Abstract: A robust automatic flight control system is designed for a small scale unmanned helicopter in this paper. We first apply system identification to the design of a high-bandwidth controller for robotic helicopters. The overall concept is to extract a complete set of non-parametric input-to-output frequency responses that fully characterize the coupled helicopter dynamics, and then apply a nonlinear search algorithm for a linear high-fidelity simulation model that matches the frequency response data set. The H_∞ loop-shaping method is used to design the inner loop of the unmanned helicopter based on the identified model. The greatest common right divisors (GCRD) method is employed to move the transfer function matrix from the real system to the target system, which is a very useful way to solve the difficulties in choosing a proper weighting matrix in H_∞ loop-shaping. Compared with the traditional method, the system using the new method exhibits a larger robust stability margin, and the decoupling and the bandwidth of the system are also improved considerably. Furthermore, it reduces the complexity and blindness for the designer to find the proper weighting matrix. The simulation results prove that the unmanned helicopter system achieves a top level control performance that conforms relevant requirements in military standards ADS-33E.

Key words: helicopter, robust control, system identification, greatest common right divisors, H_∞ loop-shaping, aircraft flying qualities