Research on Optical Flow Assisted MEMS Navigation Method based on ROF Denoising in UAVs
Ling Zhang, Zhi Xiong, JIanye Liu, Pin Lv, Jianxin Xu, NUAA University, China
Alternate Number 4
UAV is caracterized by low cost, small volume, light weight, convenient operation, and strong flexibility. It can perform specific tasks such as reconnaissance and surveillance tasks in complex and low-altitude flight environment. However, UAV may encounter drastic change of state information when disturbed by obstructions, environment changes and other interference in complex environment. So, it is needed to obtain accurate and reliable real-time aircraft motion state information, to ensure the independence and reliability of aircraft, and meet the independent and reliable needs of UAV flight in complex space-time environment.
Recently, the most navigation system applied in small aircraft like inertial / satellite integrated navigation system based on MEMS inertial devices has deficiency, and its accuracy is not high while the error accumulates over time. Moreover, as auxiliary system of MEMS inertial navigation system, GPS and Beidou satellite navigation systems are vulnerable to multipath interference and result in error mutation or failure in dead zone.
Optical flow aided MEMS navigation based on ROF denoising method is carried out, the light-change environment is set and image structure-texture is analyzed to be decomposed. According to TV image denoising model by ROF, the discrete optical flow model is built, and the textured image part is obtained, which is not affected by illumination changes. Then, the texture part is analyzed to calculate motion parameters, through which the calculated attitude information is fused with MEMS navigation information, to work out more accurate navigation attitude information, so as to improve the navigation performance of UAV, and ensure its independence and reliability. The test results show that the designed method can effectively solve the problem of illumination changes on the optical flow information, ensure the reliability of optical motion information, and improve aircraft navigation performance.
The Navigation Research Center (NRC) of Nanjing University of Aeronautics and Astronautics (http://nrc.nuaa.edu.cn) is engaged in the research of navigation technology, and has done a series of research on all source information fusion. Up to now, a great deal of achievement has been made in the following fields: Strap-down Inertial Navigation System, Inertial Integrated Navigation System, Global Satellite Positioning System, Terrain Aided Navigation System, Fault-Tolerant Technique of Multi Inertial Navigation Subsystem, Modern Optimization Filtering Theory, Precision Measurement System and so on.