Multi-trajectory prediction of 5G network for smart grid based on Transformer

5G mobile communication technology has the advantages of ultra-large connections, ultra-low latency, ultra-high bandwidth etc. Smart grid is a typical 5G vertical application field, which includes various types of 5G mobile terminals, such as power inspection robots, mobile charging piles, drones, and electric work vehicles. The mobility of such terminals inevitably causes load fluctuation of 5G network. Predicting the trajectories of these mobile terminals can provide effective support for customizing or optimizing the management of mobility, registration and handover. Most existing methods for trajectory prediction consider the input trajectories as independent, but in smart grids, the trajectories of mobile terminals have great correlation. To solve this problem, this paper proposes a multi-trajectory prediction method for 5G mobile terminals of smart grid based on deep learning Transformer. It adopts a double-layer structure. The first layer Transformer encodes the historical and current predicted trajectory for each mobile terminal, and the second layer Transformer further captures the correlated movement patterns among the terminals to improve the prediction performance. We simulate the collaboration work mode and division work mode of 5G terminals of smart grid, and experimental results verify the effectiveness of the proposed method.