Autonomous Formation Control Of Multiple Mobile Robots

Multiple mobile robots system can perform some intelligent behavior which the individual is unable to achieve with cooperation. Therefore it has received widespread attention ranging from manufacture, daily life, military to aerospace exploration. For specific and complex missions, such as transportation of large objects, exploration in unknown environment and map construction, target capture, etc., formation behavior can improve efficiency, reduce risks due to individual failure. Each member of a team can acquire more comprehensive environment information, so that resources can be allocated and applied reasonably. So the research on formation control of multiple mobile robot systems is very important.Communication connectivity control of formation in plane and formation control in three dimensional space are investigated in this dissertation. The main research results are listed as following:(1) Communication connectivity control is investigated for multiple mobile car-like robot networks. Employing algebraic graph theory, La Salle invariant and negative gradient algorithm,the distributed control laws based on distances among robots are proposed to maintain the communication connectivity of mobile car-like robots. With the control laws, robots can achieve the maximum connectivity topology and avoid collision with each other.(2) Regular polyhedron autonomous formation control in three dimensions is considered for multiple fixed-wing UAVs. Based on virtual leader and simplified agent model formulated as first-order integrator, the feedback control law is developed for distance based formation control.Velocity input for each fixed-wing UAV can be obtained by coordinate transformation.(3) Distributed motion planning is investigated for a regular polyhedron autonomous formation control of a group of miniature quadrotor vehicles. The kinematics of miniature quadrotor vehicles is simplified as second-order integrator in a framework of closed-loop cascade control system. Then the backstepping method is applied to generate the desired position,velocity, yaw angle and angular velocity trajectory for each vehicle. The outputs of the formation controller are taken as the reference inputs of the inner loop so as to implement autonomous formation control of multiple quadrotor vehicles.