Adaptive system of mutual positioning for controlling the groups of unmanned aerial vehicles

In this paper we consider the method of mutual positioning in the control system of groups of unmanned aerial vehicles (UAVs). For this purpose the exchange of coordinates is provided in the system. Each UAV has a Global Navigation Satellite System (GNSS) receiver and its own radar station with the active response. The integrated cyclogram of time in the presence of external synchronization by signals of GNSS is implemented at all interacting stations. The information exchange is organized as a fully connected system with time-division multiplexing. The structure of the transmitted data is determined. We offer a cluster system due to the rapidly increasing amount of signals in the fully connected system. It means that a group of UAVs is divided into subgroups. Each subgroup has a leader. The information exchange between leaders of subgroups is also provided. We consider the method of coordinates estimation based on the integration of measurement results obtained by using GNSS and the measurements from radars of UAVs. Mutual coordinate determination in the system has several stages: reception of the information from GNSS and airborne radars; secondary processing of GNSS data and the airborne radar information by means of Kalman filters; getting the resulting estimate by the integration of results after secondary information processing. It is researched how the choice of initial conditions and parameters of filters influences the quality of resulting estimation. The influence of update time in measuring systems with the active response on the accuracy of estimation by using Kalman filter is examined. The research is conducted by mathematical model method. The comparative analysis of two algorithms of integration is done. The relationship between computational cost of the algorithm of integration and accuracy of the resulting estimation is determined. The influence of quality of initial estimations on the accuracy of the Kalman filter resulting estimation is analyzed. At the same time weight coefficients are determined on the basis of the estimation accuracy of coordinate and ranging information filters. It is shown that information loss caused by fading influences the resulting errors of measurements. For these conditions we also consider the case of the movement of UAVs with big acceleration. It is concluded that the structure and parameters of the system should be adapted to the changing group members and conditions of obtaining coordinate information.