No abstract

In this paper, the problem of nonfragile reliable guaranteed cost control for networked control systems with time-varying sampling period sensors and actuator failures is concerned for a given quadratic cost function. For linear time-invariant controlled plant, under the assumption that the sampling period is time-varying within a certain known bound, the time delay is constant and shorter than the sampling period of the sensors, the system is transformed into a time-varying discrete time system, where the time-varying sampling period parts are transformed into norm bounded uncertainties of the structure parameter. A new linear matrix inequality (LMI)-based approach is proposed to derive a sufficient condition for the existence of nonfragile reliable guaranteed cost controller. Furthermore, the design method of the optimal nonfragile reliable guaranteed cost controller is formulated to minimize the upper bound of the closed-loop system cost. A numerical example is given to show the usefulness and effectiveness of the proposed method.

Dynamic power optimization is an important part of the power optimization for embedded parallel processing. Dynamic power estimation is the premise for dynamic power optimization. The evaluation model with fast calculating speed and high accuracy can improve the efficiency for dynamic power optimization. The dynamic power evaluation methods based on low level simulation have high accuracy. But it is very time consuming. The high level dynamic power estimation models have higher speed, but have lower calculation accuracy. For this issue, the paper proposes a dynamic power estimation method based on component which combines ow level dynamic power evaluation methods with high level dynamic power estimation methods. The dynamic power of gray level co-occurrence matrix(GLCM) and fractal dimension(FD) in the remote sensing cloud detection based on texture feature is evaluated using the proposed method. The average error of the dynamic power for GLCM and FD is 11.86% using the proposed method. The computing time for GLCM and FD using the proposed method is 0.295 times than the method based XPower.

No abstract