Decentralized Multi-UAV Formation Control and Navigation over a Self-Organizing Coordination Network

Autonomous control technology for multiple network-connected multi-rotors and simultaneous remote control by few operators has applications in disaster investigation. Challenges in constructing such a system include path-planning for multiple multi-rotors. Accurate position estimation methods such as GNSS already exist for outdoor self-positioning; however, no optimal method has been established for the path-planning problem involving multi-rotors in dynamic and complex environments. Herein, in addition to the conventional artificial potential field method, we propose an extended function that introduces a rotational force field, which enables autonomous navigation from a starting point to the target while avoiding obstacles and local minima. Additionally, we realized autonomous decentralized formation control and navigation with multiple multi-rotors. This was achieved by applying a virtual spring-damper model between the multi-rotors based on the relative distance information calculated from global positions obtained by inter multi-rotor communication. To confirm the practicality of our proposed method, we generated a two-dimensional map from the three-dimensional model data of a real urban area and conducted simulations for various multiple multi-rotor applications, successfully demonstrating formation control and navigation in complex environments.