This thesis deals with the control of a highly maneuverable miniature helicopter about trajectories, generated online, from a library of prespecified maneuvers. Linearizing the nonlinear equations describing the helicopter dynamics about the prespecified, library maneuvers results in a hybrid linear time-varying (LTV) model. Two control approaches are used to design controllers corresponding to each library maneuver: the standard L2-induced norm approach and an approach which also uses the L2-induced norm as a performance measure while accounting for uncertain initial states. Each control approach is evaluated in closed-loop simulation with a nonlinear helicopter model. The controllers are set to drive the helicopter model to track desired trajectories in the presence of disturbances such as wind gusts, turbulence, sensor noise, and uncertain initial conditions. For the specific plant formulations and trajectories presented, performance is comparable for both control approaches; however, it is possible to improve controller performance by exploiting some of the features of the approach accounting for uncertain initial states. These improvements in performance are topics for future work along with implementation of the presented approaches and results on a remote control helicopter. / Master of Science