User Selection and Antenna Optimization in Intelligent Flexible Position Antenna-Assisted Privacy-Preserved Vehicular Federated Networks

Privacy preservation and communication efficiency are two critical issues in making intelligent transportation systems (ITSs) come true. Owning to the recent advances on differential privacy (DP) and intelligent flexible position antenna (IFPA) technologies, it is promising that the aforementioned critical issues can be relieved by introducing DP and IFPA into vehicular federated networks. Still, it faces many challenges to implement such vehicular networks. On the one hand, imposing further protections on vehicular federated networks leads to extra energy consumption of vehicular users, and they may be unwilling to employ DP technology. On the other hand, vehicular user's mobility leads to continuous changes to the network topology, and therefore IFPA optimization should adapt to the variations of network topology. In this paper to overcome those challenges, an integrated compensation minimization and IFPA optimization algorithm is first proposed. The convergence upper bound of this algorithm is analyzed and some facts are revealed that there exists a good value of FL round which can strike a balance between the learning performance and training time, and also more participant vehicles will further minimize the convergence upper bound. Then, a user selection problem is formulated, subject to a limited compensation budget. A truncated scheme is proposed to optimize the vehicular user set. Finally, an integrated user selection and IFPA optimization algorithm is proposed to holistically address the implementation of IFPA-assisted privacy-preserved vehicular federated networks. Extensive simulations are conducted to demonstrate the effectiveness of the proposed algorithms. The simulation results show that the proposed algorithm can achieve significant advantages in minimizing the compensation cost and communication energy consumption, and the proposed integrated user selection and IFPA optimization algorithm can run smoothly and converge while the network topology is continuous changing.