1: \begin{abstract}

2: 	To enable communication-efficient federated learning (FL), this paper studies an unmanned aerial vehicle (UAV)-enabled FL system, where the UAV coordinates distributed ground devices for a shared model training.

3: Specifically, by exploiting the UAV's high altitude and mobility, the UAV can proactively establish short-distance line-of-sight links with devices and prevent any device from being a communication straggler.

4: Thus, the model aggregation process can be accelerated while the cumulative model loss caused by device scheduling can be reduced, resulting in a decreased completion time.

5: We first present the convergence analysis of FL without the assumption of convexity, demonstrating the effect of device scheduling on the global gradients.

6: Based on the derived convergence bound, we further formulate the completion time minimization problem by jointly optimizing device scheduling, UAV trajectory, and time allocation. 

7: This problem explicitly incorporates the devices' energy budgets, dynamic channel conditions, and convergence accuracy of FL constraints.

8: Despite the non-convexity of the formulated problem, we exploit its structure to decompose it into two sub-problems and further derive the optimal solutions via the Lagrange dual ascent method.

9: Simulation results show that the proposed design significantly improves the tradeoff between completion time and prediction accuracy in practical FL settings compared to existing benchmarks.

10: \end{abstract}

11: