that the noise of each time series analysed (e.g., absence of periodicity, presence of instrumental errors and non- uniformity of measures) did not influence the TVA algo- rithm’s forecasting capability. Also, the TVA-algorithm forecast performance on filtered time series (i.e., purged of their casual components) was higher than 80%, as documented by MAPE measurements that gave good es- timates of the actual time series (Table 1). Consistent with other studies [11,14], MAs were here used as an easy and intuitive means, even though more sophisti- cated techniques, such as exponential and/or adaptive MAs, Kalman filters, Holt-Winter filters , would have to be employed to refine the forecast results. Thus, it is surprising that such a simple algorithm is capable of producing such good predictions, but this is possible because infectious episodes are outbreaks and, there- fore, are self-exciting processes which would be ex- pected to cluster at high values. How TVA algorithm performs as we go into the future it needs to be ex- plored. To strengthen our findings, ARMA models of order (2,3), (1,3), (1,1), (2,1), (1,2) and (1,1) (the figures indicate autoregressive and moving average terms) were constructed using the training set 36-month data to provide adequate model fit for monthly ESKAPE infec- tions due to E. faecium, S. aureus, K. pneumoniae, A. baumannii, P. aeruginosa and Enterobacter species, re- spectively. However, these models allowed forecast ac- curacies of 61.11% (E. faecium), 48.65% (S. aureus), 67.17% (K. pneumoniae), 73.02% (A. baumannii), 63.01% (P. aeruginosa) and 53.42% (Enterobacter species) (data not shown), that were much lower than those ob- tained using TVA algorithm (Table 1).
To improve the estimation accuracy in the presence of time-varying errors, adaptivefilters - have been widely used. In the adaptive filtering, time-varying process and measurement error covariance matrices are estimated based on residuals. The estimated covariance matrices are utilized for time propagations and measurement updates. Related to the adaptive filtering applied to GPS/INS integration, A. H. Mohamed  proposed an adaptive Kalman filter based on the maximum likelihood criterion for the proper choice of filter weights, W. Ding  proposed an online stochastic modeling algorithm, and A. Almagbile  compared the performances of the innovation and residual based adaptive methods.
The advantage of the proposed algorithm lies in the extremely fast computation of its response due to simplicity of the used filter. As mentioned in the part Computational complexity in "Results" section, the filter requires 6 additions (or subtractions, respec- tively) and 2 divisions only to compute one output signal sample. Extremely low com- putational demands together with the highest possible efficiency of baseline wander suppression regarding to instant heart rate favour the proposed filter against the other time-varying systems presented in “Background” section. One of the most advanced adaptive filter to suppress baseline wander was presented in . However, the used bank of low pass filters requires simultaneous computation of responses of many filters in order to deliver smooth output signal when switching between filters. Further, deci- mation and interpolation filters are never ideal and they are sources not only of higher phase delay but also of errors.
Abstract—In this paper, some new schemes are developed to im- prove the tracking performance for fast and rapidly time-varying systems. A generalized recursive least-squares (RLS) algorithm called the trend RLS (T-RLS) algorithm is derived which takes into account the effect of local and global trend variations of system parameters. A bank of adaptivefilters implemented with T-RLS algorithms are then used for tracking an arbitrarily fast varying system without knowing a priori the changing rates of system parameters. The optimal tracking performance is attained by Bayesian a posteriori combination of the multiple filter outputs, and the optimal number of parallel filters needed is determined by extended Akaike’s Information Criterion and Minimum Description Length information criteria. An RLS algorithm with modification of the system estimation covariance matrix is employed to track a time-varying system with rare but abrupt (jump) changes. A new online wavelet detector is designed for accurately identifying the changing locations and the branches of changing parameters. The optimal increments of the covariance matrix at the detected changing locations are also estimated. Thus, for a general time-varying system, the proposed methods can optimally track its slowly, fast and rapidly changing components simultaneously.
complexity. In this approach they have used selective coefficients update (SCU) approach to reduce the computational complexity. The long length adaptive filter is divided in to small sub filters for making it convenient to use in acoustic echo cancellation (AEC). As TD has fast convergence and SCU has low computational complexity therefore SCU and TD are combined in this new approach to decrease the performance losses. As the convergence speed increases miss-adjustment also increases. To remove this problem hybrid algorithm was implemented. Hybrid algorithm has fast convergence speed and better performance than standard TDLMS algorithm. It provides less computational complexity. The LMS algorithm and its variants have high computational complexity if incase its filter length is large. Fourier transform based block normalized LMS (FBNLMS) was introduced to reduce the computational complexity by using discrete fourier transform (DFT). But, FBNLMS still have high computational complexity therefore, Hartley transform based normalized LMS (HBNLMS) was implemented by Vasanthan Raghavan, K. M. M. Prabhu and Piet C. W. Sommen  in February 2005, by using cosine (DCT) and sine (DST) symmetric decomposition of discrete Hartley transform (DHT) that reduces the FBNLMS computational complexity by 33%. A new Fast block-exact LMS (FELMS) was proposed by Y. Zhou, S. C. Chan and K. L. Ho  in January 2006. It is performed by using the LMS/Newton algorithm whitened input and then applying shifting property. New block-exact fast LMS update is carried out in the same way as that of LMS. This proposed algorithm has less computational complexity but they are equivalent to fast LMS in terms of mathematical stability. FWTDLMS with partial sub-band coefficients update (FWTDLMS-PU) was proposed by Samir Attallah  in January 2006. Fast wavelet transform (WT) domain LMS (FWTDLMS) algorithm is exercised to make the proposed algorithm. Elen Macedo Lobato, Orlando José Tobias and Rui Seara  in May 2008, proposed a concept of stochastic modeling. This concept was applied to TDLMS algorithm. The proposed algorithm is independent of filter order and type of orthogonal transform. Shengkui Zhao, Zhihong Man, Suiyang Khoo and Hong Ren Wu  in January 2009, stated a new approach of applying second order autoregressive (AR) process on transform domain least mean square (LMS) adaptivefilters. By applying Power normalization and data independent orthogonal transform, convergence rate of adaptive filter is ameliorated.
ABSTRACT:This paper presents design, modeling and simulation of Unified power quality conditioner system to improve the power quality. Unified power quality conditioner consists of combined series and shunt active power filters for simultaneous compensation of voltage and current. The Unified power quality conditioner system is modeled using the elements of Simulink and it is simulated using MATLAB .A new synchronous-reference- frame based control method and d-q-0 theory is used to improve the power quality at the point of common coupling on power distribution systems under unbalanced and distorted load conditions. The results are analyzed and presented using MATLAB Simulink software.
Abstract - Medical images like MRI, CAT scan, and ultrasound play role of back bone in diagnostics. Images generated from various medical diagnostic machines are transmitted through channels before used by clinical experts. Transmission through noise channel sometimes degrades and corrupts the image making it inappropriate for accurate diagnose. The aim of this article is to study andanalyse the effect of noise on medical images. Different types of noises were intentionally added such as Gaussian, Salt and Pepper, Speckle and Poisson and corrupted the medical images with varying values of mean and variance .Different filtering algorithm: Anisotropic
In this paper, we evaluate the performance limitations of subband adaptive lters in terms of achievable nal error terms. The limitingfactors are the aliasing level in the subbands, which poses a distortion and thus presents a lower bound for the minimum mean squared error in each subband, and the distortion function of the over- all lter bank, which in a system identication setup restricts the accuracy of the equivalent fullband model. Using a generalized DFT modulated lter bank for the subband decomposition, both errors can be stated in terms of the underlying prototype lter. If a source model for coloured input signals is available, it is also possible to calculate the power spectral densities in both subbands and reconstructed fullband. The predicted limits of error quantities compare favourably with sim- ulations presented.
In , an exponential-window homotopy RLS-DCD adaptive filter possessing a high per- formance and low complexity was proposed. Here, we propose a sliding-window homotopy RLS-DCD algorithm and investigate it in application to estimation of sparse UWA chan- nels. The second proposed sliding-window RLS homotopy algorithm has the same structure as the exponential-window homotopy RLS-DCD algorithm. The proposed homotopy algo- rithm is used for channel estimation in an UWA communication system. In the transmitter of the system, the guard-free orthogonal frequency-division multiplexing (OFDM) signals with superimposed pilot signals are transmitted. We investigate and compare performance of the receiver with five adaptivefilters: NLMS adaptive filter; exponential-window and sliding-window classic RLS adaptivefilters ; the exponential-window homotopy RLS- DCD adaptive filter ; and the second proposed sliding-window homotopy RLS-DCD adaptive filter. The comparison is done using signals recorded on a 14-element vertical linear antenna array (VLA) in a sea trial at a distance of 105 km with a transducer moved at a speed of 6 m/s. The proposed adaptive filter provides an improved performance and error- free transmission at a spectral efficiency of 0.33 bps/Hz.
I consider three simulation scenarios in which the key qualitative features of the Great Moderation are replicated. In the …rst set of simulation, labeled as SM1, whereas monetary shock displays a double hump-shaped pattern. A double hump- shaped pattern of monetary shock has its ground in the growing literature on Markov switching DSGE models (see, for more information, Davig and Doh, 2009; Bianchi, 2010; Liu, Waggoner, and Zha, 2010). All other structural shocks exhibit a one time permanent reduction in the volatility. Many empirical studies document a discrete drop in conditional variance of macroeconomic time series. For example, Stock and Watson (2002) argue that changes in the volatility around 1984 are comprehensive and best characterized as discrete break. The second simulation scenario (henceforth, SM2) complexi…es volatility process one step further by allowing a subset of volatilities to have independent regimes. By doing so, I are able to examine the ability of each estimation method to recover the true volatility speci…cation. The last simulation scenario (henceforth, SM3) is based on the argument that the high volatility regime is often associated with recession (French and Sichel, 1993; Hamilton and Susmel, 1994). Since there are relatively fewer number of recessions after 1990s, this may appear as volatility reduction. I let high volatility regimes manifest themselves at the NBER recession dates.
In this paper, similar to , we also consider the problem of adaptive robust state observer design for a class of uncertain nonlinear time-delay systems. However, being different from , we assume that the time-varying delays are any nonnegative continuous and bounded functions which do not require that their derivatives have to be less than one. We present a new method whereby a class of continuous memoryless adaptive robust state observers with a rather simpler structure is proposed. We also show that the proposed adaptive robust state observers can guarantee that the observation error between the observer state estimate and the true state converges uniformly exponentially towards a ball which can be as small as desired. In addition, because the adaptive robust state observers proposed in the paper are completely independent of time delays, the results obtained here may be applicable to a class of dynamical systems with uncertain time delays.
For real-time implementation of this algorithm, there is a trade-off between the number of channels and the number of filter taps per channel; due to limited computational power. If the microphones happened to be in the wrong position with respect to each other, then a non-causal solution can be caused as the result. To prevent this problem from happening the delays are introduced in one of the input channels to provide physical reliability. This beamformer algorithm has been proven to improve the overall performance of the system by 5 dB for a competing speaker, by 8 dB for a wideband semi-stationary noise, and by 20 dB for narrowband interference (Van Compernolle, Van Gerven, Broos, & Weynants, 1991). The next section will briefly discuss some of the adaptive filter algorithms that can be used for this adaptive beamformer.
However, ARMA models are applied in cases were data show evidence of a stationary stochastic process. This means that the time series’ statistical properties are all constant over time. A stationary series has no trend. That is, its variations around its mean have a constant amplitude, and it wiggles in a consistent fashion, i.e., its autocorrelations remain con- stant over time. Equivalently, short-term random time pat- terns always look the same in a statistical sense. If the con- trary stands, its generalization, the Autoregressive Integrated Moving Average (ARIMA) model can be adopted. The ”Inte- grated” indicates that the data values have been replaced with the difference between their values and the previous values, to transform the time series to a stationary one. In a recent paper by Ord´o˜nez et al. (Ordez, Snchez Lasheras, Roca-Pardias, & Juez, 2019), the authors combine time series analysis meth- ods (ARIMA) to forecast the values of the predictor variables with machine learning techniques to predict RUL of aircraft engine for more than one period ahead, from those variables. Another important category of data-driven models used are graphical models, which denote the conditional independence structure between random variables (Chen Xiongzi et al., 2011). In a recent paper, (Banghart, Bian, Strawderman, & Babski-Reeves, 2017), Banghart et al. utilize Bayesian networks (BN) to estimate the risk of the landing gear sys- tem, cockpit warning/caution annunciator panel and the en- vironmental control system turbine assembly of the Northrop Grumman EA-6B Prowler military aircraft. BN is a proba- bilistic graphical model that represents a set of variables and their conditional dependencies via a directed acyclic graph (DAG). Nodes represent variables, while arcs represent prob- abilistic relationships. For example, engine blade damage im- pacts non mission-capable time, thus an edge/arc is drawn from the respective nodes. It is a combination of graph the- ory and probability theory. It is a representation of a joint probability distribution defined on a finite set of random vari- ables that can be discrete or continuous. From a knowledge modeling standpoint, Bayesian networks can be seen as a spe- cial knowledge representation system. The advantage of BN lies in the fact that it does not rely on explicit understand- ing of causal connections within the system(s) under obser- vation, nor identification of sequences of events leading to failure. Furthermore, given their probabilistic nature, BNs prove to be a suitable technique to address the inherent un- certainty of RUL estimation. In the same view, Ferreiro et al. in (Ferreiro & Arnaiz, 2010) use BN as a predicting tech- nique and demonstrate their effectiveness by representing a
We presented an approach for facial expression recog- nition that is based on fixed, directional filters and adaptivefilters connected in a hierarchical structure. The directional filters extract primitive facial features, whereas adaptivefilters are trained to extract more com- plex features, which are then classified by an SVM. The proposed system has a classification rate of 96.2%, which is higher than existing methods tested on the JAFFE database. Furthermore, by combining several SVMs and using the mirror image, the classification rate is improved to 96.7%. For future research, we plan to train the SVM in Stage 3 simultaneously with the adap- tive filters in Stage 2.
cancellation and demonstrate it using real time speech signals. The adaptive filter design discussed here is based on two algorithms RLS (Recursive Least Square) and LMS (Least Mean Square) and a comparison has been drawn based on their performance. Adaptivefilters find application because of their dynamic nature and they work on the principle of destructive interference.
For a linear system with pure input delay, Smith predictor is introduced in . However, if the open-loop system is unstable, the Smith predictor may fail to stabilize the overall system. The limitation on the open-loop stability required by the Smith predictor in the input delay compensation can be removed by the use of a new approach called the predictor feedback . The idea behind this approach is to apply the future state which can be estimated from the current state and the past control signals, in order to compensate for the input delay. A good feature of the mentioned method is that the closed-loop system has only a finite number of zeroes. Hence, this method is also known as the finite spectrum assignment . Linear systems with both input and state delays have been investigated, and a sliding mode control scheme to achieve stabilization has been presented in . A finite dimensional feedback control law that is truncated from the traditional predictor feedback proposed in , based on the low gain feedback structure . To tackle the problems of implementation of predictor feedback controllers for input delayed systems, truncated predictor feedback method was introduced for systems with delays in their input and states . Also, Smith predictor method is applied to systems with both state and input delays in . For the delay compensation, two auxiliary dynamic adaptivefilters with adjustable gains were included in the adaptive controller part. However, due to existence of these filters, the tracking error could not be minimized. The nested predictor is another method which has been presented for state and input delays compensation  for system stabilizing problems. The large scale systems with delay in interconnection terms were investigated in  with no input delay. Also, the control law proposed in  is very AUT J. Model. Simul., 50(1)(2018)3-12
An example of a two-stage SMART design is a study that characterized cognition in nonverbal children with autism . To improve verbal capacity, participants were initially randomized to receive either a combination of behavioral interventions (Joint Attention Symbolic Play Engagement and Regulation (JASPER) + Enhanced Milieu Training (EMT)) or an augmented intervention (JASPER + EMT+ speech-generating device [SGD]). Children were assessed for early response versus slow response to the first-stage treatment at the end of 12 weeks. The second-stage interventions, administered for an add- itional 12 weeks, were chosen on the basis of the response status (only slow responders to JASPER + EMT were re-randomized to intensified JASPER + EMT or received the augmented JASPER + EMT + SGD; slow responders to JASPER + EMT + SGD received intensified treatment; all early responders continued on the same intervention). There were three pre-fixed assessment time points: at 12 weeks, 24 weeks and 36 weeks (follow-up), which were the same for all participants in the study. Compared to multiple, one-stage-at-a-time, randomized trials, SMART designs provide better ability to compare the impact of a sequence of treatments, rather than exam- ining each piece individually. For example, a SMART allows us to detect possible delayed effects in which an intervention at a previous stage has an effect that is less likely to occur unless it is followed by a particular subse- quent intervention option. The typical modeling approach for the SMART design as described by Nahum-Shani et al. includes the indicators of intervention at each stage as covariates and thus accounts for the delayed effects on the final response. In order to develop a sequence of best decision rules for each individual, various statistical learn- ing methods of estimating the optimal dynamic treatment regimens have been proposed, among which Q-learning has been developed for assessing the relative quality of the intervention options and estimating the optimal (i.e., most effective) sequence of decision rules with linear regression. For a two-stage SMART, the Q-learning approach con- trols for the optimal second-stage intervention option when assessing the effect of the first-stage intervention, and reduces the potential bias resulting from unmeasured causes of both the tailored variables and the primary out- come. A similar approach for deriving the optimal deci- sion rules for SMART is A-learning, which is more robust