| The3-Dof helicopter is a complex objects with multi-variable and non-linear characteristics and so on. Usually its exact math model is very difficult toachieve. Through the research of controlling and simulating, it can improve theapplication of control theory and technology which working on the complicatedsystem. So the research on the theory and the designing of helicopter has veryimportant practical meaning.This article has mainly taken the3-Dof helicopter model which is providedby the Googol Company as the simulation platform. As a laboratory set-up, itsstructure is easy, its appearance is intuitionistic. It is a ideal platform which canbe used to test control theory. The work is listed as follows:1、This article briefly introduces the structure and the principle of work of3-Dof helicopter model system. The mathematical model of the system has beenestablished after we analyzed the movement characteristics and mechanicalcharacteristics of it. We analyze the stablility, controllability and observability,decoupling of system.2、 Firstly, we design status feedback controller by using the poleassignment method and then we can get the input and output data. This data isused as training data. Secondly, we design neural fuzzy controller. The training isperformed using neuro fuzzy inference system. After training, we apply thecontroller to3-Dof helicopter model system, and then complete the analysis ofsimulation.3、Neural fuzzy controller is applied to the actual model for real-timecontrol. We test the three axis respectively. When doing the test of the motion ofpitch, we compared the effect of neural fuzzy control with PID control. Theresults of experiment show that the overshoot amount of neural fuzzy control is smaller than PID control in the experiment of pitch motion. The neural fuzzycontroller which we designed can do well for both given tilt angle and givenrotation angle. Its response speed is fast, and adjustment time is shorter, it has agood dynamic characteristic. |
PDF Full Text Request