UAV data delivery method based on mobile node trajectory prediction

Drone data delivery via mobile node trajectory prediction and cell-based detection probability aggregation

This technology enables UAVs to predict the future paths of multiple target nodes using movement patterns in a region of interest, determining flight paths to deliver data within a deadline. The region is divided into cells, and the UAV moves to the cell with the highest aggregate detection probability for all target nodes at each time t.

In MANET environments, constant node movement causes frequent network topology changes. This often leads to broken paths to destination nodes, resulting in data transmission failures.

This technology organizes historical node movement data into group cells—clusters of adjacent cells within communication range—to establish unique paths between origins and destinations. By clustering these, it selects the top K clusters most similar to a target node's past path. It then predicts future paths by comparing K-cluster trajectories across time slots to identify the most similar cells. When UAVs meet, they exchange data and detection probabilities. Applicable to military messaging, disaster relief, and maritime data collection, this allows drones to reach scattered terminals without infrastructure, increasing on-time data delivery rates.

‍

‍

Key Features:

• Predicting future paths of multiple target nodes based on movement patterns in a region of interest
• Determining the UAV flight path based on the aggregate cell-based detection probability for all target nodes in a region divided into multiple cells
• Summing detection probabilities for all target nodes in each cell at time t and selecting the cell with the highest value as the path for that time
• Delivering data to at least one target node within a specific deadline along the determined path

‍