Predictive Beamforming for ISAC-NOMA Enabled Vehicle Networks

As one of the important technologies of sixth generation mobile communication, integrated sensing and communication (ISAC) enables the vehicle networks to perform data transmission and target perception simultaneously. Aiming at solving the system congestion problem caused by the large number of communication devices and severe inter-user interference in 6G networks, this paper investigates a predictive beamforming scheme based on the non-orthogonal multiple access (NOMA) aided ISAC based on the historical channel space-time network (HCSTN) vehicle networks. A beamforming design problem is formulated to maximize the achievable communication rate under the constraint of sensing performance in the vehicle networks. To solve this problem, a penalty factor method is proposed to transform the original constrained optimization problem into an equivalent unconstrained optimization problem, and then a deep neural network (DNN) model is designed to optimize the system predictive beamforming matrix to maximize the system achievable communication rate. Simulation experiments show that the proposed the algorithm significantly improve the system communication rate while satisfying the perceived performance constraints, and the proposed HCSTN model has good robustness compare to other system influence factors.