船舶航向自适应复合度SVM-H^(∞)/H^(2)控制系统研究

Research of Adaptive Compound SVM-H^(∞)/H^(2) Control on the Ship Course-Keeping System

针对船舶航向控制系统中存在的海浪干扰不确定性、系统参数摄动性、非线性及耦合性等复杂性问题,提出了航向自适应复合度SVM-H^(∞)/H^(2)控制策略。利用支持向量机良好的非线性映射能力、自适应学习能力和信息并行处理能力,设计了基于系统正向模型SVM辩识器的逆模型前馈补偿控制器,并采用自适应调节权重因子前馈补偿控制策略与状态反馈H^(∞)/H^(2)鲁棒控制器构成了自适应复合度控制器,使得控制系统航向控制精度高、响应速度快、鲁棒性强、节能效果好。采用固定尺度记忆增量支持向量机使得辨识器和控制器具有信息实时更新、参数实时修正的功能,采用自适应粒子群参数寻优器保证了支持向量机参数的寻优精度和速度。数字仿真试验结果表明,所提出的航向控制系统性能优良,具有工程价值。

In order to overcome the complexity in the ship course control system,such as the uncertainty of sea wave interference,system parameter perturbation,nonlinearity and coupling,an adaptive compound SVM-H^(∞)/H^(2) control strategy is proposed.Using the excellent nonlinear mapping ability,self-adaptive learning ability and information parallel processing ability of support vector machine(SVM),an inverse model feedforward compensation controller based on the system forward model SVM identifier is designed.We also adopt the feedforward compensation control strategy with adaptive control weight factor and the state feedback H^(∞)/H^(2) robust controller to form the adaptive recombination controller,which makes the control system have high control precision,fast response speed,strong robustness and good energy-saving effect.The fixed-size memory-incremental least squares support vector machines,which is adopted in this paper,makes the identifier and controller have the functions of real-time update of information and real-time correction of parameters.The adaptive particle swarm optimization(PSO)optimizer ensures the optimization precision and speed of the SVM parameters.The results of the digital simulation test show that the proposed course control system has excellent performance and engineering value.