航空学报 > 2004, Vol. 25 Issue (5): 480-484

基于自适应不对称高斯基函数网络的可靠飞行跟踪控制

刘亚, 胡寿松   

  1. 南京航空航天大学自动化学院 江苏南京 210016
  • 收稿日期:2003-06-03 修回日期:2004-03-31 出版日期:2004-10-25 发布日期:2004-10-25

Reliable Flight Tracking Control based on Adaptive AGBF Network

LIU Ya, HU Shou-song   

  1. College of Automation Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing210016, China
  • Received:2003-06-03 Revised:2004-03-31 Online:2004-10-25 Published:2004-10-25

摘要: 针对歼击机操纵面结构故障,提出了一种基于自适应不对称高斯基函数网络(AGBFN)的可靠跟踪控制方案。该方案的特点是在传统的线性控制器的基础上,引入基于自适应AGBFN的自适应控制器,用于在线补偿由于建模误差、外扰、以及操纵面卡死或损伤所造成的影响,并且采用控制隔离技术来处理执行机构的饱和问题。完全自适应AGBFN采用不对称高斯基函数,并且可以在线更新网络所有参数,克服了传统RBF网络对称性约束,提高了网络的适应性和学习能力,从而保证歼击机的操纵品质,且输出较好地跟踪参考模型输出。飞行仿真结果表明,文中采用的控制方法使飞机在正常状态和故障状态下均可获得满意的控制效果。

关键词: 不对称高斯基函数, 自适应神经网络, 可靠控制, 歼击机

Abstract: A reliable flight tracking control scheme is developed for a fighter in the presence of control surface failures based on adaptive asymmetric Gaussian basis function network (AGBFN). An adaptive AGBFN controller is used to augment the linear controller, which can compensate the modeling error, disturbance and unknown control surface failures. The effect of actuator dynamics and limits are treated by using control hedging. The asymmetric Gaussian basis function is used for the proposed full adaptive AGBFN which eliminates the symmetry restriction of traditional RBFN and provides the neurons with higher flexibility. And the tuning rule for updating all parameters of the AGBFN is derived which ensures the stability of the overall system. By the proposed control scheme, the handling qualities of the system are maintained with the improved tracking accuracy. Simulation results show that the proposed method is suited to not only the normal aircraft, but also the fault aircraft.

Key words: asymmetric Gaussian basis function, adaptive neural network, reliable control, fight aircraft