Research On Inertial Integrated Navigation Technology For Low-cost Mini Unmanned Aerial Vehicle

The mini unmanned aerial vehicle (UAV) has small size, competitive price, and the ability of vertical take-off and landing which helps the UAV to perform the hovering and ultra-low-altitude flight. By such advantages, the mini UAVs can be seen in many applications, such as scientific research, military, counter strike, agriculture and industry. But, by constraints from payload, battery capacity and budget, mini UAVs are usually equipped with the small and low power consumption integrated navigation platform which integrates the low cost MEMS inertial sensors and satellite navigation system. Traditionally, the mini UAVs are remote controlled by operator in line-of-sight. In this case, the satellite navigation system is the only information source of current location. But now, as the UAV market keeps growing, more and more applications are referred to auto-return and mission planning, and operated in different sophisticated environments which have more requirements on continuous and accurate position determination. Therefore, further research on integrated navigation technology should be made to improve the accuracy and robustness of the navigation solution for mini UAVs.This dissertation will focus on the key steps of developing a low cost multi-sensor platform for the mini UAVs and implementing the integrated navigation system on this platform, including the hardware and software design, sensor error analysis and modelling, sensor calibration, attitude estimation, GPS/INS integrated navigation algorithms and related measurement quality control strategies. Following is the detail of this work:1. A new low cost multisensor platform, which integrated digital signal processer (DSP), MEMS inertial sensors and GPS receiver, is developed by the author for the research work in this dissertation. The soft core on this platform is a real time operating system (RTOS) DSP/BIOS, which can improve the execution efficiency and real-time performance of advanced algorithms under high payload. Moreover, it is also a good solution for the time asynchronization issues in integrated navigation system.2. Error analysis and calibration are carried out to improve the accuracy of low cost MEMS sensors. The portable integrated navigation platform usually needs to be recalibrated in field after re-assembly or in a new environment. To simplify the recalibration work, an in-field auto-calibration method based on ellipsoid fitting for magnetometer and accelerometer is proposed. Moreover, a new measurement quality control strategy based on quartile and convex hull volume is also proposed. The auto-calibration method can obtain the similar calibration result as the instrumental method. But it is more convenient and faster, and has no requirement on specific equipments. The measurement quality control strategy can detect and reject the coarse errors in measurements, and it can also verify the rationality of spatial distribution of measured data sets for ellipsoid fitting. By that, the accuracy and reliability of calibration work and the usefulness of ellipsoid fitting are all improved.3. A quaternion based adaptive extended Kalman filter (AEKF) is proposed for the attitude estimation. In this AEKF, the zero-acceleration assumption is introduced to the state process model to improve the performance. And a gyro-assisted fault detection and rejection (FDR) strategy is also proposed to improve the robustness and efficiency of magnetic disturbance detection, which can remarkably decreases the false alarm rate and false dismissal rate.4. A loosely coupled GPS/INS integrated navigation scheme with external aids is proposed for the low cost integrated navigation platform. The integrated navigation solutions are improved by introducing in the external heading and self-differential barometer height.5. A tightly coupled GPS/INS integration scheme is proposed to help mini UAVs to survive in complex environments with poor GPS signals by using the raw measurements from GPS receiver. And a quality control strategy is also proposed to detect and correct the biased measurements in such environnments. The characteristic of GPS receiver clock is analysed, and a method for clock jumper detection and compensation is presented. By that, it also optimises the initial alignment performance of the tight coupling integration. To eliminate the multipath effect, a pseudorange smoothing method based on Doppler measurement is proposed. The weight number of smoothing algorithm is optimised by using the carrier to noise rate of each available satellite signal. At last, a multi-scenario trial is carried out to test the navigation performance.