characterisation results are discussed and both the asymptotic theory and small sample power of the tests are investigated. In chapter 6, we propose a general non-parametric model for the regularly varying t a i l s of a distribution and consider the problem of estimating the parameters of the proposed model. Estimators based on the empirical characteristic function are constructed and are shown to be inferior to other estimators based on the extreme order s t a t i s t i c s . These l a t t e r estimators are shown to attain optimum rates of convergence. Two approaches to adaptiveestimation are considered. Only one of these approaches is appropriate and we conclude the thesis with a brief investigation of the small sample properties of this technique. Section 5.4 and chapter 6 are the product of jo in t work with
Le Cam (1960), Swensen (1985) and Roussas (1979) give conditions under which the log-likelihood ratio of a general stochastic process satisfies the LAN condition [Linton (1993)]; these conditions have been verified for stationary ARMA processes by Kreiss (1987), and DKW (1997) generalized the results of Kreiss (1987) to classes of non-linear time-series (location-scale) models [Wellner et al. (2006)]. In this paper, following DKW (1997), we show that the LAN property holds in the case of the reparametrized PTTGARCH(1,1) model as the reparametrized PTTGARCH(1,1) model is a non-linear time-series (location-scale) model that satisfies conditions in DKW (1997). Following DKW (1997), we also give the relevant assumptions such that semiparametric efficient adaptiveestimation is possible for π = (α + , α − , β).
The class of quantum states known as Werner states have several interesting properties, which often serve to illuminate unusual properties of quantum information. Closely related to these states are the Holevo-Werner channels whose Choi matrices are Werner states. Exploiting the fact that these channels are teleportation covariant, and therefore simulable by teleportation, we compute the ultimate precision in the adaptiveestimation of their channel-defining parameter. Similarly, we bound the minimum error probability affecting the adaptive discrimination of any two of these channels. In this case, we prove an analytical formula for the quantum Chernoff bound which also has a direct counterpart for the class of depolarizing channels. Our work exploits previous methods established in [Pirandola and Lupo, PRL 118, 100502 (2017)] to set the metrological limits associated with this interesting class of quantum channels at any finite dimension.
Money supply usually shows heteroscedastic pattern. But for most of the practical situations, we seldom know any thing about the form of heteroscedasticity then it becomes a question to correct the analysis for heteroscedasticity. When the form of heteroscedasticity is unknown, we may use adaptive estimators that give desirable results that are not possible when just OLS estimation is taken into account. Usually nonparametric approaches are used for adaptiveestimation. In these approaches nearest neighbour regression estimators give better results as compared to adaptive kernel estimators in the presence of heteroscedasticity of unknown form.
We propose a state-by-state selection method to infer the emission densities of a HMM. Using a family of estimators, our method selects one estimator for each hidden state in a way that is adaptive with respect to this state’s regularity. This method does not depend on the type of preliminary estimator, as long as a suitable variance bound is available. As such, it may be seen as a plug-in that takes a family of estimators and the corresponding variance bound and outputs the selected estimator. Note that its complexity does not depend on the number of observations used to compute the estimators, which makes it applicable to arbitrarily large data sets.
Hybrid systems have been identified as one of the main directions in control theory and attracted increasing attention in recent years due to their huge diversity of engineering applications. Multiple-model (MM) estimation is the state-of-the-art approach to many hybrid estimation problems. Existing MM methods with fixed structure usually perform well for problems that can be handled by a small set of models. However, their performance is limited when the required number of models to achieve a satisfactory accuracy is large due to time evolution of the true mode over a large continuous space. In this research, variable-structure multiple model (VSMM) estimation was investigated, further developed and evaluated. A fundamental solution for on-line adaptation of model sets was developed as well as several VSMM algorithms. These algorithms have been successfully applied to the fields of fault detection and identification as well as target tracking in this thesis. In particular, an integrated framework to detect, identify and estimate failures is developed based on the VSMM. It can handle sequential failures and multiple failures by sensors or actuators.
In a modern machining system, tool wear monitoring systems are needed to get higher quality production. In precision machining processes, especially surface quality of the manufactured part can be related to tool wear. This increases industrial interest for in-process tool wear monitoring systems. For the modern unmanned manufacturing process, an integrated system composed of sensors, signal processing interface and intelligent decision making model are required. In this study, a new method for on-line tool wear monitoring is presented under varying cutting conditions. The proposed method uses wear feature extraction based on process modeling and parameter estimation. An adaptiveestimation model of milling tool wear in variable cutting parameters is built based entirely on milling power. The adaptive model traces the properties of cutting process by combining process state signal, cutting conditions, power model. The tool wear feature is obtained from the estimated parameters of the model and carried on in the theoretical and experimental study. Experiment results have proved that changes of the parameters in the cutting power model significantly indicate tool wear independently of varying cutting conditions and it makes tool wear a recognized process with high precision.
In order to solve the problem, this paper presents an adaptive filtering algorithm based on truncated normal probability density model (abbreviated as TGPMMKF). This algorithm combines the ideas of the literature . When the maneuvering target is maneuvering with a certain acceleration, the acceleration value at the next moment is limited, and can only be in the “current” acceleration neighborhood. Base on this nature and accord to the maneuverability of acceleration components change between adjacent samples identifies targets online to achieve the online self-adaptiveestimation of the variance of the process noise, which avoids the adverse effect of the presetting of the acceleration limit value in the truncated normal probability density model on the maneuvering target state estimation accuracy. Computer simulations show that the algorithm has good tracking ability for strong maneuvering targets and has the advantage of real-time performance.
III. T RAINING S EQUENCE B ASED C HANNEL E STIMATION Based on the assumptions such as perfect synchronization and block fading, a MIMO-OFDM system is design. In training based channel estimation algorithms, training symbols or pilot tones that are known to the receiver, are multiplexed along with the data stream for channel estimation. The idea behind these methods is to develop knowledge of transmitted pilot symbols at the receiver to estimate the channel. For a block fading channel, where the channel is constant over a few OFDM symbols, the pilots are transmitted on all subcarriers in periodic intervals of OFDM blocks. The channel estimates from the pilot subcarriers are interpolated to estimate the channel at the data subcarrier. This type of pilot arrangement, given in Fig.3 is called the block type arrangement.
generalized drop-the-loser urn and generalized Fried- man’s urn design is studied for both continuous and dis- continuous outcomes [11,16,17,36]. It is shown that, under reasonable assumption about the delay, the asymptotic properties of adaptive design are not affected by the delay. In the present study, the primary focus is the comparison between commonly used test statistics for 2 × 2 tables. Based on results not shown here, a less extreme allocation with higher variation would be expected when a random delay is assumed. It is assumed that the response status of each of the patients already in the trial is available before the allocation of a new patient in our simulations evaluation.
Fault tolerance is gaining interest as a means to increase the reliability and availability of distributed energy system. In this project presents a new technique for fault detection in vector controlled induction motor (IM) drive. The proposed current estimation uses estimation uses d- and q-axes currents and is independent of the switching states of the three-leg inverter. While the technique introduces a new concept of vector rotation to generate potential estimates of the currents, speed is estimated by one of the available model reference adaptive system (MRAS) based formulations. The objective of the controller was to control the current that supply into the induction motor. The proposed method is extensively simulated in simulated in MATLAB/SIMULINK.
This work studies the problem of blind distributed esti- mation over an ad-hoc wireless sensor network (WSN). WSNs have recently generated a great deal of renewed interest in distributed computing. New research avenues have opened up in the fields of estimation and tracking of parameters of interest, in applications requiring a robust, scalable and low-cost solution. To appreciate such appli- cations, consider a set of N sensor nodes spread over a geographic area as shown in Figure 1. Sensor measure- ments are taken at each node at every time instant. The objective of the sensor is to estimate a certain unknown parameter of interest using these sensed measurements.
This paper presents adaptive link selection algorithms for distributed estimation and considers their application to wireless sensor networks and smart grids. In particular, exhaustive search-based least mean squares (LMS) / recursive least squares (RLS) link selection algorithms and sparsity-inspired LMS / RLS link selection algorithms that can exploit the topology of networks with poor-quality links are considered. The proposed link selection algorithms are then analyzed in terms of their stability, steady-state, and tracking performance and computational complexity. In comparison with the existing centralized or distributed estimation strategies, the key features of the proposed algorithms are as follows: (1) more accurate estimates and faster convergence speed can be obtained and (2) the network is equipped with the ability of link selection that can circumvent link failures and improve the estimation performance. The performance of the proposed algorithms for distributed estimation is illustrated via simulations in applications of wireless sensor networks and smart grids.
adaptively. A smoothed estimate of the state and observation noise can be respectively obtained according to steps 5 and 6 of Table 1(a), where α is a smoothing parameter and R(x) is the ramp function (R(x) = x if x ≥ 0 and 0 otherwise). These two methods for online estimation of the noise distur- bances is due to Jazwinski .
In this paper, we develop a general sequential algorithm for the M-estimate of a linear observation model. Our development is based on formulating the problem from a Bayesian perspective and using a Gaussian approximation for the posterior and likelihood function in each learning step. The sequential algorithm is then developed by determining a maximum a posteriori (MAP) estimate when a new set of training data is received. The Gaussian approximation leads naturally to a quadratic objective function and the MAP estimate is an RLS-type algorithm. We have discussed the quality of the estimate, issues related to the initialization and estimation of parameters, and the relationship of the proposed algorithm with those of previous work. Motivated by reducing computational cost of the RLS-type algorithm, we develop a family of LMS-type algorithms by replacing the covariance matrix with a scaled identity matrix. Instead of updating the covariance matrix, we update the scalar which is set to preserve the determinant of the covariance matrix. Simulation results show that the learning performance of the proposed algorithms is competitive to that of some recently published algorithms. In particular, the performance of proposed LMS-type algorithms has been shown to be very close to that of their respective RLS-type algorithms. Thus they can be replacements for RLS-type of algorithms at a relatively low computational cost.
The system in the period t, t = 0,1 … begins to function from the moment when Federal cargo company provides the input of the repair depot with operating influences such as plan and resource. Simultaneously the input is influenced by the environment in the form of a stochastic hindrance. The repair depot chooses an output. Federal cargo company observes the output, estimates it and extrapolates the estimate to the following period. Further, Federal cargo company defines resource and plan for the next period on the basis of this estimation using procedures of planning and regulating. On comparing of the actual output to the plan, the stimulus of the repair depot is defined. After that the system’s functioning in period t is over, and there comes period t+1, and so on.
Abstract In a wide variety of parametric uncertainties and external disturbance estimation and attenuation methods uncertainties and disturbance are lumped together and an observation algorithm is employed to estimate the total disturbance. While in certain cases of application can be required the separate estimation or identification of uncertain parameters itself and a disturbance. In the paper the separate and simultaneous identification (estimation) of unknown and/or changing parameters and an external disturbance of a linear object is considered. For this porpose the known procedure of synthesis of adaptive observer for estimation of parameters and state coordinates of a n-th order linear object is used taking into account the influence of the scalar external disturbance operating on this object. Developed adaptive observer provides asymptotic stability of processes of separate and simultaneous identification of uncertain parameters and an external disturbance of a n-th order linear object. Asymptotic stability of proposed observer is proved by Lyapunov’s direct method. As an example of using of the offered adaptive observer the structure and algorithm of joint identification of the moment of inertia (parameter) and the torque of resistance (external disturbance) of mechanical load of dc electric drive model are obtained. Asymptotic stability of processes of joint identification of the parameter and external disturbance of drive model is proved. Simulation results of identification processes and their using for control system adaptation are shown on graphs of transition processes. Designed algorithm for the joint identification of parameter and external disturbance of plant provide adaptive stabilization of desirable dynamic properties of control system with adaptive observer.
In this paper , we are interested to investigate the feasibility of using adaptive observer technique to realize monitoring of drilling operations in oil wells from the essential and critical operational problem views , being considered in the work. The resulting developed estimation and monitoring systems will be implemented and evaluated in simulation environment on the basis of accessible operational data from candidate oil wells.
Abstract —The remaining useful life (RUL) of transformer insulation paper is largely determined by the winding hot- spot temperature (HST). Frequently the HST is not directly monitored and it is inferred from other measurements. However, measurement errors affect prediction models and if uncertain variables are not taken into account this can lead to incorrect maintenance decisions. Additionally, ex- isting analytic models for HST calculation are not always accurate because they cannot generalize the properties of transformers operating in different contexts. In this con- text, this paper presents a novel transformer condition as- sessment approach integrating uncertainty modeling, data- driven forecasting models and model-based experimental models to increase the prediction accuracy and handle uncertainty. The proposed approach quantifies the effect of measurement errors on transformer RUL predictions and confirms that temperature and load measurement errors affect the RUL estimation. Forecasting results show that the extreme gradient boosting (XGB) algorithm best cap- tures the non-linearities of the thermal model and improves the prediction accuracy amongst a number of forecasting approaches. Accordingly, the XGB model is integrated with experimental models in a Particle Filtering framework to im- prove thermal modelling and RUL prediction tasks. Models are tested and validated using a real dataset from a power transformer operating in a nuclear power plant.