Robust Beamforming and Power Allocation for Multi-Antenna C-V2X Networks

In this paper, we study the beamforming and power allocation problem in the scenario of multi-antenna C-V2X networks with uncertain channel state information (CSI). We attempt to minimize the beamforming vector of base station and the transmit powers of vehicle users, whilst ensuring probabilistic form constraints of quality of service (QoS). Then, a support vector clustering (SVC)-based uncertain set learning approach is developed to model the uncertain CSI into solvable convex set shapes. Based on the modeled convex set, a robust counterpart transformation approach is proposed to transform the highly intractable probability constraints into solvable polyhedral constraints. Then, a generalized equivalent transformation approach is developed to transform the probability constraint problem into a semi-definite resource management problem. To improve the solving efficiency of the resource management problem, we further propose an Enhanced SVC (ESVC) algorithm to simplify the spatial structure of the polyhedral constraints. Finally, the experiments are simulated to compare our proposed algorithms with other traditional and non-robust algorithms.