UAVs Group Control Based on the Relative State Space Method

Abstract

In this paper, the problem of fixed-wing unmanned aerial vehicles (UAVs) group control was solved using the relative state space method. To achieve this goal the authors created a flight dynamic model of a fixed-wing UAV. Then it was linearized using MATLAB built-in routine and the standard PID autopilot was synthesized. The control inputs for the UAV's inverse kinematic model were calculated based on relative state space method, which is decentralized control of the multi-agent system, fault-tolerant in the sense that the failure of individual agents does not lead either to the failure of the entire system or to the inability to further build and maintain the formation. Each agent has autonomy, i.e. the ability to control part of the system's global state. This method is a bio-inspired algorithm based on the model of living organisms' motor neurons network. In comparison with the "leader-follower" method, the relative state space approach involves the construction of a control hypersurface in the relative state space instead of just following "leader" commands. In addition, this method is resistant to atmospheric disturbances in comparison with virtual structure approach and it has a low computational complexity. Initially the relative state space method was developed only for linear control objects without taking into account their dynamics. Therefore, it was modified by the authors to be applicable to nonlinear control objects (for example, fixed-wing UAVs). The math modeling in MATLAB/Simulink shows successful solution of the problem. Further research will focus on optimization of the formation building time, considering the vehicles collision avoidance and creating path-following algorithms based on the above method.