Neuro-aided H2 Controller Design for Aircraft under Actuator Failure

Wang, Zhifeng; Xiong, Gang

doi:10.1007/978-3-642-28308-6_61

Cited by 2 publications

(1 citation statement)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Extended MRAN (EMRAN) [13] is a faster implementation of MRAN where the parameters of the most dominant neuron only are updated. Based on EMRAN several neural aided controllers were designed for autolanding a typical modern high performance aircraft under severe winds and unknown actuator failures [2,12,[14][15][16][17]. The conventional PID, LQR, ∞ H or 2 H controllers were used as baseline controllers for on-line learning of the networks.…”

Section: Introductionmentioning

confidence: 99%

Improved neural-aided sliding mode controller for autolanding under actuator failures and severe winds

Ismail

Pashilkar

Ayyagari

et al. 2014

Aerospace Science and Technology

View full text Add to dashboard Cite

In this paper, the fault tolerant capabilities of the neural aided sliding mode controller for autolanding under actuator failures and severe winds developed earlier are improved significantly by incorporating a novel anti-windup strategy and a phase compensation scheme. This controller further increases the size of the fault tolerance envelope for various types of control surface stuck faults and provides complete coverage at every point within the envelope boundaries. Earlier work by the authors showed the existence of a neural-aided sliding mode controller which could handle a wide range of actuator stuck faults. One of the major drawbacks of this earlier controller is that it does not ensure that all points within the range of minimum and maximum bounds of the fault tolerance envelope are covered. The anti-windup proposed in this paper is a generalization of the scheme used for proportionalderivative-integral controllers to the cascaded trajectory following controllers designed by the authors. This scheme can handle requirements of state limiting as well as multiple redundant control surface saturation. The proposed anti-wind up design is a simplification over the command filter approach used for adaptive backstepping. The approach is demonstrated for a fixed-wing aircraft undergoing unknown actuator stuck failures and subject to severe wind disturbances during autolanding. An example of three control surface failures (both ailerons and rudder) handled by this controller is also presented.

show abstract