Stabilization with data-rate-limited feedback: tightest attainable bounds

This paper investigates the stabilizability of a linear, discrete-time plant with a real-valued output when the controller, which may be nonlinear, receives observation data at a known rate. It is rst shown that, under a nite horizon cost equal to the mth output moment, the problem reduces to quantizing the initial output. Asymptotic quantization theory is then applied to directly obtain the limiting coding and control scheme as the horizon approaches in nity. This is proven to minimize a particular in nite horizon cost, the value of which is derived. A necessary and su cient condition then follows for there to exist a coding and control scheme with the speci ed data rate that takes the mth output moment to zero asymptotically with time. If the open-loop plant is nite-dimensional and time-invariant, this condition simpli es to an inequality involving the data rate and the unstable plant pole with greatest magnitude. Analagous results automatically hold for the related problem of state estimation with a nite data rate. c © 2000 Elsevier Science B.V. All rights reserved.