By considering an analogy between information theory and predicate logic in artificial intelligence (AI) systems, we obtain interesting results in the field of both AI and information theory. First, we define the self-information and the entropy of formulas in predicate logic. Since inference of logic is regarded as information transformation, the information structure of logic is clarified by using mutual information based on the analogy with source coding. Next, we introduce a new concept of hierarchy of information, which have not been very clear in information theory, to discuss the similarity between inference and source coding. Finally, a new theoretical method is proposed for processing uncertain knowledge which can not be treated by the ordinary logic from the foregoing information theoretical concepts.

This paper deals with the recent trends of reseaches on intelligent image coding technology focusing on model-based analysis synthesis coding. By means of the intelligent image coding scheme, we will be able to realize epock-making ultra-low-rate image transmission and/or so-called value-added visual telecommunications. In order to categorize the various image coding systems and examine their potential applications in the future, an approach to define generations of image coding technologies is presented. The future generation coding systems include the model-based analysis synthesis coding and knowledge-based intelligent coding. The latter half of the paper will be devoted to the recent work of the authors on the model-based analysis-synthesis coding system for facial images.

This article briefly reviews the elementary concepts of the quantum communication theory. Specific topics on some of the optical communications by utilizing quantum nature of light wave will be discussed, especially for applications to information transmission of classes of squeezed state exceeding the present limitations, and for quantum coding theory relevant the photon communications.

Trade-offs between storage and error in the simple binary-valued retrieval problems are analyzed by using the rate distortion theoretic method from the viewpoint of the trade-offs evaluation model. These problems, the table lookup question, the exact match question, and the set of all binary-valued questions, have been proposed as the typical basic model of the information retrieval systems to evaluate the memory cost, the access cost, the state cost, the logic cost, and their relationships. The evaluation critera of memory-error trade-offs are the elastic condition and the excess of information. The former means that drastic savings of the amount of storage with a small error tolerance is feasible. The latter means the measure to evaluate the degree of the possibility for the system to achieve elastic condition. As a result, although the set of all binary-valued questions admits the possible highest excess of information, its statistical property is almost equivalent to that of the table lookup question, and only the exact match question satisfies the elastic condition. Furthermore, considering the conjunctive model which means the combination of each questions for the basic model, the property of elasticity or inelasticity for the basic model is changed from that of the basic model in accordance with the degree of the combination of the questions.

Recently, a fail-safe principle in wide sense was proposed by one of the authors, in order to accomplish the flexible security against accidents or failures. This new principle is established by the following idea: i.e., if the tolerable range of the deviated output response as a behavior of system is definitely estimated, in advance, and if the response can be found in this range, we can always use this response as available. Even if the deviation range of the output response extends far from the tolerable one, it is sufficient that the system itself has a methodology to find the available response in the tolerable range, and the system behaves by this response. This new concept of an extended type of the fail-safe principle is a significant application of the mathematical theory of nondeterministic fluctuations of systems which was presented by the authors. In this paper, the foundation is given to this principle by developing its mathematical theory, in detail.

This paper discusses branch metric computation in the main decoder placed in an SST (Scarce State Transition) Viterbi decoder. The basic assumptions that all the message sequences are equally likely and that the channel is memoryless do not hold for the main decoder in an SST Viterbi decoder, when an inverse encoder or a pseudo-inverse encoder is used as a pre-decoder. Therefore, in contrast to a conventional method, an MAP (Maximum A Posteriori probability) estimation method itself, which is the starting point of the maximum likelihood decoding, has been applied to branch metric computation. Then, it has been clarified that the conventional branch metric computed in a usual SST Viterbi decoder happens to be equal to the branch metric derived using the MAP estimation method only for systematic codes. For general non-systematic codes, in particular, it has been found that it is impossible to decompose a path metric into such branch metrics that follow the original code trellis structure, because the branch metric derived at time k is also dependent on future state transitions.

This paper proposes a new evaluation method for ARQ schemes. The decoded bit error rate versus E_b/N₀ and the bandwidth expansion ratio are adopted as criteria to measure performance capability. The method makes it possible to compare the performance capabilities of various error control schemes on the same basis. Since this method clarifies the E_b/N₀ and the bandwidth required for obtaining the desired BER, it gives some useful indication on deciding which scheme to use in a particular application. FEC, ARQ, Type- hybrid ARQ, and Type- hybrid ARQ, each of which uses a binary BCH code with a common code length, are taken up as examples for analysis. The numerical results show that well designed hybrid ARQ schemes can considerably reduce E_b/N₀, at a cost of a small bandwidth increase. Based on the foregoing analysis, this paper also proposes a constant-rate ARQ scheme with finite buffers. This scheme offers a constant throughput rate and a constant delay time as a conventional FEC scheme. It requires less E_b/N₀ than an FEC scheme. The numerical results are presented for some examples. The results on the fixed bandwidth expansion ratio show that the proposed scheme with a binary BCH code requires less E_b/N₀ than an FEC scheme using a convolutional code and soft decision Viterbi decoding. The decoder of the proposed scheme is much simpler than a Viterbi decoder. This proposed rate-constant ARQ scheme is suitable for application to those systems in which a feedback channel is available and a constant throughput rate and a constant delay time are required.

In this paper we propose a novel scheme for fault detection in Random-Access Memories (RAM) using pseudorandom M-sequences. Only static pattern-sensitive (weight sensitive) faults are considered and three different versions of the scheme, based on the fourth and the fifth order M-sequences, are discussed. The conditions required to obtain complete test set are derived. Since the proposed schemes use simple hardware like linear feedback shift registers for test pattern generation, the schemes are suitable not only for simple on-line testing, but also for built-in self test of memory IC.

With popularity of radio communication, voice scrambling has received increased attention for communication privacy. It has been known that pre-emphasis at the transmit side and de-emphasis at the receive side improves the received signal to noise power ratio (SNR) both for scrambled and nonscrambled signals. In this paper, optimum pre- and de-emphasis are theoretically investigated for FM and PM transmission including frequency-domain scrambling for voice security. Using a linear system model, the optimum frequency-characteristics for pre- and de-emphasis filter are given as a function of a voice signal spectrum, demodulator output noise spectrum, and the scrambling method. Comparisons of the maximum attainable SNR are made for various case including plain voice (non scrambling) and secured voice scrambled by spectrum inversion, reverberation and FFT spectrum transformation, assuming simple voice and noise spectra.

An AR model is one of linear stochastic equations, which is characterized by the eigenvalues and eigenvectors. Since the poles in the AR model correspond to the eigenvalues in linear equations, the weights of poles in the AR model correspond to the eigenvectors in linear equations. The AR type model has essentially two types poles; system poles and virtual poles corresponding to system zeros. These poles can be distinguished by observing the weight of each pole in the partial fraction expansion of the AR model transfer function. The rules for separation of AR poles are: (a) If the weight of an AR pole is constant for AR model order change, the AR pole is a system pole. (b) If the weight of that is inversely proportional to the AR model order, the AR pole is one of virtual poles.

In this paper, a stochastic signal processing method of analog type for the prediction of power state variable is first discussed by considering the effect of an internal mechanism of an instrument with mean squaring operation. Next, in the actual case of quantized observation contaminated by an additive random noise, a wide sense digital filter estimating the power state variable of stochastic systems is proposed. These methods are applied to the actual data measured in acoustic environment.

As is well-known, the ordinary regression analysis method is confined to a simplified linear model of the estimation based on the Gaussian property and a least squares error criterion. Then, usually the prediction is done through the transformation based on this regression function. In this paper, a new trial for the regression analysis is proposed especially in the form matched to the complexity of physical phenomena and stochastic signal detection under the existence of background noise. Furthermore, the prediction of the output probability distribution is done based on the regression relationship with less information loss. Finally, the effectiveness of the proposed method is confirmed experimentally by applying it to the actual acoustic data.

In optical coherent communications with homodyne detection system, it is one of the most essential problems to synchronize the phase of local oscillator with that of received signal on the receiver. In general, the phase-locking performance is limited by frequency fluctuation and quantum noise of light source. So far some optical phase-locked loops have been proposed in order to optimize phase-locking performance. Their main purpose is to minimize phase error variance of systems with frequency fluctuation and quantum noise transformed into the electric region. To improve the phase-locking performance under the same situations, this paper proposes a new optical phase-locked loop with received quantum state controller, called a Squeezed-PLL, which can reduce the impact of quantum noise in the optical region. This system can eliminate the effect of the vacuum noise due to the beam splitting. Finally, the general signal to noise ratio of data-branch is shown, and phase-locking performance of the Squeezed-PLL is verified by computer simulation.

We show the general formula of the signal to noise ratio in chains with amplifiers and attenuators using nonclassical properties of light. From the results, we show that quantum noise can be effectively controlled. For the application we show that the optical fiber transmission system with conventional optical amplifiers can be designed as if the noiseless optical amplifiers are used.

Sorting is a fundamental problem of information processing both in theoretical interests and in practical applications. Enormous efforts have been made to develop high speed sorting algorithms, especially suited for SIMD or MIMD type parallel computers built in VLSI technology, especially for an nn 4-neighbor mesh-connected parallel computer. In this paper, we deal with a sorting algorithm assuming an eight-neighbor mesh-connected parallel computer with no wraparounds in rows or columns. The machine consists of N=nn processors (processing elements) which are arranged in a two-dimensional array and can communicate, if exist, with 8 neighbor processors. Our algorithm fully makes use of the communication capability, that is, we introduce a primitive operation which compares and exchanges four elements, simultaneously, between four processors. The algorithm sorts nn elements into shuffled row-major order and yields the sorting time of 2(4n-2log n-5)t_r + (log²n-log n+1)t_c. where t_r and t_c are defined as the times for a unit routing step and a comparison, respectively. It is shown that the algorithm is faster than ones for a 4-neighbor mesh-connected computer presented earlier by Nassimi and Sahni or Kumar and Hirschberg. The speedup is due to adequate combinations of the primitive operation in an 8-neighbor mesh-connected array with the property of shuffled row-major indexing. This paper is organized in the following manner: Firstly, we illustrate the structure and the function of an 8-neighbor mesh-connected parallel computer. Next, we describe and discuss our sorting algorithm precisely. Finally, we derive the sorting time of this algorithm and compare it with those of the algorithms developed for a 4-neighbor mesh-connected computer.

Some concepts in information theory are tried to apply to the study of genes. The mutual entropy is used to define a measure indicating the similarity between two genetic sequences. The alignment of sequences is briefly discussed. Some phylogenetic trees are written by using the entropy measure. According to this results, usefulness of information theory is discussed in the study of genes such as molecular evolution.

A nonrecursive algorithm for address generating during the Hilbert scan of image buffer is presented. The technique is based on two level look-up tables. For large images, the additional storage is less than one percent of image buffer size. The present implementation of the algorithm is only about two times slower than the raster scan. Three level lool-up tables are generated if image data array is too big to fit image buffer and it is accessed from a mass storage. Software tools were developed for look-up tables generating and for the convenient use of the algorithm from C programs.

This letter presents an automatic repeat request/adaptive forward error correction (ARQ/AFEC) hybrid system which adaptively alters the error-correcting capability according to channel state using a nonsystematic Reed-Solomon code. Frame efficiency is also presented.

Aperture field of a small offset cylindrical reflector antenna is computed by using boundary-element method (BEM). The numerically obtained aperture field (both amplitude and phase) of the antenna at the testing frequency of 12 GHz is compared with the corresponding measured data, and a good agreement-between these two fields is observed.

Boundary-element method (BEM) is applied to compute near-field strength of a parallel-plate small-aperture horn antenna which is blocked by a conducting circular cylinder. The computed field strength of the antenna at the testing frequency of 12 GHz is compared with the corresponding measured data (amplitude), and fairly good agreement is obtained.

It is well know that charges stored in condenser decay exponentially in time through resistor in CR-circuit. But in some types of condensers the exponential decay at initial stage breakes down, and the speed of discharge slows down. There the decay has long time tail characteristic to fractal time random walk.

We successfully fabricated a very narrow gap in niobium film using electron beam lithography and RIE. The width of the gap was about 40 nm. We applied this process to the fabrication of Nb-PbIn-Nb planar Josephson junctions, and one of them showed the Dayem effect.

Sampling theorem for all bandlimited distributions which converges in the sense of distributions is presented. An inter-extrapolation formula with the most general sampling points on the negative time axis is provided and especially for equally-spaced sampling points an explicit formula is given, where it is proved that the Newton interpolating polynomial through the finite number of sample values converges with the Mittag-Leffler summation. These formulas are both proved to be convergent in the topology of distribution, hence a consistent theory is accomplished within the scope of distributions. As many linear operations on signals such as differentiation are continuous with respect to the distribution topology, the given formulas exhibit great facility when applied to signal analysis.

This paper presents an efficient algorithm for solving bipolar transistor networks. In our algorithm, the network equation f (x)=0 is solved by a homotopy method, in which a homotopy h (x, t)=f (x)-(1-t) f (x⁰) is introduced and the solution curve of h (x, t)=0 is traced from an obvious solution (x⁰, 0) to the solution (x^*, 1) which we seek. It is shown that the convergence of the algorithm is guaranteed by fairly mild conditions. A rectangular subdivision and an upper bounding technique of linear programming are used for tracing the solution curve. Our rectangular algorithm is much more efficient than the conventional simplicial type algorithms. Some numerical examples are given in order to demonstrate the effectiveness of the algorithm. The advantages of the rectangular algorithm are as follows. (1) Convergence is guaranteed by fairly general conditions. (2) There is no need to evaluate Jacobian matrices. (3) There is no need to invert matrices except for the first step; only pivoting operations are necessary. (4) The replacement rule of vertices is very simple. (5) The computational complexity is markedly reduced compared with the simplicial algorithm. (6) The computational efficiency can be greatly improved by choosing the grid sizes of the rectangular subdivision pertinently according to the nonlinearity of the equation.

An iterative numerical method is used to synthesize a given non-series-parallel minimal LC network with a prescribed driving-point function. Converting the given network to a series-parallel network by a certain reduction method produces the initial element values. The concept of sensitivities of a network's evaluating equations with respect to the element values is used to achieve a convergent solution. The procedure is applicable to canonic and non-canonic networks alike. The element values can be determined to any degree of accuracy desired.

In this paper two classes of multi-state systems are proposed and the concept of modules of those systems is clarified from mathematical aspects. The purpose of this paper is to investigate some properties of modules on basis of the minimal elements which determine the structure of systems. Furthermore, we discuss a criterion, what we call test for modularity", for deciding whether or not a subset of components is a module, and develop three modules theorem" which is well-known in the theory of binary-state coherent systems.

Trellis coded modulation (TCM) is a new and attractive communication technique which yields significant coding gain with no degradation of the bandwidth efficiency. So far, researches concerning the TCM have concentrated upon minimizing the error-event probability (EEP). However, for the various bitwise data transmission systems, decreasing the post-decoding information-bit error rate (IBER) is more important than only minimizing the EEP. This paper is concerned with a simple but important problem that how to improve the IBER in TCM systems. The IBER of a TCM system depends not only on the EEP but also on the mapping of a binary m-tuple to a 2^m-ary channel signal. A signal mapping called Set Partitioning mapping has been well adopted in the previously known TCM systems. In this paper, another well known signal mapping, the Gray-code mapping, is investigated as a promising signal mapping in TCM, and compared with the Set Partitioning mapping. Through a computer search, a number of optimun trellis codes with Gray-code mapping are found. It is shown that these codes have the minimum Euclidean distances as large as those of the Ungerboeck codes and yet yield smaller EEP and IBER.

The accumulation of the thermal noise can be considered to be one of the most important parameters for the effective design of a multirelay transmission system. A theoretical model for prediction of the distribution of this parameter has recently been proposed. The method has assumed the log-normal form for the unconditional point rainfall rate distribution. In the present work an extension of the previous method is proposed, to include the case of the gamma distribution which is more applicable to Japan and other regions with similar climatic conditions. Numerical results are compared with experimental data taken from a 20 GHz multirelay system located in Japan.

This paper presents a method for analyzing a Markov process of the generalized loss system which includes the , /M₁, M₂/S, the , /, /S and the , /M₁, /S loss systems. The generalized loss system is a loss system with multi-class of batch arrival customers. In the system, a customer requires more than one server at the same time. The analyzing method makes use of the existence of a product form solution for a Markov process. This paper also presents a theorem for the , /M₁, M₂/S (PBAS) loss system. The theorem asserts that the steady-state probability of the total customers of the , /M₁, M₂/S (PBAS) loss system can be derived as that of the M ^x/M/S (PBAS) loss system with the same traffic intensity and batch size distribution. For the , /M₁, M₂/S (WBAS) loss system, two approximation analysis methods that assume the product form solution and use the theorem of the PBAS type system are propesed to obtain the blocking probability. The analysis for the generalized loss system and approximation analysis for the , , M₂/S (WBAS) loss system are applied to the design and dimensioning of resource sharing systems in ISDN.

This paper describes theoretical estimation of power efficiency improvement by adopting amplitude and phase modulation (APSK) with a honeycomb constellation (HC) instead of multilevel QAM modulation. Nonlinear distortion caused by a power amplifier is considered in the estimation, and nonlinearity of the amplifier is approximated by third and fifth order nonlinearity. To eliminated the difficulty in carrier reconstruction, a pilot carrier injection method is assumed for the APSK with the HC. However injecting the carrier reduces the power efficiency improvement, so dependence of power efficiency on the injected carrier level is estimated, theoretically. SNR of the recovered carrier as a function of the pilot carrier level is also estimated, experimentally. From these two estimations, an optimum pilot carrier level is determined for a 64 APSK system wit the HC. The possibility of reducing maximum available power of the amplifier by 2.0 dB is confirmed at the optimum pilot carrier level that corresponds to an offset of 1/4 data space in constellation. At the optimized level, SNR of the recovered carrier is 40 dB, which guarantees satisfactory operation of the system.

The function of optical wavelength selection is realized using a laser diode module which has an incorporated single mode fiber pigtail and an incorporated photodiode. One signal can be selected from two-wavelength multiplexed light signals with a channel spacing of 0.23 nm, and a crosstalk level of 30 dB. The laser diode in a module is tuned to the injection light, and the signal to noise ratio necessary for receiving 1 Gb/s signal is obtained when the received light power into fiber of the laser diode module is more than -35.9 dBm.

Mesh-connected computers (MCC's for short) are an important class of physically realizable parallel processors, since many scientific problems can be naturally mapped on them and because their regular structures and simple nearest-neighbour interconnections are particularly suitable for VLSI implementations. In former days iterative arrays and cellular automata were studied and recently special attention has been paid to the study of systolic arrays as a model of parallel computation on VLSI implemented MCC's. These abstract computational models constitute a family of MCC's. In this paper we study the effects of broadcasting bus systems augmented with a mesh-connected computer. First we develop a direct proof technique for the elimination of broadcasting buses. Then, as an application of the technique, we will show that a rich variety of broadcasting bus systems on one- and two-dimensional arrays can be eliminated without any loss of time efficiency. No-time-loss elimination of broadcasting buses on one-dimensional arrays has been shown by a different technique by Ibarra et al., but without our technique, it would be more difficult, but not impossible, to get the same results that we presented newly in this paper.

The difficulties of test generation and test application for VLSI chips have made the built-in self testing attractive. In the scheme, exhaustive or random test patterns are usually applied for logic blocks. However, for the testing of a multiple-output combinational circuit of which outputs depend only on partial input lines, the application of all input patterns may not be necessary in order to exercise the whole circuit. This paper proposes methods of evaluating the length of test sequences that can provide almost exhaustive test patterns to those output-wise subcircuits. Conventional linear feedback shift registers are used as the test pattern generator. Characteristics of pseudorandom sequence and truly random sequence as well are studied, in order to make the difference clear. the probabilities of achieving complete exhaustive testing are also estimated by simulation, showing the validity of the probabilistic model of LFSR sequence. It is shown that the probabilistic approach is effective compared with other deterministic ones.

A polynomial for automatic contour vector coding of the binary picture pattern is proposed. The proposed polynomial is the expansion of the Bezier polynomial to have the more flexibility. It is easy to be implemented and the generated curve is easy to be estimated from the given control points. This advantage is confirmed by the automatic contour coding of both English and Japanese fonts.