Análise da estabilidade e do desempenho de sistemas de controle via redes de comunicações com atraso aleatório.

Abstract

In this thesis we analyze the performance and stability of a Networked Control System (NCS) that presents random delays in the communication process among the system's devices connected to the network medium. These delays may degrade the system's performance and even destabilize i t . Random delay models are motivated by both network features (protocols) and timing features of system's devices (event and time driven networks). The complexity of control systems analysis and design is increased when a random delay model is adopted. This happens because the system is no longer time-invariant, so that new analysis tools must be investigated. To help the design and the analysis of NCS with random delays, it is proposed in this thesis a stability analysis method based on probability density functions (pdf) of the closed-loop eigenvalues. By using the pdf it is possible to determine the probability of having unstable modes for a given system configuration. Another useful tool is the state covariance. This tool allows analyzing the evolution of the probabilistic behavior of the system's variables through the time. The main contribution of this thesis is in the combination of pdfs, covariance and cost functions in order to compare available control strategies.