| Inertially stabilized platform (ISP), also known as gyroscope stabilized platform (GSP), is the core device of navigation, guidance and measurement equipment. It can efficiently isolate carrier turbulence, maintain the platform’s stability, and allow the optical equipment to maintain line of sight (LOS) or track and target the selected object. High precision inertially stabilized platform is a pivotal part of modern warplanes, hence developing stabilized platform with high precision, super long-range identification ability and real-time image production, has been the heated topic in the worldwide arm race. Assuming that the hardware used in the platform is unchangeable, this thesis takes a closer look at how to enhance the platform’s stability accuracy by increasing the isolation ability through improvement done on the control strategy.Within the inertially stabilized platform’s multi-gimbal control system, the velocity-stabilized loop that holds the gyro, is the core component that isolates the carrier turbulence. In applied engineering, linear controllers such as PID controller, lead-lag controller etc have sophisticated design and pronounced efficiency, hence are widely seen in velocity-stabilized loop. This type of controllers usually has good precision, however it is difficult to further improve its performance. Mainly because (1) an inertially stabilized platform normally works under low frequency low speed environment, but linear controller’s ability to prevent non-linear friction is limited, causing friction to become the main cause of error in velocity-stabilized loop. (2) During the design of traditional controllers, which rely on the linear model of the controlled object that is identified under frequency domain, the non-model parts and the effects that are brought upon during parameter change are not considered. Based on above, the main research objection of this thesis is to apply appropriate control strategies onto the original control system to overcome the errors on the stabilized platform and further improve the velocity-stabilized loop’s isolation performance.The thesis is based on the joint project (NO. KD1012210167) between University of Science and Technology of China and Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Science. The research target is military used two-axis four-gimbal airborne inertially stabilized platform with micro radian. The main objective is to enhance the inertially stabilized platform’s velocity-stabilized loop’s isolation performance, through research such as modelling analysis and controller design. The main researches of this thesis include the following work:(1) The design of inertially stabilized platform’s velocity-stabilized loop’s high precision model in order to build a solid base for design of the control strategies conducted later on. First the velocity-stabilized loop’s mechanism linear model is established through the usage of direct current torque motor’s kinetic equation, and further simplified by application of the characteristics of the current loop. Linear model identification is conducted by using data obtained during actual experiments, and where the non-linear friction characteristics that the platform faces under low frequency low speed environment is being observed. Based on Stribeck friction model and targeting the characteristics of the platform, two types of refined friction models are introduced to describe the non-linear friction in the different types of stabilized platforms. The velocity-stabilized loop’s synthesis non-linear model is obtained, by combining the mechanically derived linear model and the improved Stribeck friction model. Taking into consideration of the model’s non-linearity and multi-parameter characteristics, genetic algorithm (GA) that has global optimization ability is used to identify the parameters. The model that has been identified is verified through root mean square error (RMSE) and isolation performance, and the model’s high precision performance and applicability are confirmed. Additionally, Simulink in MATLAB is used to build the velocity-stabilized loop’s simulation system.(2) Based on the identified model, the non-linear friction feed-forward compensation strategy is designed, optimized and applied to the actual inertially stabilized platform. According to the non-linear friction model part, the corresponding feed-forward compensation strategy is designed and applied to the simulation system. The feed-forward compensation design is then optimized to counter certain aspects during actual engineering application. Then the feed-forward compensation strategy is applied to the modified platform’s turbulence tracking system and actual experiments are conducted to verify the strategy’s effectiveness. Lastly the compensation strategy is applied to the actual platform’s turbulence isolation system and its performance is analyzed through actual experiments.(3) Based on the friction feed-forward compensation strategy, two types of intuitive model reference adaptive controllers (MRACs) were built. The first type of MRAC is based on the state space equation in continuous time, and the second type is designed based on gradient descent in discrete time. Both types of MRACs’effectiveness and the ability to tolerate non-linear friction were tested by conducting simulation experiments on the carrier’s turbulence tracking system. As during actual usage, the turbulence signal of the isolation system cannot be measured, therefore the carrier’s turbulence observer has to be derived through the kinematic equation of the platform’s multi-gimbal. Lastly both types of MRACs are applied to the stabilized platform’s turbulence isolation system, and actual experiments are conducted and their performance is analyzed.(4) To standardized the calculation of the control strategy, the design of direct MRAC within the velocity-stabilized loop had to be reconsidered based in discrete time. Through model matching and assumed conditions the adjustable controllers’ existence and uniqueness are confirmed, and according to the matching equation the control law and the parameter estimation’s observation equation are derived. Hence, MRAC’s standardized structure is obtained, where its controller and adaptive law are relatively independent during design. Improved projection algorithm and fading memory recursive least square algorithm were selected to design the adaptive law, and the performances of both types of MRACs were verified in the turbulence tracking system’s stimulation experiments. Through the usage of a composite control structure diagram, the MRAC’s stableness and convergence conditions are analyzed. By applying both types of controllers to the actual platform’s turbulence isolation system, actual experiments and performance analyze are conducted. Lastly this thesis compares the principle, structure, stimulation results and practical experiments results of the four types of MRACs that were introduced.This thesis has a relatively solid applied engineering background when taking a look at the airborne inertially stabilized platform’s research and experimental methodology, and conclusion. Hence it has good reference usage for high precision internally stabilized platform’s modelling analysis, error compensation and model reference adaptive control research. |
PDF Full Text Request