| omputer controlled systems are being widely used in industry. This has been made possible by the rapid progress in microelectronics and computers in the last couple of decades. However, the theoretical developments on the analysis and design of such digital controllers have not kept pace with their applications. The popular classical techniques either design a continuous-time controller and discretize it or design a controller for a discretized approximation of the continuous-time process. While the first method leads to good performance of the digital control system only for very high (often unrealistic) sampling rates, the second method completely ignores inter-sample behavior.;We approach digital control systems as hybrid systems containing both continuous-time and discrete-time signals. Our methods exploit the fact that a digital control system is essentially a time-varying periodic system with a special structure that resembles a time-invariant system. The signal and plant uncertainties are treated directly in continuous-time. The techniques that we have developed are novel in that they handle inter-sample behavior directly.;We have considered the characterization and computation of various system norms for sampled-data systems. The results appear to be the most natural and direct generalizations of well-known results for linear time-invariant systems and they can be easily implemented using standard numerical analysis software. We have defined a notion of... |
