A Dispersed Federated Learning Framework for 6G-Enabled Autonomous Driving Cars

Sixth-Generation (6G)-based Internet of Everything applications (e.g. autonomous driving cars) have witnessed a remarkable interest. Autonomous cars using federated learning (FL) has the ability to enable different smart services. Although FL implements distributed machine model training without requirement move data devices centralized server, it its own implementation challenges such as robustness, server security, communication resources constraints, and privacy leakage due capability malicious aggregation infer sensitive information end-devices. To address aforementioned limitations, dispersed (DFL) framework for is proposed offer robust, resource-efficient, privacy-aware learning. A mixed-integer non-linear programming (MINLP) optimization problem formulated jointly minimize loss in accuracy packet errors transmission latency. Due NP-hard non-convex nature MINLP problem, we propose Block Successive Upper-bound Minimization (BSUM) based solution. Furthermore, performance comparison scheme with three baseline schemes been carried out. Extensive numerical results are provided show validity BSUM-based scheme.