| Being in small size, low speed and limited payload capacity, most conventional flight navigation systems and strategies are no longer practical for Micro Aerial Vehicles (MAVs).The miniature equipments are required in complex indoor environments emergency rescue, exhibition tour, navigation for the blind or visually impaired where Global Positioning Systems (GPS)can't work. The MEMS sensors have been becoming increasingly important in the aerospace, communication, navigation, transportation and safety areas because of their miniaturization, high integration, low power consumption,and low cost features. The study of the measurement and control technologies in navigation applications with a micro modular navigation system based on MEMS sensors has a great scientific and practical significance.This paper focuses on the key technologies of attitude measurement,navigation control and localization based on the navigation platform. Attitude using quaternion method under dynamic conditions was calculated with the inertial/magnetic sensors .The motion acceleration was eliminated to improve the calculation accuracy. With reliable attitude information, a practical navigation algorithm was proposed to control the MAV to achieve autonomous navigation and complete desired flight tasks. The three dimensional indoor pedestrian navigation was also discussed to achieve localization where GPS signal is not available. A universal and modular micro navigation system was proposed with MEMS sensors. The multi-problems of weight, size, power consumption, computing speed, anti-interference,versatility and scalability were considered in system establishment. A highly integrated navigation platform was designed with six functional modules under the integration and modularization principles. The system scalability was improved to set aside several interfaces with the support of the host controller computer. MEMS sensors errors were specifically compensated.The experiments of attitude determination, flight navigation and pedestrian localization were carried out to based on the navigation system.On the basis of summarizing the main work, the innovations of this paper are as follows:(1) A dynamic attitude determination algorithm was achieved based on MEMS inertial/magnetic sensors. The quaternion fuzzy adaptive Kalman filter algorithm was used to calculate the dynamic attitude. The measurement noise matrix under dynamic conditions was adjusted to reduce the impact of dynamic motion acceleration. The corresponding dynamic processes for MAV navigation and pedestrian navigation were analyzed in details. The MAV motion acceleration was estimated by airspeedmeter, and the centrifugal acceleration was analyzed with the force balance method. The expert rules were designed to adjust acceleration errors observation matrix in severe dynamic scenarios, the impact of measurement error was weakened to augment the attitude measurement. This attitude algorithm can overcome the ordinary Kalman filter attitude algorithm divergence problem in big posture change.(2) A practical control structure and navigation algorithm were proposed and improved with the navigation system, and applied to the servey of Jingjiang waterways. The decoupling of heading and roll directions was used in stabilization control to improve the security. An improved nonlinear guidance method was described for horizontal navigation control. The nonlinear controller can automatically adapt to ground velocity change, which is usually caused by gust disturbance, thus the UAV has good wind resistance characteristics. Autonomous navigation and specified flight tests were achieved with the proposed navigation method. The deviation of the trajectory in the survey of Jingjiang waterways with the SKY09P system was below 20m and the altitude deviation was below 10m, which satisfied the trajectory tracking demand.(3) A modified IEZ+ PDR (INS-EFK-ZUPT+ Pedestrian Dead Reckoning) localization algorithm was proposed with the navigation platform to realize the indoor three-dimensional navigation. With sensor errors completely compensated, the displacements could be achieved with acceleration secondary integration, and the attitude information was acquired with acceleration, magnetic field and angular rate fusion. The stance phase was detected by the "AND" logic with acceleration and angular velocity to improve detection reliability, synergism height was the combination of pressure altitude and acceleration height. The displacements were achieved to obtain a relative error about 1% of the horizontal traveled distance and below 0.1 meter on the vertical plane. |
PDF Full Text Request