This paper is a concise overview of the main motivations behind the rise of periodic control theory, its achievements and some of its applications.

View full abstract

In this paper, we consider a subspace model identification for closed-loop systems based on stochastic realization approach. At first, by the joint input-output method the closed-loop identification problem is reformulated to an open-loop one, which can be solved by the recently proposed realization based subspace method ([3]). Then an effective model reduction step is adopted to obtain a plant model of the original system order from the estimated higher-order model, which is possibly unstable. We include two numerical examples to illustrate the effectiveness of the proposed method.

View full abstract

In this paper, we review the subspace identification method based on a preliminary orthogonal decomposition, which is shown to be more robust and reliable than N4SID like methods in critical cases [2, 5]. We then briefly describe a method of identifying a system operating in closed-loop based on the orthogonal decomposition method. Numerical results are included to show an improvement of performance in critical cases.

View full abstract

This paper presents the use of four continuous-time (CT) identification methods combined with Subspace-based State-Space System IDentification algorithms in order to identify continuous-time state-space models directly. These four CT methods cover the three classes of the CT methods: the linear filters, the modulating functions and the integral methods. Moreover, in order to point out the efficiency of the methods, a simulation study is carried out.

View full abstract

In this paper, we present the recursive subspace identification based on the subspace extraction via Schur complement. The subspace extraction via Schur complement is the alternative derivation of the estimate of the extended observability matrix in MIMO Output-Error State space model identification (MOESP) algorithm. A relationship between the least squares residual and the Schur complement matrix obtained from input-output data is shown, and the recursive implementation for the subspace-based state-space model identification (4SID) methods are presented. The proposed algorithm also apply to an instrumental variable (IV) based 4SID method.

View full abstract

In this paper, the mechanism is investigated how monetary policy is transmitted to the real economy in Japan The system is constructed in a form of VEC(Vector Error Correction) model and impulse responses of real GDP are calculated under situations such that the shock of call rate influences to real GDP with two channels of money and credit. Furthermore, the credit channel is classified by two kinds of bank loans: those to big enterprises(denoted by bloan) and those to small-to-medium-sized enterprises(denoted by sloan). Impulse responses of bank loans are also investigated.

View full abstract

The carrier phase ambiguity resolution is important for fast and high precision relative GPS positioning. In this paper, three ambiguity resolution techniques: Fast Ambiguity Resolution Approach (FARA), Least Squares Ambiguity Search Technique (LSAST) and Least squares AMBiguity Decorrelation Adjustment (LAMBDA), which can be applied to the single frequency data are reviewed from statistical view points. We compare the techniques by applying each to the single frequency double difference observations with some modifications. Also comparisons using the real receiver data are made.

View full abstract

Relationship between accuracy and factor selection of on-off decision making based on conditional probability was investigated. Criterion of the factor selection in the algorithm is introduced and the relationship between the criterion and accuracy of the decision making is investigated. Numerical example shows the usefulness and the effectiveness of the proposed method.

View full abstract

The authors previously proposed an on-line length measurement system for straight pipes using stationary waves which were generated in the pipes. The system modeled several modes of the stationary wave as state variables of a linear dynamic system and realized an on-line high-accurate pipe length measurement. This paper presents length measurement systems for curved and branch pipes by extending the pre-proposed measurement system for straight pipes. Experiments will show that an accurate pipe length measurement is realized with the methods.

View full abstract

From the viewpoint of stochastic inverse problem, identification of transfer functions in feedback systems has been examined by the previous papers[1][2]. In this paper it is applied for a new approach to electroencephalography (EEG). Some transfer functions between measurement regions of EEG are identified by our method of inverse problem in a stochastic feedback system. In our method for stochastic inverse problem an innovation model must be self-consistent: Since an innovation model equivalent to correlation functions has minimum phase and suitable properties from the theoretical considerations[3], innovation models with their properties are selected for dynamic analysis of EEG. Hence, a stochastic feedback model in the stationary process can be determined from EEG time series data of measurement regions, and we can obtain transfer functions between measurement regions.

View full abstract

Least-squares (LS) method is the most widely used method for the parameter estimation. However when applying directly the LS method to estimate parameters in presence of correlated noise, the LS method will be an asymptotically biased estimation method and be unable to give consistent estimates. There have been many studies to solve the problem of bias and one of them is bias-eliminated least-squares (BELS) method. In this paper, we examine the property of the BELS method and discuss the equivalence of BELS method and IV method under certain condition.

View full abstract

In this paper, a scheme to identification and control is proposed for main-steam temperature raising in a super heater system of thermal power plant. When modeling the system, a quasi-ARMAX model[1] is effectively used by considering an appropriate treatment of external inputs in the super heater system. Then the validity of constructed model is verified through identification tests. In particular, it is found that the model validity becomes more evident for the high nonlinear case. Finally, an adaptive control system for main-steam temperature raising is designed based on a nonlinear predictor model in the framework of linear stochastic control theory.

View full abstract

To solve a scientific problem by using a computer system, it is necessary to integrate various elements of computational technologies such as numerical simulation, database, network technology, human interface, and so on.

We have studied and developed Web Computing Aid Tools (WEBCAT) in order to simulate the scientific problems. This problem-solving environment (PSE) was designed based on the Web technology. WEBCAT functions as a Web application-specific server. Entering the parameter from the Web client to generate a source program in high-level language for a numerical simulation. Program resources are generated and compiled by clicking the images on the navigator homepage. The virtual experiment system is composed of a network-linked Web server that consists of the integration of various technologies. This system is able to extend a function by linking the equipment on the network. The data I/O system is equipped with a Virtual Reality device. VR device can be utilized for the 3-dimensional stereo-display of the data, which is transmitted via the visualization server. In addition, this system can generate the program, which analyzes the data was transferred from the Web device on the network. This report summarizes the overview of the system.

View full abstract

Many space and/or time patterns are observed in the various fields of enginering. For example, the spiral and target patterns in Belousov-Zhabotinski(BZ) reaction and the orderly convection pattern like scrolls or polygons in the Rayley-Benard convection are well-known as the typical examples of the pattern formations. Up to date, the excellent works of the pattern formations have been performed, however, models in the pattern formations were considered in the deterministic framework. In a realistic situation, the disturbance exists in greater or lesser degree. So, in this paper, taking the disturbance into consideration, we study the stochastic modeling of BZ reaction and the influence of the disturbance on the pattern formations through the digital simulation experiments.

View full abstract

A representation of fractal patterns is presented for coding complex random patterns in noisy imagery. Target patterns are assumed to be generated as the attractors of not-yet-identified contraction mappings.For capturing fractal attractor exactly Gaussian array is introduced on continuous image model. By local analysis, most probable points of unknown attractor are extracted within the framework of entropy maximization. In this continuous-discrete representation, background noise is shown to be eliminated via input-and output-filitering. Filtered discrete image is demonstrated to yield computable information enough to design a set of reduced affine mappings to regenerate observed attractors.

View full abstract

In this paper, we propose the visualization system for speech signals as well as for sound, we name this system as the KanNon system with meaning the system of seeing sound in Japanese, and this system can be useful for handicapped persons in hearing, as the training tool. As the method of seeing sound, we apply the sound spectrogram called as a voiceprint, which is a three-dimensional representation of spectral intensity of sound, in different frequency band, and over time is portrayed, and we show the sound spectrogram applying the Burg method and display common properties of Japanese utterance by different speakers. Finally, we describe the composition of the KanNon system that realize a real-time processor of speech signals.

View full abstract

An improved algorithm for tracking a maneuvering target with variable course and speed is established by introducing the jerk model to the target's dynamics. The effectiveness of the algorithm is shown by simulations and experiments using real data, comparing with the one which was previously proposed by the authors.

View full abstract

We consider a sensor allocation problem for the Kalman-Bucy filter within an information theoretic framework. For the signal and the observation of Kalman-Bucy filter, the mutual information between them is determined by the power of the drift-term in the innovations process and we cannot make the mutual information larger without increasing the power of the innovations process. Under a constraint on the mean square power of the drift-term in the innovations process, a set of equations is derived to compute the optimal gain matrix for the sensors which minimizes the least-squares estimation error.

View full abstract

State estimation of dynamical systems is one of the most important problems in control systems engineering. This paper deals with a stochastic approximation type nonlinear filter for state estimation of nonlinear stochastic dynamical systems. Though sufficient conditions for the convergence of this nonlinear filter has been provided, they are still restrictive. We relax the restriction on the observation systems by introducing the idea of randomly varying truncations. some modifications on the algorithm is also provided to accelerate the convergence.

View full abstract

In this paper we consider an H_∞ Kalman filtering problem for linear weakly coupled stochastic systems composed of N subsystems. The H_∞ Kalman Filter is a robust optimal filter which eliminates the effects of a bounded noise and a variation of system parameters. To solve the Kalman filtering problem we use the Hamiltonian approach, which introduces the transformation for the exact closed-loop decomposition of the H_∞ Kalman filtering Hamiltonian system. Using the proposed method, we can obtain N reduced-order Hamiltonian subsystems. In addition, we obtained the exact solution of the algebraic Riccati equations in terms of the solutions of the corresponding reduced-order subsystem algebraic Riccati equations. The introduced transformation produces a lot of savings especially for on-line computation since it allows parallel processing of information with lower-order Kalman filters.

View full abstract

The Genetic Algorithm(GA) with using previously proposed Edge Assembly Crossover(EAX) have shown a good performance for Traveling Salesman Problems(TSPs). Many crossovers for TSPs have been proposed so far. We first examine advantages of EAX by comparing it with other crossovers, and identify some properties to obtain a good performance. So previously proposed EAX was not refined well, the performance of GA using EAX can be improved by some simple modification. In this paper we give the modified version of EAX. Appling modified GA using EAX to well-known benchmarks in TSPLIB95, modified GA succeeded in finding optimal solution for almost problems up to 4461 cities.

View full abstract

We propose a new population-based Evolutionary Algorithm, which uses real-coded representation and normal distribution crossover-like mutation for generating next searching points. This Gaussian distribution is formed based on the position relationships between an individual and its neighbors, and is not carried with self-adapting parameters as inheritable traits. This algorithm causes emergence of clusters of individuals within the population, as the result of evolutions of each individuals without intent to cluster. Through searching independently, that emergent clusters introduce various solutions that include optimum at the same time, even if the problem has strong local minima.

View full abstract

This paper discusses the Genetic Algorithms for Noisy Fitness functions (GANF), that is, the Genetic Algorithms (GAs) for optimization of fitness function having random fluctuation. First, promising applications of the GANF are examined. Considering practical applicability of the GANF, requirements for the GANF are clarified. Several methods of the GANF proposed so far are overviewed, and the genetic algorithm with memory-based fitness evaluation (MFEGA) proposed by Sano and the author is described more in detail. Further, possible extensions of the GANF are also discussed.

View full abstract

In this paper, we adopt a genetic-based machine learning (GBML) approach to realtime assignment problems, e.g., a container assignment problem, and propose a method of generating and selecting rules for assigning each container to a desirable position in a container yard. In applying the GBML, we use the Pitts approach, where the set of rules (rule-set) is represented symbolically as an individual of genetic algorithms, and the fitness of an individual is calculated based on the sum of time required for replacement to fetch each container in a prescribed order. We actually carried out some computational experiments, which indicate that the method of applying the GBML is effective for finding good rule-sets.

View full abstract

Most of software availability models so far have paid attention to whether or not the software system is available only at a given time point. This paper discusses software availability measurement considering continuous use. In particular, the following two measures are derived: the interval software reliability defined as the probability that the system is continuously available for a given time period, and the conditional mean available time defined as the mean operable time measured from a time point at which the system is operating. The stochastic behavior of the system alternating between up and down states is described by a Markov process. Then, the software reliability growth process and the upward tendency of difficulty in debugging are also incorporated in the model. Finally, we show numerical examples of software availability analysis.

View full abstract

Many software systems have been produced under host-concentrated development environment. However, at present, the software development environment has been changing into distributed development environment because of the progress of network computing technology[1, 4]. Software systems produced in such distributed one tend to increase in size and complexity. Many software reliability growth models (SRGM's) based on a nonhomogeneous Poisson process (NHPP) have been proposed by many researchers. Most of all SRGM's which have been proposed up to the present treat the event of fault-detection in the testing and operational phase of software development as a counting process. However, if the size of the software system is large, the number of faults detected during the testing phase becomes large, and the change of the number of faults which are detected and removed through each debugging becomes sufficiently small compared with the initial fault content at the beginning of the testing. Therefore, as mentioned above, we can use a model based on a stochastic process with a continuous state space in such distributed development environment. In this paper, we propose an SRGM describing a fault-detection process during the system testing phase of the distributed development environment by applying a mathematical technique of stochastic differential equations of Itô type[5, 6]. We also derive the maximum likelihood estimators of the unknown parameters of the model. Furthermore, we compare our model based on stochastic differential equations with the existing SRGM's in terms of goodness-of-fit for actual data sets.

View full abstract

In this paper we discuss the problem of allocation of testing-effort. We consider two types of tests. The first is a module test and the second is an integration test. The integration test will be conducted successively after the module test. The amount of testing-effort is limited because of its great expense. We will formulate our model and will develop an efficient algorithm for it. Finally a numercal example will be given.

View full abstract

In this paper, a preventive replacement policy considering warranty is proposed. Throughout the warranty period, a manufacturer copes with the first failure of the objective system by replacing it with a new one free of charge, but he conducts minimal repairs to the succeeding failures without charge. An expected cost per unit time is formulated, when a user preventively replaces the system at time T from the initial purchase. We clarify the sufficient conditions under which a unique finite preventive replacement time that minimizes the expected cost exists. Numerical examples assuming a personal computer are also presented.

View full abstract

We deal with a situation where a worker processes two kinds of jobs. Job A (JA) can be processed only at a discrete time queueing system. He/She must join the queue when he decides to process JA and stays there until he completes JA. Job B (JB) is completed after processing several steps and each step needs a

unit time. At the end of each step, he observes the queue length and decides whether he joins the queue. If JB is completed, he joins the queue to process JA. If he decides to join the queue, he processes the rest of JB after JA. There is a time limit to process both jobs and the objective is to maximize the probability to finish the two jobs within the time limit. We prove a monotone property of the optimal policy by a dynamic programming formulation.

View full abstract

In this paper, we characterize the idle-period distribution for the GI/G/1 queueing system with vacation. Based on the properties of the idle-period distribution, we develop the upper bounds of the waiting time and the approximation formulae of the idle-period distribution.

View full abstract

In order to discuss the relationship between the lower and higher frequency components of source spectra, we deal with impulse train model as source time function, because spectral characteristics of source time function depend on occurrence times of impulse function which corresponds to small extent on the fault. Considering the successive rupture of small extent on a fault plane, we represent the source time function by time series which consist of random sequence of impulses. Under this assumption, the spectral characteristics of source time function are obtained analytically and numerically from the stochastic viewpoints: namely, on one hand, the trend of impulse train determines the frequency characteristics of phase spectrum in lower frequency range, and on the other hand, the fluctuation from the trend settles higher frequency range. Furthermore, it is shown that the spectral properties of source time function can be determined using only two parameters which are number of impulses N and probability density function (PDF) of occurrence time of impulse f_T(t).

View full abstract

Nowadays, owing to the widely applied electronic technology equipments, we are always exposed to electromagnetic wave radiated from them. At the same time, we are affected by the sound noise and light wave and so these should be evaluated totally. In such environment, it is necessary to pay attention to the fact that some accumulative and multiplicative effects are caused by them both physically and psychologically. For this purpose, first, it is important to know the correlation information among them as deeply as possible. However, since originally these fluctuate respectively in a non-Gaussian distribution form, it is difficult to find this mutual correlation by employing only the usual standard type correlation analysis in terms of linear and/or nonlinear type regression functions. In this paper, a new trial of hierarchically extracting such mutual correlation is proposed based on the conditional probability density function form derived from a statistical type Hermite series expansion expression of two-variate joint probability density function. The proposed method has been confirmed by applying it to the electromagnetic, sound and light leaked from VDT.

View full abstract

In the measurement of actual sound environment, the observed data show various types of probability distribution forms, and often contain fuzziness due to the existence of confidence limitations in measuring instruments, permissible errors in experimental data, and quantized errors in digitized observation. In this study, by introducing the well-known fuzzy theory, a state estimation method based on the above fuzzy observations is theoretically proposed from both the static and dynamic viewpoints. The validity and the effectiveness of the proposed method are confirmed experimentally by applying it to the evaluation of road traffic noise.

View full abstract

In this research, co-operative control by use of synchronous positioning control technique is proposed to realize co-operation of robot arms. The proposed method is expansion of the previous proposed contour control by using synchronous positioning control for synchronization of two links of one robot arm. Precise co-operative control of performer MK2 and performer MK3 robot arms are realized by use of the proposed method.

View full abstract

We have constructed a failure diagnosis system using a quick identification method named as SANQ system identification [1], and verified the effectiveness and performance under the deterministic situation. However when the application of this method is considered, it is necessary to investigate the performance under the stochastic situation caused by, e.g. disturbances and measurement noises. In this paper, we will verify characteristics and tolerance of the SANQ method under a stochastic situation, through simulation studies for a ship propulsion system. And we will try to modify the algorithm of the SANQ method so as to improve its performance under such a situation.

View full abstract

For the problem of consistently estimating the transfer function model in the presence of colored output noise, a three-step estimation procedure has been previously developed. This procedure exploits the input/output correlation information with respect to the correlation time in the first and second steps by setting the higher model order than the actual one. The bias-compensated least squares (BCLS) estimates are finally obtained by applying the compressive parameter transformation with the actual model order in the third step. This paper extends the above procedure to the case of colored input noise. As a result, it is shown that the extended procedure yields the BCLS estimates as well as in the case of the output noise. The effectiveness of the procedure is demonstrated by a simulation study.

View full abstract

A new adaptive identification approach is proposed for a time-varying FIR system in which the system parameters change as piecewise polynomial function of time. By appropriately incorporating the internal model corresponding to the polynomial degree into the parameter adjustment dynamics of the new algorithm, its BIBO stability can be assured by the aid of the small gain theorem and modified passivity theorem. Its effectiveness of the proposed LMS and RLS types of algorithms is validated by comparing to ordinary approaches in adaptive direction-of-arrival (DOA) estimation for fast moving vehicles.

View full abstract

In this paper, an identification problem of uncertain parameters which are involved in the mathematical model is investigated for a class of stochastic uncertain systems. First, it is shown that the parameter identification can be achieved by solving the state estimation and the least-squares problem simultaneously. Secondly, a modified Gauss-Newton method is proposed, showing that it has contraction property. Finally, a numerical example is shown to demonstrate the identification method proposed here for a process modeled by a first-order system with time-delay.

View full abstract

In this paper, we present a filtering model on a default risk related to mathematical finance. We regard the first hitting time at zero of a one dimensional process which starts at some positive number and is not directly observed as the time when a default occurs. We discuss the conditional law of the hitting time under imperfect information. We use the reference measure change technique and some formula on a kind of conditional expectation to obtain a so-called hazard rate process. It is also discussed what the relation between the hazard rate process and the conditional law of the hitting time is like.

View full abstract

In the context of financial engneering, practitioners tend to price almost every option by Black-Sholes formula or some modified ones. This means they consider f(S¹_T,…,S^m_T) to be somehow log-normally distirbuted, regardless of f. The aim of this talk is to explain this by some approximating procedure. Moreover, we will discuss the quasi pricing of compound options. Not only we give an explicit formula dor a compound option, we also introduce a concept of stationarity, which are given by the form of a integral equation. This latter problems remais unsolved.

View full abstract

In this paper, the image identification and restoration algorithm using only observable noisy blurred image based on the so-called semi-causal models is discussed. Our image identification is performed by applying the EM algorithm which is carried out by using the Kalman smoother estimates effectively. Also, we show compensation algorithms due to the approximation of Toeplitz matrices into circulant matrices in our image models.

View full abstract

This paper deals with the blind deconvolution problem of MIMO-FIR channels driven by source signals which are temporally second-order uncorrelated but fourth-order correlated and spatially second- and fourth-order uncorrelated. Recently, we proposed a criterion for solving the blind deconvolution problem [8], [9]. However, we have not yet shown an algorithm to implement our proposed criterion. One of our objectives in the present paper is to propose an iterative algorithm for achieving the blind deconvolution by using this criterion. To achieve this objective, we consider an approach called the non-deflation method. We propose an iterative algorithm based on the method. Moreover, we present a deflation algorithm using this criterion for the blind separation of MIMO constant channels.

View full abstract

We propose the method to reduce the redundancy between features extracted from images by the infomax algorithm. The method consisted of the addition of noise and the feature selection based on kurtosis. The addition of noise to images contributes to the reduction of redundancy and make it easy to set a criterion of features selection. We adopt ETL-4 database sets of hand-written Japanese Hiragana characters, and performed the recognition experiment with the extracted features. Results of the recognition experiment showed that it was possible to reduce the redundancy of features by the proposed method.

View full abstract

The predictor-based least squares algorithm (5) derived from the relation of the backward predictors as the fast RLS (FRLS) algorithm [1] gives the exact solution of the least squares problem as the RLS, the FRLS algorithms do. One of its advantages is the possibility of reducing its complexity to O(MN) where N is the length of the filter if the input signal is autoregressive of order M(≪ N) by truncating the prediction filters in the PLS algorithm. This algorithm is named the fast PLS (FPLS) algorithm [7] and is proven to be equivalent to the fast Newton transversal filter (FNTF) algorithm [4]. Since the FPLS and the FNTF algorithms do not rigorously solve the LS problem, they have a different convergence rate from the LS algorithms such as the RLS, the FRLS and the PLS algorithms. This paper theoretically gives the convergence rate, which is larger (slower) than that of the LS algorithms.

View full abstract

Several properties of the filtered-error (filtered-E or adjoint) least mean square (LMS) algorithm, such as the stability condition, the upper limit of the step size and the variance of the error signal are quantitatively evaluated and compared with the those of the filtered-X LMS algorithm. For this purpose, the averaged system and the ordinary differential equation (ODE) method are applied to the frequency domain expression of both algorithms. From the averaged system, it is demonstrated that the stability condition and the upper limit of the step size of the filtered-E LMS algorithm are same with those of the filtered-X LMS algorithm. On the other hand, from the ODE analysis, it is shown that the excess mean square error of the filtered-E LMS algorithm is two thirds of that of the filtered-X LMS algorithm.

View full abstract

This paper describes an early attempt to build a new multi-scale stochastic model on multiwavelet. Previously such stochastic model are defined either on a single or multiple binary tree as well as on an ordinary wavelet structure. Several interesting results are obtained due to the specific properties of multiwavelet which are different from its wavelet counterpart. Smoothing of a certain fractal signal on the proposed model will be given.

View full abstract

We study the splitting-up method for solving the Zakai equation for the diffusions with noise correlation. Using this approach, the numerical integration is decomposed into a stochastic step and deterministic one. In the stochastic step, we need to solve a first order hyperbolic equation. By using the method of a stochastic characteristic curve, the explicit form of the solution can be easily obtained. This approach is applied to solve the stochastic volatility estimation problem which arises in the mathematical finance filed.

View full abstract

This paper is concerned with the stochastic modeling of phase transition phenomena and its behavior analysis by numerical simulations. First, classification of phase transition phenomena is explained. Phase transition phenomena are often observed in various kinds of fields of engineering, including chemical, biological and fluid engineering. To construct precise models of phase transitions and their analysis are one of very important problems in engineering. Usually, phase transition phenomena are modeled by the order parameter, which is classified into the conserved and nonconserved type. In this paper, using two types of the order parameter, we propose models of two kinds of phase transitions. One is the model of phase separation phenomena and the other is the model of the order-disorder transitions. Thermal fluctuation and disturbance such as impurities in chemical substances play an important role in phase transitions, so in this paper, we propose stochastic models of phase transitions and analyze behavior of phase transitions through simulation experiments.

View full abstract

The author has proposed a kind of fuzzy random vectors(FRVCs in short) which may be appropriate for modeling the data with vagueness and randomness simultaneously[1]. Those FRVCs are considered as the vague perceptions of random phenomena. The purpose of this paper is to propose another type of FRVCs which are not vague perceptions of random phenomena but vague themselves. This another type of FRVCs may be appropriate for modeling human feelings, economic conditions and so on, whose typical properties are that they are obtained as the synthetical verbal expressions of complex random phenomena. First, using the set representation approach, the reasonable definition of fuzzy random variables(FRVCs) is proposed. Secondly, The statistical moments for the proposed FRVCs and their estimates are investigated. Finally, some of estimates are proposed for the statistical moments of FRVCs.

View full abstract

Fractional Brownian motion (fBm) provides a useful fractal model to represent natural landscapes. The fractal dimension of an fBm is closely related to its spectral exponent, which is the essential parameter of the power spectral density (PSD) of fBm. In this paper, a maximum likelihood (ML) spectral estimation procedure for an fBm defined on a multi-dimensional space is derived. The likelihood function is shown to be represented in terms of a periodogram and the PSD model, instead of a covariance matrix. This representation facilitates the ML estimation in the multi-dimensional case. In an application of this, we estimated the spectral exponent of the moon surface and constructed a detailed simulated moon surface.

View full abstract