Security design for integrated communication and computation networks through cooperative cognitive radios

In this letter, a novel integrated communication and computation framework is proposed through cooperative cognitive radios, which includes a communication primary network (PN) and a computing secondary network. Specifically, two primary communication users intend to exchange their confidential information with the relay assistance of cognitive access point (CAP), meanwhile, multiple sensors share the spectrum of PN to transmit their concurrent sensing data to CAP. The CAP is assumed to be a potential eavesdropper and to ensure the secure communication of primary users (PUs) and computational performance of the CAP, it is proposed to maximize the sum achievable secrecy rate of PUs by jointly designing the transmit precoding, relay beamforming matrix, transmit scaling factors and aggregation beamforming vector, while satisfying the constraint of the computational mean square error of the CAP. A weighted minimum mean square error (WMMSE)‐based block coordinate descent (BCD) algorithm is developed to solve the formulated problem, and simulation results are provided to validate the effectiveness of our proposed schemes. In this letter, we propose a novel integrated communication and computation framework through cooperative cognitive radios, which includes a communication primary network (PN) and a computing secondary network. Specifically, two primary communication users intend to exchange their confidential information with the relay assistance of cognitive access point (CAP), meanwhile, multiple sensors share the spectrum of PN to transmit their concurrent sensing data to the CAP. We assume the CAP is a potential eavesdropper and to ensure the secure communication of primary users (PUs) and computational performance of the CAP, we propose to maximize the sum achievable secrecy rate of PUs by jointly designing the transmit precoding, relay beamforming matrix, transmit scaling factors and aggregation beamforming vector, while satisfying the constraint of the computational mean square error of the CAP.