The Low-Cost Integrated Navigation System For Small Rotor UAV

Navigation technology is the foundation of flight control, and also the key point to the UAV’s safety. The rotor UAV navigation system is mainly based on the low-cost Inertial Navigation System (INS) and the Global Positioning System (GPS). The integrated system which is mentioned above can take advantages of both the subsystems, and improve the precision and reliability of the navigation information. However the GPS signal is easy to be interfered by obstacles block, or electromagnetic wave interference, as a result, it may cause outliers in the measurments, even make the GPS unlocked. So for the low-cost aircraft, how to improve the stability and the reliability of the navigation system when GPS is abnormal, has great theoretical importance and practice value.On the basis of previous work, with an eight-rotor UAV as an object, this paper focuses on the integrated navigation algorithm. The integrated navigation algorithm for the low cost system is proposed, and the situations when GPS signal is abnormal and unlocked are studied.The main work and contributions of this thesis are summarized as follows:(1) The inertial navigation system’s information (posture, speed, location) updating algorithm and error propagation model are studied, the system model of the INS/GPS integrated navigation system was setted up.(2) On the basis of the full state navigation system model, considering the low-cost computing unit of the rotor UAV, a hierarchical navigation structure based on the extended Kalman (EKF) is designed. The integrated navigation system is decomposed into two groups:attitude plus velocity group, and the single position group, aiming to reduce the order of the filter and finally improve the real-time performance.(3) Based on the position estimation system, a robust position estimation algorithm is proposed to solve the problem when GPS signals are abnormal. This method is proved that it can effectively identify the outliers, and improve the hover range into about 1 meter. What’s more this mothod can be appllied into different filters.(4) The theory of artificial neural network is studied and the RBF neural network and the Extreme learning machine(ELM) are introduced. The ELM algorithm is introduced into the integrated navigation system to solve the problem when GPS is unlocked. When GPS is locked, the neural networks learn the model through the navigation parameters according to the motion state of the UAV, when GPS is unlocked, neural network is appllied to predict the observations of Kalman filter. The simulation results show that ELM has a faster training speed compared with the RBFNN, and the errors accumulate slowly, so it can meet the requirement of the navigation system well and provide a strong guarantee for the flight safty when GPS is unlocked.