A novel relative navigation algorithm for formation flight

Abstract

For higher accuracy and better performance, a novel relative navigation framework named full parallel distributed architecture is presented, using inertial navigation system, global positioning system, and data link. Aiming at multiple aircraft, it is designed to enable each plane to serve as a fusion center. Also this structure enhances the collaboration between aircraft by sharing the relative navigation information. It breaks the limitation of the single fusion center method and can adapt to the reconfiguration of formation. A two-stage filtering estimation algorithm based on Kalman filter is developed to determine relative position, attitude, and velocity between the formation aircraft. Each vehicle contains not only a local filter but also a relative state filter, which helps to improve the accuracy of the relative state error model. The relative navigation system is designed as a closed loop system with parallel processing and real-time performance. Simulation results compared with the traditional centralized filtering method indicate that the approach provides better estimates and restrains the error divergency effectively.