In the nature, it is common to observe flocking phenomena such as bird flocks and animal herds. With flocking behaviors, the individuals will follow the moving pattern of their group without colliding with the others. Many researchers have been attracted to the formation or flocking of multiple intelligent agents in order to imitate the behavior in a bird flock. Although there exist many realizations on the topic, it is still a challenge to implement a fully autonomous and distributed measurement and control system, which enables each agent to perform positioning and controlling with onboard sensors and processors instead of global positioning sensors and ground stations.

To realize an autonomous and distributed measurement and control system, some flocking control laws such as the “Vic...[ Read more ]

In the nature, it is common to observe flocking phenomena such as bird flocks and animal herds. With flocking behaviors, the individuals will follow the moving pattern of their group without colliding with the others. Many researchers have been attracted to the formation or flocking of multiple intelligent agents in order to imitate the behavior in a bird flock. Although there exist many realizations on the topic, it is still a challenge to implement a fully autonomous and distributed measurement and control system, which enables each agent to perform positioning and controlling with onboard sensors and processors instead of global positioning sensors and ground stations.

To realize an autonomous and distributed measurement and control system, some flocking control laws such as the “Vicsek” model and the “Cucker-Smale” model can offer real-time control commands to the agents of a flock rather than planned trajectories. As for the sensor selection in order to complete the task, the Ultra-WideBand (UWB) sensors can measure the distance and communicate concurrently. Besides, traditional sensors such as cameras and inertial measurement units (IMU) are also needed.

In this thesis, some previous flocking control models as well as positioning methods are reviewed first. Second, a visual-inertial-UWB based trilateration algorithm and a pure UWB based algorithm are proposed to deal with the relative localization problem. After obtaining the relative position and velocity estimation, a flocking model modified from the “Cucker-Smale” model and the “Cucker-Dong” model, is used for the flocking control. Finally, an unmanned aerial vehicle (UAV) and an unmanned ground vehicle (UGV) are implemented for the experiment. It is shown that the UGV can follow the movement of the UAV and the velocity converges, which indicates the feasibility of the proposed relative positioning algorithm as well as the modified flocking control model.