Research On INS/GNSS/Visual Integrated Navigation Technology For Small UAVs

Navigation system is an important part of unmanned aerial vehicles(UAVs),which plays a key role in the flight platform.Aiming at the navigation and positioning requirements from small UAVs in complex environments,this thesis studies the deeply integrated navigation technique using the middle/low-accuracy INS(Inertial Navigation System)/GNSS(Global Navigation Satellite System).By establishing nonlinear error models for the INS/GNSS integrated navigation system,integrated navigation filtering methods like nonlinear filtering and noise adaptive filtering are researched to improve the performance of the inertial aided satellite deeply integrated navigation system.Aiming at practical mission requirements of UAVs,visual tracking-based integrated navigation technique is further researched.By introducing a visual auxiliary system,both environment modeling and dynamical target tracking under unknown environments are realized,which further improve the engineering performance of UAVs.The main works and innovative contributions in this thesis are summarized as follows.1.Aiming at the situation of large attitude angle errors,the velocity,position and attitude error models of the nonlinear strapdown inertial navigation system are established using the quaternion method.The optimization problem of the strapdown attitude algorithm is analyzed,and the third-order Taylor recursion expansion for solving the corresponding quaternion differential equation is derived.Comparing studies between the derived method and the previous forth-order Rouge-Kuta method are also provided.2.Aiming at effects of the uncertain integrated navigation filtering model and noise,the noise adaptive filtering algorithm based on the maximum likelihood estimation is developed,which enhances the stability and performance of the integrated navigation system in the situation of unknown system model and noise characteristics.3.Aiming at position and velocity measurement errors as well as the inefficient computing capacity of GNSS,a position/velocity sequential UKF filtering method is proposed,which has been successfully applied in the initial alignment and navigation calculating of small UAVs,especially for the large misalignment angle condition.4.Aiming at application needs of the integrated navigation under high dynamic and strong interference environments,inertial aided satellite deeply integrated naviga-tion technology is researched.The performance analysis for the inertial aided loop is realized based on the characteristics of satellite loop.An architecture of deeply integrated navigation based on INS is proposed,and further,the simulating analysis and system development are realized,indicating that the proposed method enhances the dynamic adaptation of satellite navigation as well as improves the disturbance rejection capability,significantly.5.Aiming at handling the failure of GPS under complex electromagnetic environment and thus the unsatisfactorily low precision of pure inertial navigation,a visual tracking-based integrated navigation algorithm is proposed.This navigation algorithm is based on an improved target color kernel-based tracking algorithm,which utilizes multidimension color features to describe the moving target as well as to remove the redundant information,realizing the accurate tracking to the target.The tracking information is utilized to obtain the velocity of UAVs,which is then sent to the inertial navigation system to realize integrated navigation.The proposed method can improve the navigation precision in the absence of GNSS.6.To verify the proposed integrated navigation methods including the sequential STF-AUKF-based INS/GNSS integrated navigation,the semi-coupled deeply INS/GNSS integrated navigation and the visual assistant integrated navigation,various onboard flight tests have been implemented.Experimental results indicate that the proposed integrated navigation methods can efficiently improve both robustness and precision of the navigation system of UAVs.