Transfer of objects by robots is a main issue in the robotics nowdays. Ground robots are able to carry heavy loads, however in rugged and slippery environments, displacement of objects is difficult by ground robots. Therefore, the use of aerial robots can be suitable. In this research, the aim is carrying a certain weight load by four quadrotors. Quadrotors are located at four vertices of a square and the load is connected by rope to each of them. The equations of motion of the quadrotors are considered completely and without simplification. To express the relationship between the load and the quadrotors, the ropes are considered as springs, thus they are pulled out and retracted during the mission. Formation control of the group is done in centralized and decentralized methods. In the centralized method, the virtual structure is used and in the decentralized method, none of the quadrotors knows the information of the others and decides itself to carry out the assigned mission. When there are external forces (wind), the performance of both centralized and decentralized controller has also been studied. To avoid collision with environmental barriers for team of quadrotors is considered in both well known and unknown environments. To avoid of quadrotors collision with obstacles in the known environment, the optimum A * algorithm is used and an algorithm based on the use of distance-sensing sensors is designed in an unknown environment. An extended Kalman filter is used to estimate quadrotor’s speeds and use them in the obstacle avoidance algorithm. Cutting the rope of one of the quadrotors is applied to the system as a fault and the performance of the centralized and decentralized controller is checked in this situation. Also in this research external impact on the load is applied and in this case, the quadrotors must eliminate its effect by moving back and forth and reduce load fluctuations. Keywords : Quadrotor, Suspended load traort, Formation control, Centralized and decentralized control, Obstacle avoidance, Kalman filter, Fault occurrence.
