Frequency Identification Of UAV Based On CIFER

Rotorcraft can take off and land vertically, and can perform flight ranging from hovering to airplane-like cruising with agility and maneuverability. These qualities have made them indispensable vehicles for a variety of applications. With the development of microelectronic techniques and control systems, there is a growing interest in using unmanned small-scale rotorcrafts, which have extended to some civil use, such as urban traffic monitoring, spraying of agriculture and forestry, complex terrain mapping. It is the ever increasing applications that create the project.However, current unmanned small-scale rotorcrafts fail to exploit the vehicles'full potential because of deficient flight control systems. The design of high performance control systems for a vehicle with complex dynamics requires a mathematical model that accurately describes the vehicles'dynamics. And hence, in the thesis, we try to get linear model and obtain some physical parameters based on the available nonlinear model.The thesis has introduced the algorithm of CIFER to detail. CIFER is a frequency identification method based on the aircraft sweep frequency data. It is an excellent frequency identification method, which lies in the comprehensive identification process. It includes several steps: preprocessing, Fourier transformation, multi-input process, multi-widows process, state-space identification, time domain verification. The most advantage of this identification method is that it will provide a multi-input multi-output result. To a coupled unmanned small-scale rotorcraft, it can offer a more accurate model.After processing sweep frequency data of Raptor90's each channel by CIFER, I got accurate relations of each input and output. From the coherence of each input and output, we knew lateral channel is coupled with longitudinal channel and yaw channel at the same time. First of all, identify the SISO transfer functions of each channel. From the verification result, we can see these functions can basically describe the channel characteristics. However, the more important task of these SISO functions is to help set initial values of MIMO state space functions. The key part of the thesis is to identify and verify the state space functions of lateral-longitudinal, lateral-yaw coupled channels. Linearizing the normally used nonlinear model of small-scale rotorcraft, we can get the linear functions with physical parameters, then simply it to the same forms as identified function; get some physical parameters of nonlinear model, which is helpful for us to understand the physical structure of small-scale rotorcraft.