization is presented, and a system-level evaluation of secure compression is implemented on hardware platform to demonstrate the effectiveness of our scheme. The results suggest that the use of cross-layer secure compression can promote energy-efficiency.