Based on physiological properties of synapse, soma and axon, this paper presents and analyses a model of neural circuit which can approximately simulate input-output relation, strength-duration curve, adaption and nonlinear connection of real neuron. The obtained result; show that the model approximates to realistic principles of neural computation better than the available neural network. The impulse -coded WTA (Winner takes all) network constructed with the above model finds the winner more effectively th...

A fault fuzzy diagnostic system (FFDS) based on neural network and fuzzy logic hybrid is proposed. FFDS consists of two modes: a fuzzy inference mode and a rules learning mode. The fuzzy inference rules are stored in the memory layer. The excitation levels of the memory neurons reflect the matching degree between the input vector and the prototype rules. In the rules learning mode, the rules can be produced automatically through the cluster process. As a application case of this diagnostic system, the fault diagnosis experiment of the rotating axis is simulated.

This paper gives moment s description of 3-D object and presents similar moment matching detection method for binary discrete objects from 3D-space of series slices. Theoretical analysis and experimental results show that it saves computer memory extremely and can easily be realized, and it decreases calculation and improves detecting ratio. Number of objects, start section and end section centers of each object are obtained after processed whole series slices.

An optimal tolerance design method is presented. The method uses an initial objective function in the design process, and a close-center objective function for reducing the design time. The orthogonal array optimization method is used in order to simplify the method. The center values and tolerances of design parameters can be determined simultaneously by searching out the radius and center of the biggest inscribed globe. Finally, it is necessary to design center value and tolerance synthetically for reaching the aim of optimal design. Two examples are presented to illustrate that the method is feasible and has superiority.

The concepts of complementary cofactor pairs, normal double-graphs and feasible torn vertex sets are introduced. By using them a decomposition theorem for first-order cofactor C(Y) is derived. Combining it with the modified double-graph method, a new decomposition analysis-modified double-graph decomposition analysis is presented for finding symbolic network functions. Its advantages are that the resultant symbolic expressions are compact and contain no cancellation terms, and its sign evaluation is very...

The development of modeling for access way of loop LAN is reported. Moreover, the mathematical analysis and simulating experiment for feedback M/G/ 1[K]/S/FCFS model are also given.

The generalised magnitude-symmetry theorem for lossless network is extended, the phase information previously omitted is supplemented. The extended theorem is equivalent to the unitary condition for lossless network rigorously. Two illustrations are given for the applications of the theorem to analysing the properties of lossless network.

The expression of dyadic Green s function for one and two cylindrical layers of chiral medium are derived by using the method of scattering superposition, for which the electric current sources are placed both inside and outside of a chiral cylinder and a cylindrical chirodome. Then the radiation characteristics of a point dipole antenna on the axis of the chiral cylinder and cylindrical chirodome are analysed. The results show that the polarized states of radiated fields can be changed by choosing the size of chiral cylinder or the thickness of cylindrical chirodome. Also, the expressions of dyadic Green's function given in thid paper can be directly use to analyse the radiation characteristics of cylindrical chirostrip antenna.

By applying the perturbation theory and complex-point method, theoretical research of the measurement of complex permittivity of uniaxial anisotropic materials, by means of an electromagnetic open resonator has been made, and the double refraction phenomenon due to the anisotropy of measured dielectric materials has been quantitatively analyzed. Finally, at 8mm band, an automated electromagnetic open resonator measurement system was used, and measurements have been made on some single-crystal quartz speciments.

A class of clock-cross-controlled sequences over GF(p^α) (p ≥ 2 is a prime number, α >1 is a positive number) using the quadratic character of GF(p^α), i.e., LSR_g[d₀,d₁,d₂]-clock-cross-controlled sequences is introduced. When the feedforward function g(x₁,x₂,…,x_n), which is used as the controlling shift clock, is a quadratic form of GF(p^α), the properties of the period and the linear complexity of this sequences are discussed.

Evaluation by running window smoothing is used for the digital pro-cessing of the polarization of geophysical ULF signals. The observed signals are resolved into two orthogonal complex components so that it is no longer necessary to consider the phase and amplitude of the signals simultaneously.

A new idea of representing the mutual injection locking phenomena in the short-slot hybrid power combiner by a model of negative nonlinear network is proposed. This model thoroughly represents the nonlinear characteristic of the combiner. After the nonlinear equations of the combining system are derived, a method of optimizing the parameters of the combined oscillators to achieve a maximum power combining efficiency is given. The optimum results agree with experimental ones quite well.

The topological expressions of the circuit elements with meshs are put forward. Then, direct algorithm of mesh equations of the active multi-winding coupled network, which is planar, is formulated. The number of the network equations equals the mesh number of the network. It is less than the equations of the modified nodal approach. The procedure of forming equations of this algorithm is simpler than that of the current nodal and the loop approach.

A unified model of low temperature current gain of polysilicon emitter bipolar transistors based on effective recombination speed method is presented, incorporating bandgap narrowing, carrier freezing-out and tunneling of holes through interface oxide. The modeling results based on the unified model are in good agreement with experimental data.

A uniformly valid approximate solution of a kind of nonlinear wave equations is studied, the research results indicate that the solution of this kind of equatiors can be represented by Airy function approximately. The usually used W. K. B. approximation is the first order approximation of the present result in the region far away from the turning point of refractivity. At the turning point of refrac-tivity, the present result is still valid.

Vector On-Surface Radiation Condition (VOSRC) and Moment Method (MM) are applied to analyzing the performance of a wire antenna in the presence of a conducting sphere. The results are in good agreement with that obtained by generalized multipole technique (GMT).

A matrix method used in multilayer stack of dielectric films is applied to planar dielectric optical waveguides. A simple and applicable method for obtaining characterristic equation is presented.

Macromodeling methods of large scale integrated circuits used in the colour TV CAD system are introduced. The prospects of application and the ways of development of these methods are discussed.