Multi-Drone Path Planning Using Gradient Descent and Potential Fields For Warehouse Applications

This study presents an approach for multi-drone path planning in warehouse environments using a combination of the Gradient Descent Method and Artificial Potential Fields (APF). The objective is to ensure efficient and collision-free navigation of multiple drones through complex warehouse spaces, avoiding obstacles such as storage racks. The Gradient Descent method is employed to guide the drones toward their target locations, while the APF technique integrates attractive forces toward the goal and repulsive forces from obstacles to dynamically adjust the drones’ paths. Simulations were conducted using MATLAB to evaluate key parameters such as path length and navigational time, with variations in obstacle configurations. To validate the approach, real-time experimental trials were performed using custom-built drones equipped with Pixhawk flight controllers and obstacle detection sensors. Results from both simulation and experiments demonstrated high accuracy, with an error of less than 5% in path length and navigational time. The findings show that this combined method is effective for autonomous drone coordination in real-world warehouse applications, ensuring safe, efficient, and adaptive navigation.