علي ابراهيمي

In this thesis, a new method for flocking quadcopters is introduced. For this purpose, the idea of two-level control has been used, in which the high-level controller is the same as the flocking algorithm and acts as the optimal path designer of quadcopters. Moreover, tracking of the generated desired path is performed by a low-level controller. In this thesis, the main focus is on the high-level controller. Therefore, a new leaderless flocking algorithm is introduced, where new potential functions are generated using fuzzy logic to achieve proper lattice and maintain the desired distance from the obstacles. The introduced potential functions have a minimum value in the lattice positions. Therefore, the control signal tries to minimize its value by using the gradient-descent method to reach the optimal desired path. In this thesis, a safety radius is defined for each agent such that with the proposed flocking algorithm, the quadcopters do not enter each other's safety region. In order to avoid obstacles, the idea of virtual agents, called beta agents, is used on the border of the obstacles; and to follow the desired path of the group, virtual gamma agent is used. The stability and convergence of structural and transitional dynamics of the system in the proposed method are investigated with and without obstacles. Finally, the proposed method is eva‎luated through simulations of five quadcopters under different conditions. The results show that The proposed method provides better performance in creating a lattice and maintaining obstacle distance and also greater resistance to external disturbances with and without obstacles.

،حركت گروهي، , سيستم هاي چندعاملي , ربات‌هاي پرنده , سيستم فازي , پرهيز از مانع

Flocking , multi agent systems , quadcopter , fuzzy , Obstacle avoidance

Ali Ebrahimi

Mohammad Farrokhi