that the reference will be known in advance so that the controller can be designed accordingly. For instance, the feedback law can be chosen to give stability relative to a nominal trajectory as was discussed in Chapter Five. There are several aspects of the approach taken here which deserve investigation. Some of these are: (1) Analyze the situation which occurs when the internal model system's poles are not placed at the correct locations. For example, what is the consequence of placing poles at ąj2 when they are actually needed at ąj2.1. (2) Find a way to bound the tracking error when (A.2) is not satisfied. Also show the relationship between the error bound and the number of terms taken from a Fourier series expansion of the true state and input. (3) Develop a feedback law which either gives global stability or increases the region of stability for the system NCT. Since the resulting feedback may be nonlinear, further conditions are likely to be imposed for a solution to the servomechanism problem. One scheme which warrants investigation is the use of prestabilizing nonlinear feedback on the plant. (4) Investigate the effect of the proposed scheme on the transient behavior of the system. Although it is true that the steady-state error can be improved with this scheme, the use of an internal model system of large dimension may degrade transient performance.