| The paper aims at auto-leveling system's and auto-leveling technology's actuality and develop with brief introduction and discussion at first. Then the thesis introduces mathematic module for electromechanical auto-leveling system and proposes fixation height excess adjust correction leveling strategy, and make its rigidity model simulate in SIMULINK of MATLAB. Considering some factors, such as the transmogrification of auto-leveling flat, stroke clearance of lead screw legs and hyperstatic and so on, have grade effect in leveling precision, and it is difficult to found a precise mathematic module for this factors. So we found a neural network error compensated model makeuse of the ability-distinguishing an unknown function. Then the neural networkerror compensated forecasting model is founded with combining rigidity model with neural network error compensated model. We can get the conclusion that the control structure has high control precision and rapid convergence ability though simulating in MATLAB. In the aspect of weak leg question, we obtained the law of the weak leg caused by protraction changes respectively through the analyses in the statics of simplified system. Considering the different of control sequence and control law in the control process, divided the control system to two parts: weak leg judgment and protraction changes pre-estimate. So that debased the system's complexly. Built the WJFLC and the EQFLC control models constructed two step control system, and defined the inputs quantizing gene of WJFLC and the EQFLC. Simulation shows that this control construct can un-weak the weak leg quickly and won't cause strongly shake of the platform, when there is no senor besides leveling senor. Simultaneity the study's fruits applied in the radar-antenna truck. And program is used to control the vehicle's balance; the whole program is easily, humanization, intelligential.
|
PDF Full Text Request