Vision Based Flocking Control for Multi-Agent Systems in 3-D Space.

We consider the problem of designing distributed controllers for multiple mobile agents to achieve flocking motion in three-dimensional space based only on measured position. A flocking algorithm is proposed under which each agent has the important performance of cohesion maintenance and collision avoidance (CMCA) as basic control goals. In addition, velocity consensus will asymptotically attain. To achieve flocking motion, velocity information can be estimated by a velocity observer which only depends on position information, and using potential function which provides attractive/repulsive bounded force to ensure CMCA. The process of designed control law is presented by means of backstepping method. Simulation results emphasize the effectiveness of the proposed flocking control law.