Sensing-Transmission-Energy Trade-Off for IRS-Assisted Cognitive UAV Communication Systems

In this paper, an energy-efficient intelligent reflecting surface-assisted cognitive unmanned aerial vehicle communication system (IRS-CUAVCS) is studied, where the moving CUAV acts as a bidirectional sensing-transmission (BST) node, while IRS is arranged to enhance the BST capability of CUAV and suppress the interference between CUAV and primary networks. The sensing-transmission-energy trade-off problem is formulated to maximize the energy efficiency of IRS-CUAVCS by jointly optimizing the CUAV’s sensing duration, transmitted power, trajectory, and the IRS’s BST phase shift matrices (PSMs) under imperfect channel state information. To solve the complicated non-convex optimization problem, the original problem is decoupled into five subproblems, and S-Procedure is adopted to tackle the channels uncertainties. For sensing duration and transmitted power subproblems, their optimal solutions can be obtained through golden section search and quadratic transformation methods, respectively. For IRS’s PSMs subproblems, the suboptimal solutions can be obtained via sequential rank-one constraint relaxation and Cholesky decomposition. For non-convex fractional programming CUAV’s trajectory subproblem, the suboptimal solution can be obtained via successive convex approximation technique. Finally, an offline-online based framework is designed to optimize these variables alternate iteratively under the two-stage BST capability enhanced mode. Simulation results show that the proposed scheme outperforms the existing schemes significantly.