Leveraging NOMA and artificial noise in secure STAR-RIS-assisted UAV-MEC networks

This paper investigates a mobile edge computing (MEC) network empowered by non-orthogonal multiple access (NOMA) to enhance security and reduce computational load. The system comprises an unmanned aerial vehicle (UAV), a simultaneously transmitting and reflecting reconfigurable intelligent surface (STAR-RIS), two access points (APs), and a passive eavesdropper. The STAR-RIS employs an energy splitting (ES) protocol, enabling its elements to simultaneously reflect and transmit signals. To enhance physical-layer security, the UAV generates artificial noise (AN) to interfere with the eavesdropper, while optimally leveraging the STAR-RIS for directional signal delivery to legitimate APs. The system performance is evaluated based on the secure computation probability, reflecting the likelihood of successful and confidential task offloading. Numerical results demonstrate significant improvements in system security by jointly optimizing UAV altitude and AN power allocation