Filter-based Model Research Articles

Correction: Coppieters de Gibson, L.; Garner, P.N. Training a Filter-Based Model of the Cochlea in the Context of Pre-Trained Acoustic Models. Acoustics 2024, 6, 470–488

The authors would like to make the following corrections to the original publication [...]

Shoreline morphological change prognostic model based on spatiotemporal framework imagery data on the northern coast of Java, Indonesia

This study focuses on identifying the degree of shoreline changes on the north coast of Java Island (Indonesia), particularly the Batang Regency segment, using multiple data sources: Landsat 9 and 8, Landsat 5 TM, and PlanetScope spanning from 2000 to 2023. The degree of change was obtained by statistically calculating net shoreline movement (NSM), linear regression rate (LRR), and end-point rate (EPR) using the digital shoreline analysis system (DSAS) software. The three rate-of-change statistics were used to calculate the multidecadal shoreline change prediction or projection for the next 10 and 20 years using a Kalman filter-based model with a simple data assimilation technique. Based on these results, NSM indicated that the shoreline experienced a total accretion and erosion of 361.04 and 111.58 m, respectively, from 2000 to 2023. Over these years, the highest accretion rate reached 16.44 m/year (according to LRR) or 15.83 m/year (EPR), while the maximum erosion rate was up to 11.34 m/year (LRR) or 9.43 m/year (EPR). The use of two data sources with different spatial resolutions produces varying statistical values, which depend on the degree of detail and complexity of the data derived from the source imagery. The coastline prognosis model produced with the Kalman filter shows that the entire coastline in Batang Regency will most likely experience an average accretion and erosion of 18.03 and 21.42 m, respectively, in 10 years, while that in the next 20 years will be 26.38 and 33.3 m, respectively.

Training a Filter-Based Model of the Cochlea in the Context of Pre-Trained Acoustic Models

Auditory research aims in general to lead to understanding of physiological processes. By contrast, the state of the art in automatic speech processing (notably recognition) is dominated by large pre-trained models that are meant to be used as black-boxes. In this work, we integrate a physiologically plausible (albeit simple filter-based) model of the cochlea into a much larger pre-trained acoustic model for speech recognition. We show that the hybrid system can be trained and evaluated with various combinations of fine-tuning and self-supervision. The results broadly show that the system automatically yields structures that are known to work well. Moreover, these structures lack artifacts that were apparent in (our) previous work using less sophisticated neural models. We conclude that the hybrid structure is an appropriate way to proceed in auditory research, more generally allowing the work to take advantage of larger models and databases from which it would not otherwise benefit.

Extended State Kalman Filter-Based Model Predictive Control for Electro-Optical Tracking Systems with Disturbances: Design and Experimental Verification

A modified Extended State Kalman Filter (ESKF)-based Model Predictive Control (MPC) algorithm is introduced to tailor the enhanced disturbance suppression in electro-optical tracking systems. Traditional control techniques, although robust, often struggle in scenarios with concurrent internal, external disturbances, and sensor noise. The proposed algorithm effectively overcomes these limitations by precisely estimating system states and actively mitigating disturbances, thus significantly boosting noise and perturbation control resilience. The primary contributions of this study include the integration of ESKF for accurate system state and disturbance estimation in noisy environments, the embedding of an ESKF estimation-compensation loop to simulate an improved disturbance-free system, and a simplified modeling approach for the controlled device. This designed structure minimizes the reliance on extensive system identification, easing the predictive control model-based constraints. Moreover, the approach incorporates total disturbance estimation into the optimization problem, safeguarding against actuator damage and ensuring high tracking accuracy. Through rigorous simulations and experiments, the ESKF-based MPC has demonstrated enhanced model error tolerance and superior disturbance suppression capabilities. Comparative analyses under varying model parameters and external disturbances highlight its exceptional trajectory tracking performance, even in the presence of model uncertainties and external noise.

A real-time biphasic Kalman filter-based model for estimating human core temperature from heart rate measurements for application in the occupational field.

Early identification of hypothermia or hyperthermia is of vital importance, and real-time monitoring of core temperature (CT) of the workers exposed to thermal environments is an extremely valuable tool. From the existing literature studies, the model developed by Buller et al. in their study of 2013 that generates real-time estimates of CT from heart rate (HR) measurements using the Kalman filter (KF) shows good potential for occupational application. However, some aspects could be improved to reliably handle the existing very wide range of workers and work activities. This study presents a real-time CT estimation model, called the Biphasic Kalman filter-based (BKFB) model, based on HR measurement, with characteristics suited to application in the occupational field. Thirteen healthy subjects (six female and seven male) were included in the study to perform three consecutive tasks simulating work activities. During each test, an ingestible CT sensor was used to measure CT and a HR sensor to measure HR. The KF methodology was used to develop the BKFB model. An algorithm with a biphasic structure was developed using two different models for the increasing and decreasing phases of CT, with the ability to switch between the two based on an HR threshold. CT estimates were compared with CT measurements, and with respect to overall root mean square error (RMSE), the BKFB model achieved a sizeable reduction (0.28 ± 0.12°C) compared to the Buller et al. model (0.34 ± 0.16°C). The BKFB model introduced some modifications over the Buller et al. model for a more effective application in the occupational field. It was developed using data collected from a sample of workers (heavily weighted toward middle-aged, not very fit, and with a considerable fraction of female workers), and it also included two different modeling of CT (for the up- and down-phases), which allowed for better behavioral modeling in the two different stages. The BKFB model provides CT estimates reasonably in comparison to the measured intra-abdominal temperature values in both the activity and recovery phases but is more practical and easier to use for a real-time monitoring system of the workers' thermal states.

Kalman filter-based model predictive control for drag-free satellite under actuator failures and input saturations.

This paper studies the problem of fault-tolerant control for the drag-free satellite subject to actuator failures and input saturations. Particularly, a new Kalman filter-based model predictive control method for the drag-free satellite is proposed. Based on the developed dynamic model and the Kalman filter strategy, a new fault-tolerant design scheme is presented for the challenging situation where the satellite is subject to measurement noise and external disturbance. With the designed controller, the robustness of the system can be guaranteed, and the problem caused by actuator constraints and faults can be solved. Finally, the correctness and effectiveness of the proposed method are verified by numerical simulations.

Improved Particle Filter-Based Modeling With Robotic GMAW for Weld Forming Prediction

Improved Particle Filter-Based Modeling With Robotic GMAW for Weld Forming Prediction

Effect of turbulence modelling and non-condensable gas on cloud cavity dynamics

Turbulence modelling plays an important role in the prediction of unsteady and highly dynamic characteristics of cavitating flows. The models with turbulent viscosity modification as Density-Corrected Method (DCM) and Filter-Based Model (FBM), can deal with the dynamics of cavitating flows. Taking it into account, the application of these models to the cavitating flows of water with dissolved air (i.e. third phase) to predict the unsteady phenomena are worth to be analysed and compared. Due to poor self-oscillation characteristics of RNG k-epsilon model in modelling of the cavitating flow, in the present study, the DCM and FBM are implemented to perform the simulation of cavitating flow with consideration of dissolved air. Also, the numerical results are compared with the experimental data. The experiments are conducted in a rectangular test section equipped with Clark Y hydrofoil providing cavity visualisation, instantaneous pressure and vibration fluctuations. The simulations are carried out for different cavitation numbers with and without dissolved air. The Fast Fourier Transform and Continues Wavelet Transform are implemented to extract and compare the shedding frequency of experiments and numerical predictions. In addition, the influence of modification models on the cloud cavitation shedding evolution, dominant shedding frequency, vorticity, pressure, velocity profile, lift/drag coefficient is evaluated.

Prediction Model of Bus Arrival and Departure Times Using AVL and APC Data

The emphasis of this research effort was on using AVL and APC dynamic data to develop a bus travel time model capable of providing real-time information on bus arrival and departure times to passengers (via traveler information services) and to transit controllers for the application of proactive control strategies. The developed model is comprised of two Kalman filter algorithms for the prediction of running times and dwell times alternately in an integrated framework. The AVL and APC data used were obtained for a specific bus route in Downtown Toronto. The performance of the developed prediction model was tested using “hold out” data and other data from a microsimulation model representing different scenarios of bus operation along the investigated route using the VISSIM microsimulation software package. The Kalman filter-based model outperformed other conventional models in terms of accuracy, demonstrating the dynamic ability to update itself based on new data that reflected the changing characteristics of the transit-operating environment.A user-interactive system was developed to provide continuous information on the expected arrival and departure times of buses at downstream stops, hence the expected deviations from schedule. The system enables the user to assess in real time transit stop-based control actions to avoid such deviations before their occurrence, hence allowing for proactive control, as opposed to the traditional reactive control, which attempts to recover the schedule after deviations occur.

A modified Kalman filter-based model for core temperature estimation during exercise and recovery with/without personal cooling interventions

One of the most important parameters to consider while analysing heat related illness is the human body core temperature. It is therefore required to monitor real-time core temperature in an ambulatory situation for industrial workers, athletes, military personnel, and a variety of other people. As conventional methods of monitoring core temperature are unsuitable for ambulatory use or prohibitively expensive, it is necessary to develop a reliable, inexpensive and non-invasive method for real-time measurement of core temperature. Kalman filter based models for core temperature estimation are found promising to meet these requirements. In this work, we have highlighted one of the major shortcomings of the existing extended Kalman filter model and proposed a modification. The modified extended Kalman filter based model has better performance compared to the existing model in estimating the core temperature as it considers two separate measurement models for exercise and recovery periods. The proposed modified model is validated using human trials data which includes participants wearing regular clothing and different types of personal cooling clothing. It is found that the proposed modified model can predict core temperature more accurately (bias= −0.03 to −0.06°C, LoA=±0.14 – ±0.32°C, and RMSE=0.09–0.16°C) as compared to the existing model (bias=−0.03 – 0.08°C, LoA=±0.26 – ±0.33°C, and RMSE=0.14°C–0.18°C). Personal cooling intervention minimises core temperature rise significantly and thus helps in heat stress mitigation. It is observed that the proposed modified model can estimate core temperature more accurately for cooling intervention cases and intermittent working cases as compared to the existing model. It can be used effectively in an ambulatory setting, but for situations where higher accuracy is required and invasiveness of the method is not a concern, direct contact method of core temperature is always preferred as the proposed model has varying accuracy levels between individuals.

Validation and simulation of cavitation flow in a centrifugal pump by filter-based turbulence model

The simulation of cavitation flow is sensitive to the turbulence model and cavitation model. To improve the accuracy of numerical simulation, a centrifugal pump with specific speed of 56.5 is taken as the research object, and the Filter-Based Model (FBM) was used to correct the RNG turbulence model. Zwart-Gerber-Belamri (ZBG) cavitation model, Kunz model and Schnerr-Sauer model were employed for the simulation of cavitation flow in a centrifugal pump. According to the comparison, the results calculated by the ZGB model were 8% closer to the experimental results in cavitation predictions. The cavitation occurs around the impeller ring and balance holes firstly, and then the vapor develops and cover 80% of the blade suction side. By using Q-criterion, the distribution of the vortex core region under different cavitation conditions were investigated. The volume of high intensity vortex under severe cavitation condition is 30%∼50% higher than that under other conditions. Furthermore, the distribution of the radial force of the impeller corresponds to the number of blades, and the cavitation affects the resultant force direction of the impeller. With the intensification of cavitation, the radial force of impeller increases gradually, and the regularity of radial force direction will be disorder.

Evaluation of modified turbulent viscosity on shedding dynamic of three-phase cloud cavitation around hydrofoil – numerical/experimental analysis

PurposeThis paper aims to focus on the cavitating flow around the Clark-Y hydrofoil when the dissolved air is taken into account as the third phase. As the RNG k-epsilon model yields poor prediction due to overestimation of viscosity, the modification approaches including density corrected method, filter-based model and filter-based density correction model are used, and the turbulence model is modified. Also, the numerical results are compared with the experimental data.Design/methodology/approachThe cavitating flow is known as a complex multi-phase flow and appeared in the regions where the local pressure drops under saturation vapor pressure. Many researches have been conducted to analyze this phenomenon because of its significant impact on the erosion, vibration, noise, efficiency of turbomachines, etc.FindingsThe experiments are conducted in a rectangular test section equipped with Clark-Y hydrofoil providing cavity visualization, instantaneous pressure and vibration fluctuations. The simulations are carried out for different cavitation numbers with and without dissolved air. The Fast Fourier Transform, continues wavelet transform and temporal-spatial distribution of gray level are implemented to extract and compare the shedding frequency of experiments and numerical predictions and cavitation evolution. It is concluded that the flow structure, shedding frequency and time-averaged characteristics are highly influenced by the dissolved air. Also, the numerical prediction will be more satisfactory when the modified turbulence models are applied.Originality/valueTo the best of the authors’ knowledge, the originality of this study is the modification of the turbulence model for better prediction of cavitating flow, and the validation of numerical results with corresponding experimental data.

The Impact of Sparse Coding on Memory Lifetimes in Simple and Complex Models of Synaptic Plasticity

Models of associative memory with discrete state synapses learn new memories by forgetting old ones. In the simplest models, memories are forgotten exponentially quickly. Sparse population coding ameliorates this problem, as do complex models of synaptic plasticity that posit internal synaptic states, giving rise to synaptic metaplasticity. We examine memory lifetimes in both simple and complex models of synaptic plasticity with sparse coding. We consider our own integrative, filter-based model of synaptic plasticity, and examine the cascade and serial synapse models for comparison. We explore memory lifetimes at both the single-neuron and the population level, allowing for spontaneous activity. Memory lifetimes are defined using either a signal-to-noise ratio (SNR) approach or a first passage time (FPT) method, although we use the latter only for simple models at the single-neuron level. All studied models exhibit a decrease in the optimal single-neuron SNR memory lifetime, optimised with respect to sparseness, as the probability of synaptic updates decreases or, equivalently, as synaptic complexity increases. This holds regardless of spontaneous activity levels. In contrast, at the population level, even a low but nonzero level of spontaneous activity is critical in facilitating an increase in optimal SNR memory lifetimes with increasing synaptic complexity, but only in filter and serial models. However, SNR memory lifetimes are valid only in an asymptotic regime in which a mean field approximation is valid. By considering FPT memory lifetimes, we find that this asymptotic regime is not satisfied for very sparse coding, violating the conditions for the optimisation of single-perceptron SNR memory lifetimes with respect to sparseness. Similar violations are also expected for complex models of synaptic plasticity.

Hodrick–Prescott filter-based hybrid ARIMA–SLFNs model with residual decomposition scheme for carbon price forecasting

Accurate carbon pricing guidance is of great importance for the inhibition of excessive carbon dioxide emissions. Aiming at improving forecast performance, a number of carbon price forecasting models have been proposed based on the combination or multiscale hybrid frameworks. However, most of these hybrid models cannot easily cast a perfect reflection of erratic fluctuation in carbon trading schemes due to lack of judgment on the trend or inaccurate trend reconstruction. In this study, a novel filter-based modeling with Hodrick–Prescott (HP) filter, that can identify repeated up and down structural features around a certain carbon price, negotiates the obstacle of the parallel–series hybridization concerning the linear and the nonlinear model identification. The residual decomposition scheme with adaptive noise is carried out on the random and nonlinear component for error correction to filter-based models. Moreover, Bayesian optimization adjusts the structure of seven single-hidden layer feedforward neural networks (SLFNs) and the inputs to provide the best generalization performance. The proposed filter-hybrid model using kernel extreme learning machine as the final nonlinear integrator has better stability to the parameters, and has the superiority over the parallel–series and allocation-based models from a statistical perspective. Comparing with existing data-driven models, our proposed model is competitive in view of prediction accuracy and time cost in the majority of carbon futures trading cases.

Tracking Unmanned Aerial Vehicles Based on the Kalman Filter Considering Uncertainty and Error Aware

Recently, Unmanned Aerial Vehicles (UAVs) have made significant impacts on our daily lives with the advancement of technologies and their applications. Tracking UAVs have become more important because they not only provide location-based services, but are also faced with serious security threats and vulnerabilities. UAVs are smaller in nature, move with high speed, and operate in a low-altitude environment, which makes it conceivable to track UAVs using fixed or mobile radars. Kalman Filter (KF)-based methodologies are widely used for extracting valuable trajectory information from samples composed of noisy information. As UAVs’ trajectories resemble uncertain behavior, the traditional KF-based methodologies have poor tracking accuracy. Recently, the Diffusion-Map-based KF (DMK) was introduced for modeling uncertainties in the environment without prior knowledge. However, the model has poor accuracy when operating in environments with higher noise. In order to achieve better tracking performance, this paper presents the Uncertainty and Error-Aware KF (UEAKF) for tracking UAVs. The UEAKF-based tracking method provides a good tradeoff among preceding estimate confidence and forthcoming measurement under dynamic environments; the resulting filter is robust and nonlinear in nature. The experimental results showed that the UEAKF-based UAV tracking model achieves much better Root Mean Square Error (RMSE) performance compared to the existing particle filter-based and DMK-based UAV tracking models.

MA Filter-based Modeling and Control of Smoothing Photovoltaic Power Fluctuations by Inverter Air Conditioners

Renewable energies (RES) will receive higher attention in the future, and continuously increase the penetration rate in the power system to achieve low carbon goal. However, the volatility and intermittent nature of RES generation caused by the uncertainty of the weather have brought huge challenges to the safe and stable operation of the power system. IACs have become a substitute for traditional power plants to accommodate RES by changing their operating power. This paper proposes a control method based on the MA filter to smooth photovoltaic generation power fluctuations utilizing inverter air conditioners (IACs). The photovoltaic power data is filtered by the MA filter in real time firstly. An output power with little fluctuations and one power deviation with high frequency fluctuations are obtained. The power deviation with high frequency fluctuations will be smoothed by IACs, while the remaining power deviation will be smoothed by ESSs. The case studies show a smooth power output is obtained for the power system after filtering.

Extended state Kalman filter-based path following control of underactuated autonomous vessels

Path following control of underactuated autonomous vessels remains a challenging issue in recent years due to its inherent underactuation and nonlinearities as well as the widely existing disturbances in the marine environment. In order to accommodate all the difficulties simultaneously, a novel extended state Kalman filter, which adopts the idea of extended state observer in estimating and compensating system lumped disturbance and optimizes the filter gain in a real-time fashion using Kalman filter technique, is constructed to estimate system states and disturbances in the presence of model uncertainties and measurement noise. Based on the estimated states and disturbances, an enhanced model predictive controller is proposed to steer the underactuated autonomous vessels along a predefined path at a desired speed after considering system state and input constraints. Simulation results have proved the superiority of extended state Kalman filter over traditional extended state observer and extended Kalman filter under various disturbance and noise scenarios. Moreover, the comparison results with conventional proportion-integration-differentiation controller have demonstrated the feasibility and efficacy of the proposed extended state Kalman filter-based model predictive controller in both set-point tracking and disturbance rejection.

Enhanced DNNs for malware classification with GAN-based adversarial training

Deep learning based malware classification gains momentum recently. However, deep learning models are vulnerable to adversarial perturbation attacks especially when applied in network security application. Deep neural network (DNN)-based malware classifiers by eating the whole bit sequences are also vulnerable despite their satisfactory performance and less feature-engineering job. Therefore, this paper proposes a DNN-based malware classifier on the raw bit sequences of programs in Windows. We then propose two adversarial attacks targeting our trained DNNs to generate adversarial malware. A defensive mechanism is proposed by treating perturbations as noise added on bit sequences. In our defensive mechanism, a generative adversary network (GAN)-based model is designed to filter out the perturbation noise and those that with the highest probability to fool the target DNNs are chosen for adversarial training. The experiments show that GAN with filter-based model produced the highest quality adversarial samples with medium cost. The evasion ratio under GAN with filter-based model is as high as 50.64% on average. While incorporating GAN-based adversarial samples into training, the enhanced DNN achieves satisfactory with 90.20% accuracy while the evasion ratio is below 9.47%. GAN helps in secure the DNN-based malware classifier with negligible performance degradation when compared with the original DNN. The evasion ratio is remarkably minimized when faced with powerful adversarial attacks, including $${\textit{FGSM}}^r$$ and $${\textit{FGSM}}^k$$ .

Research and Application of Several Key Techniques in Hyperspectral Image Preprocessing.

This paper focuses on image segmentation, image correction and spatial-spectral dimensional denoising of images in hyperspectral image preprocessing to improve the classification accuracy of hyperspectral images. Firstly, the images were filtered and segmented by using spectral angle and principal component analysis, and the segmented results are intersected and then used to mask the hyperspectral images. Hyperspectral images with a excellent segmentation result was obtained. Secondly, the standard reflectance plates with reflectance of 2 and 98% were used as a priori spectral information for image correction of samples with known true spectral information. The mean square error between the corrected and calibrated spectra is less than 0.0001. Comparing with the black-and-white correction method, the classification model constructed based on this method has higher classification accuracy. Finally, the convolution kernel of the one-dimensional Savitzky-Golay (SG) filter was extended into a two-dimensional convolution kernel to perform joint spatial-spectral dimensional filtering (TSG) on the hyperspectral images. The SG filter (m = 7,n = 3) and TSG filter (m = 3,n = 4) were applied to the hyperspectral image of Pavia University and the quality of the hyperspectral image was evaluated. It was found that the TSG filter retained most of the original features while the noise information of the filtered hyperspectral image was less. The hyperspectral images of sample 1–1 and sample 1–2 were processed by the image segmentation and image correction methods proposed in this paper. Then the classification models based on SG filtering and TSG filtering hyperspectral images were constructed, respectively. The results showed that the TSG filter-based model had higher classification accuracy and the classification accuracy is more than 98%.

A Hybrid Intrusion Detection with Decision Tree for Feature Selection

Due to the size and nature of intrusion detection datasets, intrusion detection systems (IDS) typically take high computational complexity to examine features of data and identify intrusive patterns. Data preprocessing techniques such as feature selection can be used to reduce such complexity by eliminating irrelevant and redundant features in the dataset. The objective of this study is to analyze the efficiency and effectiveness of some feature selection approaches namely, wrapper-based and filter-based modeling approaches. To achieve that, a hybrid of feature selection algorithm in combination with wrapper and filter selection processes is designed. We propose a wrapper-based hybrid intrusion detection modeling with a decision tree algorithm to guide the selection process. Five machine learning algorithms are used on the wrapper and filter-based feature selection methods to build IDS models using the UNSW-NB15 dataset. The three filter-based methods namely, information gain, gain ratio, and relief are used for comparison to determine the efficiency and effectiveness of the proposed approach. Furthermore, a fair comparison with other state-of-the-art intrusion detection approaches is also performed. The experimental results show that our approach is quite effective in comparison to state-of-the-art works, however, it takes high computational time in comparison to the filter-based methods whilst achieves similar results. Our work also revealed unobserved issues about the conformity of the UNSW-NB15 dataset.