zbMATH — the first resource for mathematics

Controllers for LQG self-tuning applications with coloured measurement noise and dynamic costing. (English) Zbl 0592.93036
The paper studies a single input/single output LQG control problem by a frequency domain type of analysis. The cost function includes dynamic weighting elements, which allows integral action to be introduced and robustness characteristic to be modified. The output of the plant y, which is affected by a disturbance d, must follow the reference signal r, whereas the total observed output is \(y+n\), where n is the measurement noise. Here d, r and z are linear transformations of white noise. Such a model is motivated by some industrial application (ship steering problem) given in the paper. Simulation results are also presented. The control problem is solved with a polynomial system approach. The author discusses also how to incorporate an estimation algorithm for the plant parameters and/or the measurement noise polynomial to build a self-tuning controller. The reference and the disturbance signals are supposed to be almost constant, so that it is possible to cancel these terms with an integral action and identify only the measurement noise.
Reviewer: M.Piccioni

93C40 Adaptive control/observation systems
93C05 Linear systems in control theory
93E20 Optimal stochastic control
93B50 Synthesis problems
93C55 Discrete-time control/observation systems
93E12 Identification in stochastic control theory
Full Text: DOI