Research On Information Measurement System And Some Problems

As there have been no a relatively unified theory in current measurement systems, in this dissertation, the thoughs and methods of direct information measurement are proposed based on entropy sets including various entropies, information amount and entropy production. An information measurement system is built according to the principle of maximum information acquisition. It not only avoids the information loss problem in measurement process based on the signal, and achieves a theoretical framework based on entropy measurement basis. The major contributions of this dissertation are specifically stated as follows.This dissertation first describes the measurement system and proposes the block diagram of the information-driven measurement system, with which the mathematical model of this system is built. And these basic characteristics of this information measurement system are summarized through discussing the relationship range among the measuring information amount, the error entropy and the entropy of source.Second, the essence of the Clausius entropy, the Bolzmann entropy and the Shannon entropy are analyzed within the framework of the information energy. And the result shows that the three kinds of entropy have the same soul. Combinating the variation of entropy, entropy flow and entropy production, this dissertation proves the equivalence between the maximum information entropy and the principle of maximum entropy production in non-equilibrium system. The concept of information power is introduced and the electrical network theory is used to analyze the coupling relationship between the information source and sensor. Furthermore, the information model of the sensor is proposed.Third, the linear network entropy theorem and linear network measurement information theorems are proposed through analysesing the variation of information entropy of the linear network. On this basis, the changes of measurement information through the linear network which contained two different kinds of noise are discussed respectively. These theorems are applied to the selected frequency network, Z transform and the process of information quantization. Fourth, the information processing algorithm based on maximum entropy is studied. The maximum entropy fuzzy clustering method is attemped to partition information sets, and constructed the general objective function whose convergence is proved. The algorithm has been applied to the analysis measurement data. The maximum entropy entropy distribution theorem and the principle of maximum entropy spectral density analysis are analyzed from the perspective of information measurement. The algorithm of independent component analysis applied to a hybrid data processing is described, and simulation experiments are given.At the end of this dissertation, the measurement errors are analyzed using the information entropy. The logarithm relationship between error entropy and noise power, error entropy and uncertainty are derived respectively, and the related performance evaluation base is built in information measurement system.