Research On The Anti-interference MIMU/GPS Vehicle Integrated Navigation Algorithm Based On IDNN-EKF

Along with the continuous development of science and technology,inertial navigation is no longer the unique noun of military weapons,in civilian UAV,intelligent mobile communications equipment,navigation equipment and other fields is also widely used.MEMS with its small size light,weight and low power consumption,to further promote the development of the micro inertial navigation system,making the use of inertial navigation is more common.MIMU is a kind of MEMS inertial components based autonomous inertial navigation system,it has the advantages of low cost,can be mass production,portable and so on.The vehicle integrated navigation systems based on the MIMU has been the focus of research at domestic and overseas because of its application universality.However,due to the MIMU is conditioned by the level of production technology,the inertial sensor precision is low,and the error accumulate rapidly,lead to low precision and poor stability of navigation results.So,MIMU can not complete the navigation mission independently,we usually using GPS for compensation,to form the vehicle MIMU/GPS integrated navigation system.However,due to the GPS signal is influenced by external factors,the phenomenon of GPS signal become abnormal during the vehicle navigation is frequent,it affect the navigation accuracy seriously.Therefore,how to overcome the effects of GPS signal abnormalities,effectively improve the precision and stability of the vehicle integrated navigation,is the main researching content of the recent years in the field of vehicle integrated navigation.Based on the introduction of present situation of the development and research,and the basic principle of vehicle MIMU/GPS integrated navigation system,and the classic information fusion algorithm,on the basis of information fusion algorithm for complex environment caused by GPS signal abnormalities,we focus on the investigation of MIMU/GPS information fusion and present an anti-interference MIMU/GPS vehicle integrated navigation algorithm based on IDNN-EKF.On the basis of EKF to fusion the navigation information,we utilize IDNN to aided navigation system,and according to the driving features of the vehicle build constraint equations to improve the stability and accuracy of the input delay neural network in the prediction stage during the GPS outages,and thus reduce the vehicle integrated navigation system of navigation error,improve the overall accuracy and stability of the system.When the GPS signal is good,using EKF to MIMU and GPS information fusion and filtering,to correct the navigation information,and at the same time to assist input delay neural network to carry on the training study,determine the parameter,establish the model.When GPS signal is detected abnormal,the input delay neural network will be entered into the phase of prediction,and instead of EKF to forecast the velocity and position errors,to correct the navigation information.To detect whether the GPS signal is normal,so that to switch working state in time.In this paper,we use the?~2 detection.In addition,to solve the problem that the input delay neural network stability and accuracy degradation,which because of the GPS signals caused by abnormal continued for a long time,we according to the movement of vehicles in the navigation feature to establish constraint equation for input delay neural network constraints.Finally,on the basis of the simulation experiments,we use XSENS MTi-G micro inertial sensor and GPS receiver in the vehicle test.Experimental results show that compared with the traditional vehicle integrated navigation algorithm the anti-interference MIMU/GPS vehicle integrated navigation algorithm can effectively overcome the GPS signal abnormalities,and improve the overall accuracy and stability of the navigation system.