Experimental verification of formation control by model predictive control considering collision avoidance in three dimensional space with quadcopters
Kenta Yamamoto, Kazuma Sekiguchi, Kenichiro Nonaka · 2017
This paper proposes a formation control method for multi quadcopters considering collision avoidance using model predictive control (MPC) in three dimensional space. To achieve the formation control, the term of consensus is included in MPC. Collision avoidance among control objects is considered by adding the term of repulsive force to index function of MPC. Thus this paper achieves formation control by combining consensus based control with collision avoidance control in MPC. In general, the repulsive force is described as a nonlinear function. It is difficult to add the nonlinear function to the term of index function because the computation time can increase. Hence it's not appropriate to implement to practical system. Thus the term of repulsive force is approximated as a linear function to achieve real-time control. In addition, the model of control object is assumed as linear system. The validity of proposed method is verified via experiment with three quadcopters in three dimensional space.