This paper presents a Dual-Security-Assured Computation Offloading (DSACO) scheme for space-air-ground networks (SAGNs), which exploits the integrated sensing, communication, and computing (ISCC) capability of the low Earth orbit (LEO) satellite to support secure and efficient computation offloading. In the proposed scheme, the air-ground mode serves as the default edge-processing strategy due to its low latency and energy consumption, while the LEO satellite senses the malicious aerial eavesdropper and protects the ground device (GD)-to-UAV links through directional anti-eavesdropping jamming. When such protection becomes insufficient, the system switches to direct GD-to-LEO offloading via dedicated frequency bands. Accordingly, the secure offloading process is formulated as a worst-case average long-term energy minimization problem under sensing uncertainty. To solve the resulting dual-timescale mixed-integer nonlinear problem, we develop a hierarchical multi-agent deep reinforcement learning (H-MADRL) framework for the joint optimization of sensing duration, UAV trajectories, computation offloading, and resource allocation. Simulation results demonstrate that the proposed scheme significantly enhances system security while maintaining offloading efficiency, and that the H-MADRL framework outperforms benchmark methods.

Backscatter communication is emerging as a promising solution for enabling low-power and large-scale IoT applications. However, it faces challenges in terms of widespread deployment, wireless resource management, and quality of service (QoS). In this paper, we first propose a secure transmission architecture to integrate the backscatter network with the existing 5G/IoT infrastructure. Next, we introduce a multi-device scheduling and time-switching strategy aimed at optimizing both capacity and secure throughput. In the time-switching scheme, BDs primarily operate in symbiotic mode without requiring additional spectrum, but can dynamically switch to opportunistic spectrum access mode when necessary, with adaptive time allocation to improve QoS. For multi-device scheduling, BDs are assigned to function as a master transmission node, cooperation node, or spoofing/jamming node, thereby enhancing the system’s resistance to proactive eavesdropping. The optimization problem is formulated to minimize spectrum resource usage while ensuring QoS and following the energy constraint. To solve this, we introduce an enumeration-based interior-point algorithm (EIA) and design a novel progressive greedy algorithm (PGA). The EIA method provides optimal solutions, while the PGA algorithm achieves high-quality suboptimal solutions with lower complexity. Extensive simulation results demonstrate that the proposed strategy stands out in ensuring QoS, enhancing security, and reducing spectrum resource usage.

Physical layer security (PLS) in symbiotic radio (SR) systems is primarily considered for passive eavesdropping scenarios. However, overlooking the impact of proactive eavesdroppers poses significant risks. In this paper, we focus on secure transmission in SR systems under proactive eavesdropping conditions. A PLS strategy is investigated for a full-duplex facultative symbiotic radio (FD-FSR) system. First, we introduce an innovative FSR protocol. It allows backscatter devices (BDs) to dynamically switch between cognitive and symbiotic patterns. Next, we develop a multi-device scheduling method. It adaptively assigns BDs as transmitters, cooperators, or jammers to enhance the secrecy rate. We formulate the pattern switching and BD scheduling as a mixed integer programming problem (MIP). To solve this, we first decompose it into binary decision-making and multi-variable optimization sub-problems. Then, a low-complexity two-stage optimization strategy is employed. Numerical results demonstrate that our proposed strategy significantly outperforms existing schemes.

In post-disaster scenarios, it is challenging for affected users to transmit data as quickly as possible before their residual energy (RE) is exhausted. Besides, the problem of limited users’ RE causes severe transmission delay unfairness within the network. In this paper, we propose a novel two-phase scheme, called energy-aware unmanned aerial vehicle (UAV) relay transmission (EURT), to balance transmission delay of users and network fairness. Specifically, in the first phase, we pair users two-by-two based on their RE and minimize the maximum transmission delay among all pairs by jointly optimizing the bandwidth allocation, transmit powers, and UAV altitude. In the second phase, we design a cognitive radio (CR) inspired rate-splitting multiple access (RSMA) scheduling strategy to obtain the optimal power splitting factor for each pair. This strategy considers the user with a lower RE value in each pair as the primary user (PU) and the other one as the secondary user (SU), then minimizes the transmission delay of the SU while ensuring the quality of service (QoS) of the PU. Furthermore, we propose a novel evaluation framework to explore the degree of impact of RE and channel state information (CSI) on network delay fairness. Simulation results demonstrate that: i) the proposed EURT algorithm effectively improves performance metrics of networks in terms of transmission delays, throughput, energy consumption and energy efficiency; ii) The proposed algorithm achieves a trade-off between the minimum delay and the optimal network fairness by adjusting the QoS threshold of the PU.

Due to the constraints on power supply and limited encryption capability, data security based on physical layer security (PLS) techniques in backscatter communications has attracted a lot of attention. In this work, we propose to enhance PLS in a full-duplex symbiotic radio (FDSR) system with a proactive eavesdropper, which may overhear the information and interfere legitimate communications simultaneously by emitting attack signals. To deal with the eavesdroppers, we propose a security strategy based on pseudo-decoding and artificial noise (AN) injection to ensure the performance of legitimate communications through forward noise suppression. A novel AN signal generation scheme is proposed using a pseudo-decoding method, where AN signal is superimposed on data signal to safeguard the legitimate channel. The phase control in the forward noise suppression scheme and the power allocation between AN and data signals are optimized to maximize security throughput. The formulated problem can be solved via problem decomposition and alternate optimization algorithms. Simulation results demonstrate the superiority of the proposed scheme in terms of security throughput and attack mitigation performance.