Hybrid adaptive control

Synthesis of a controller for a random parameter linear system is difficult when the state measurements contain noise. Although representational results have been derived, no explicit solution has been obtained in any reasonably general circumstance. An approximation to the quadratic-optimal regulator problem is presented for a situation in which there is an unconventional measurement architecture. The solution is in a form quite similar to that obtained in the complete observation case, but the gain equation is made more complicated by the presence of noise.