Lateral Control Law Design for Helicopter Using Radial Basis Function Neural Network

As fixed-parameter control can not satisfy control requirements when helicopter is aviating in large scale flight envelop, this paper proposes a new control law design to adjust parameters on-line. Firstly a parameter-mapping approach is developed to design flight control parameters at certain flight conditions according to the desired system performance. Then parameters obtained at given conditions are used to train Radial Basis Function Neural Network (RBFNN). Thus RBFNN can generalize the given flight conditions information and output appropriate control parameters which will meet control requirements for any current flight condition within the given flight envelop. Simulation results indicate this control law design is feasible and effective.