Research On Quad-rotor Integrted Navigation Based On Low Cost Sensor

In the rapid commercialization process of the multi-rotor unmanned aircraft,a product with high stability,low cost,superior performance plays a vital role in the occupation of the market.This fact put forward harsh requirement for the UAV's navigation system.How to integrate these navigation sensors' data well to build a stable navigation system which can tolerate faults effectively is a significant part during the research.It possesses unparalleled commercial value and practical significance.Information for navigation mainly comes from micro inertial sensors,GPS,barometer and magnetometer.All of them have its corresponded output model and characteristics of error.For the inertial sensor,we use Allan variance analysis to analyze the random error and calibrate it according to the corresponding output model.For the special status of inertial navigation in the integrated navigation system,the information update algorithm and error transfer characteristic of strapdown inertial navigation system are deduced in detail.In relation to this fact,this paper simplifies the update algorithm and error transfer characteristic according to the sensors' characteristics and application scenario.Based on the above subsystem models,the attitude information fusion model and the velocity/location information fusion model are constructed respectively.The attitude information fusion model uses the information of accelerometer and magnetometer to compensate the attitude information integrated by gyroscope.The second fusion model uses GPS and barometer to calibrate the inertial navigation output.At the same time,we use centralized filter and federal filter to integrate velocity/location information respectively and compare the features of them especially the advantage of federal filter in relation to the ability to tolerate the fault of subsystem.Ultimately,in order to verify the effectiveness of the proposed integrated navigation system and the advantage of the federal filter related to fault-tolerant performance.We set the trajectory and calculate the output of navigation subsystem inversely and introduce the corresponding error.Finally we can complete the simulation for verification.