Stability Analysis of a Nonlinear Plant with Limited Information

This paper studies the characteristics of a specific model-based control scheme in which sampled and corrupted information is used to govern the state of a partially known plant. More precisely, it is assumed that signals can be transferred from a controller to the plant continually, while information can be transferred from the plant to the controller only at sampling times. Information coming from the plant is assumed to be corrupted by an unknown disturbance, that may represent the effect of quantization, delay, measurement errors. Using an approximate model of the plant and the available information, the controller generates a feedback signal that is used to govern the state of the plant. The specific control objective is that of keeping the state of the plant bounded. The performances of the control scheme depend, clearly, on the length of the sampling intervals and, therefore, one is interested in finding conditions on the length of the sampling interval which makes possible to achieve the control objective. The paper gives two results in this direction, stating sufficient conditions on the length of the sampling interval for guaranteeing boundedness of the state of the plant in two different, very general situations.