Failure Detection and Adaptive Compensation for Fault Tolerable Flight Control Systems

This paper presents an advanced aircraft flight control system which is capable of providing failure detection, identification, and adaptive compensation in the event of the damage of control surfaces. The new approach is based on the detection of immediate changes of correlation between pitching, yawing, and rolling rates. The failure compensation mechanism is realized by an adaptive wavelet neural net-based proportional-integral-derivative (PID) controller. The fault-tolerant mechanism is activated to regain effective control of the damaged aircraft whenever a control surface failure is detected. Design of a real-time executive kernel based on the RT-Linux operating system is experimentally conducted to realize and verify functions of our proposed design.