Low-power CSD linear phase FIR filter structure using verticalcommon sub-expression

A high-speed/low-power canonic signed digit (CSD) linear phase finite impulse response (FIR) filter structure using vertical common sub-expression is proposed. In the conventional linear phase CSD filter, the horizontal common sub-expression method (see Hartley, 1996) has been widely utilised due to the inherent symmetrical filter coefficients. However, use has been made of the fact that the most significant bits of adjacent filter coefficients in the linear phase filter are also equal since they have mostly similar values. Through the example, it is shown that the proposed structure is more efficient in the case where bit precision of implementation is lower     

Learning Markov chains in fractal compression of image data

In a recent paper [17] we proposed a stochastic algorithm which generates optimal probabilities for the decompression of an image represented by the fixed point of an IFS system (SAOP). We show here that such an algorithm is in fact a non trivial example of Generalized Random System with Complete Connections. We also exhibit a generalization which could represent the solution to the inverse problem for an image with grey levels, if a fixed set of contraction maps is available. Received: 1 July 2002     

荧光波长对共焦显微镜成像特性的影响

导出了共焦显微镜中不同荧光波长情况下的荧光功率传输函数、三维脉冲响应函数（3D－PSF）和三维光学传递函数（3D－OTF）。结果表明，不同的荧光波长对共焦显微镜的空间截止频率、分辨率、光学传递函数存在明显的影响。当激发波长与荧光波长的比值下降到一定程度时，可以看到明显的失锥现象。     

Design of a 3780-point IFFT processor for TDS-OFDM

This correspondence presents a design of 3780-point IFFT processor for TDS-OFDM terrestrial DTV transmitter using FPGA. It demonstrates the algorithm design and error analysis of the processor, which can achieve a throughput of 7.56M complex IFFT operations per second. This design meets the signal-to-quantization noise ratio requirement of the TDS-OFDM system. It consists of two FPGA and one dual-port RAM. The data stream pipeline algorithm is implemented     

Blowup of solutions for a class of non‐linear evolution equations with non‐linear damping and source terms

We consider the blowup of solutions of the initial boundary value problem for a class of non‐linear evolution equations with non‐linear damping and source terms. By using the energy compensation method, we prove that when p>max{m, α}, where m, α and p are non‐negative real numbers and m+1, α+1, p+1 are, respectively, the growth orders of the non‐linear strain terms, damping term and source term, under the appropriate conditions, any weak solution of the above‐mentioned problem blows up in finite time. Comparison of the results with the previous ones shows that there exist some clear condition boundaries similar to thresholds among the growth orders of the non‐linear terms, the states of the initial energy and the existence and non‐existence of global weak solutions. Copyright © 2002 John Wiley & Sons, Ltd.     

Quantum Cosmology in CGBD Theory

We apply the theory developed in quantum cosmology to a model of charged generalized Brans–Dicke gravity. This is a quantum model of gravitation interacting with a charged Brans–Dicke type scalar field which is considered in the Pauli frame. The Wheeler–DeWitt equation describing the evolution of the quantum Universe is solved in the semiclassical approximation by applying the WKB approximation. The wave function of the Universe is also obtained by applying both the Vilenkin-like and the Hartle–Hawking-like boundary conditions. We then make predictions from the wave functions and infer that the Vilenkin's boundary condition is more reasonable in the Brans–Dicke gravity models leading a large vacuum energy density at the beginning of the inflation.     

An approximate method for computing performance-related reliability of gracefully degrading systems

Using existing methods, the computation of performance-related reliability (PRR) of large-scale gracefully degrading systems is very tedious and time consuming. In this paper, the behaviour of such systems is respectively modeled as two types of diffusion processes according to their reconfiguration coverage. If the coverage is 1 (i.e. the reconfiguration is always successful), it is modeled as a regular diffusion. If, on the other hand, the coverage is less than one, it is modeled as diffusion with killing. Kolmogorov backward equations for regular diffusion processes and for diffusions with killing are then applied to compute the PRR. The methods have been applied in several examples, and the results satisfactorily agree with the accurate results.