| With the rapid development of modern aviation industry, vehicle speed,maneuverability, ceiling have been greatly improved inboth civil and military fields.weare pursuit of higher, faster and better flexible performance,at the same time, it developedalong the direction of aerospace vehicle. In this context, as a aerocraft driver–humansdo not up to such a high intensity. So, UAV is created and appears to have an powerfuldevelopment.The essence of UAV is a kind of flying robot. Compared with the traditional "manned aircraft ", its characteristic and advantage is mainly manifested in: the UAV hasthe advantages of zero casualties, low operation cost, flexibility and strongadaptability.However, there is also a lot of problem on the research of UAV, such as longdevelopment cycle, high capital investment, consume huge manpower and so on.Therefore, the UAV simulation also has a very important significance, it not only cangreatly shorten the development cycle, save the cost, but also can be used to validate theUAV autopilot, decision-making, task planning, task coordination, feasible andcorrectable of strategy and tactics algorithm.Flight simulation of the UAV system mainly includes the air dynamics simulation ofmathematical model, the flight scene simulation, autonomous control and task planning.The main content of this paper on which the former two aspects of the related research:one is the mathematical modeling of the aerodynamic characteristics of UAV Based onRBF neural network; two is the simulation for visualization of UAV using OSG graphicsrendering engine.In view of the modern UAV aerodynamic layout and dynamic characteristics of moreand more complex, it is difficult to realistic effect that simulation is carried out using thissimplification, idealization of the former model. Therefore, this paper proposes a fittingand approximation of the aerodynamic modeling method for flight data based on RBFneural network. This method can make full use of flight data during the flight record, onaerodynamic characteristics of modeling, not only can reflect the modern aircraft hascomplicated aerodynamic aerodynamic characteristics, but also for single frameindividual targeted, it can have a definite object in view.In visual simulation, this study uses OSG rendering engine as a development tool,which is an open source, mature, widely used graphics repertoire. Using OSG todeveloping flight visual simulation system can develop flight visual simulationconveniently and agilelywith high-efficiency. The paper mainly studies the constructionof large-scale flight terrain, the closed loop communication between visual system andMatlab as well as the parameter gathering of such hardware as pole and helm by VPB.Finally, the results of joint debugging of aerodynamic model and visual system show that: the UAV flight simulation system, which was established based on the methoddescribed in this paper, meet the requirements ofthe visual system and hardwareparameters acquisition and so on. The flight simulation system is high fidelity and strongtargeted, could be used for UAV research or personnel training or other related fields. |
