Research On Neural Network Sliding Mode Control Of Airborne Photoelectric Stabilized Sighting Platform

Airborne photoelectric stabilized sighting platform is an integrated equipment integrating optical technology,electronics technology,precision machinery technology,automatic control and communication technology,which is often equipped outside the carrier.It can effectively reduce or avoid the adverse impact of carrier vibration on the photoelectric sensor.Through the use of photoelectric sensors to scan the target,and the acquisition of image information and data for real-time feedback,so as to assist the flight crew quickly and accurately complete the search and detection of the target.Since the airborne photoelectric stabilized sighting platform system is mounted on the aircraft,the attitude change,vibration and the reaction force of the air on the frame during the flight will cause interference to the visual axis,and thus have a great impact on the clear imaging of the observation equipment.Considering the above factors,it is of great practical significance to improve the control precision of airborne photoelectric stabilized sighting platform system by researching advanced control strategy under the premise of saving cost.This paper mainly studies from the following aspects:(1)Taking the airborne photoelectric stabilized sighting platform with two axes and four frames as the research background,various factors affecting the stability of the airborne photoelectric stabilized sighting platform were studied through the mechanism analysis of the airborne photoelectric stabilized sighting platform,and the mathematical model of the airborne photoelectric stabilized sighting platform was established with the method of system identification.(2)For airborne photoelectric stabilized sighting platform velocity loop of rapidity and stability requirements,combining Sliding Mode Control,(Sliding Mode Control,SMC)plus interference to the system and dynamic modeling has good robustness and adaptability of completely,and the advantages of implementation method is simple,the Sliding Mode variable structure Control method is used to realize the speed Control loop.The simulation results show that the sliding mode variable structure control can realize fast tracking of the control input,but the speed tracking error is large,and the increase of the initial state error leads to the sharp increase of the control input.It can be seen from the velocity tracking curve and control input curve that the system tracking has obvious chattering problem.The existence of chattering problem will affect the overall stability of the system and even lead to the instability of the system.(3)By using RBF Neural Network(Radial Basis Function Neural Network,the RBFNN)as the compensator to improve the sliding mode control law,considering actuator is a kind of limited input of the Neural Network sliding mode variable structure control method.The simulation results show that this method can track the angular velocity of the airborne photoelectric stabilized aiming platform quickly,and the control input can quickly jump out of the saturation limit of the motor and become smooth,which effectively reduces the chattering problem in the system when the sliding mode variable structure control method is used alone,and improves the control effect.