Neural Network Control for Image Stabilization System on Helicopter in Remote Sensing

Image stabilization system (ISS) is a device used in remote sensing which acquires stable images of remotely located objects. To get stable images, the ISS should reject the outer disturbance. The ISS has nonlinear characteristics such as structural errors, friction and disturbances. Although PID controllers are optimally designed, PID controllers may not perform well as expected when disturbances are present. To reject disturbances, the neural network controller along with PID controllers is used to compensate for uncertainties in online fashion. Experimental results show that performance of stabilization is well improved by a neural network controller