Research On Integrated Navigation Key Techniques In Complex Environment

Owning to the rapid development of communication technology and computer technology,a navigation system of higher precision,higher anti-interference performance and higher reliability in complex environment is required urgently.Satellite navigation system is the most widely used navigation and positioning means at present.However,its performance deteriorates sharply in complex environments such as canyons,tunnels,tall buildings and closed forests.All kinds of single navigation system,like inertial navigation system(INS),celestial navigation system(CNS)and barometer,have their limitations and are difficult in accurate navigation and positioning in complex environment.As a result,intergrated navigation technology,which combines multiple navigation sensors with optimized configuration and complementary performance,has become a research hotpot for now.The strapdown inertial navigation system(SINS)is taken as a reference to research the multi-sensor integrated navigation technology in the thesis.While the multi-source data fusion and the filtering algorithm are studied,the fault detection algorithm and system reconstruction algorithm are integrated to develop a multi-sensor integrated navigation system of unmanned aerial vehicle based on the federated architecture.The main research content includes:(1)To solve the problem of low performance of single sensor and GNSS/SINS integrated navigation in complex environment,an integrated navigation model of federated architecture is proposed.The strapdown inertial navigation output is corrected by using position information provided by satellite navigation system,doppler velocity information and magnetometer angle information.In addition,an improved adaptive filtering algorithm is proposed to improve the filtering accuracy and a simplified information fusion algorithm is proposed to reduce the computation and the effect of singular value.The proposed model can effectively reduce the impact of single sensor performance degradation in complex environment and get more accurate and stable resuts.(2)To solve the problem that whether different information allocation algorithms have different influence on the performance of integrated navigation system in this thesis,a variety of information allocation algorithms are proposed after verifying that the fault toterance performance of the distributed filter is better than that of the centralized filter.We study the accuracy and fault tolerance performance of various algorithms in fault-free condition and in fault condition to choose the most suitable information alloction algorithm for the model proposed above.(3)To solve the problem that the fault of the sub-filter in integrated navigation system affects the global accuracy in complex environment,a real-time fault diagnosis mechanism based on SVM is proposed by introducing machine learning algorithm and combining with federated architecture.The fault subsystem is isolated,then the whole system would be rebuild to make sure that the rest of the subsystems work properly and stably when one subsystem fails.At the same time,the machanism of information fusion and information allocation using fault score is studied to increase the filtering accuracy and further improve fault tolerance performance of the system.(4)A robust integrated navigation mechanism based on GNSS/SINS/DVL/CNS is established,which provides a feasible scheme for accurate and stable navigation in complex environment,by studying the filtering algorithm,fusion algorithm and the fault detection algorithm of the multi-sensor integrated navigation model.The performance of the mechanism is verified by experimental simulation.