Research On Small UAV Flight Dynamic Model Identification And Its Robust Control

Small UAVs are lightweight, flexible, low cost, safe and reliable. They have broad applicationprospects in the field of military and civilian. In this paper, UAV aerosonde is choosed to study theflight dynamic model. Firstly the input-output data is obtained by simulation stimulus, then by usingthem to estimate model parameters through identification methods. Robust control law for the UAV’slateral model is designed in this paper. The main works are as follows:(1)The subspace algorithm and prediction error method are used to identify the small unmannedaircraft vehicle’s flight dynamic model in level flight state.Two-step identification method is used toget the UAV dynamic model in level flight state. Firstly the subspace algorithm is used to obtaininitial optimization model, and then use the prediction error method to obtain the optimal modelwith priori knowledge provided by dynamic model. This study focuses on the theoretical analysis andpriori knowledge, so the result can correctly describe the mechanical properties andhas well-defined physical meaning and applicability, it is also easier accepted by engineers andtechnicians.(2)Matlab GUI is used to establish a set of identification software according to the selectionof identification algorithm. The software includes five modules: UAV identification data input module,data processing module, identification module, result analysis module and validation module, thefeatures of each module are designed with recognition of the need for flight test data.(3)Mu synthesis approach is used to design the small UAV lateral robust controller. The robustcontroller which is built using mu synthesis has good robust performance, and it can also avoida large conservative in design result. The order of robust controller is reduced as the application inengineering. Taking into account the UAV appears apparent changes of height caused by winddisturbance, fuzzy controller is designed to maintain the height of UAV, and the model is simulated.(4)Robust control law is compiled to source code which can be executed by DSP chip basedon Matlab\simulink platform. A lot of time for DSP chip programming don’t need by using thismethod. It can generate a complete DSP source code only relying on Matlab\simulink platform. Thesource code generated is verified to show its effectiveness.