Improve Estimation Performance Research Articles

Noise‐robust structural response estimation method using short‐time Fourier transform and long short‐term memory

AbstractStructural response estimation based on deep learning can suffer from reduced estimation performance owing to discrepancies between the training and test data as the noise level in the test data increases. This study proposes a short‐time Fourier transform‐based long short‐term memory (STFT‐LSTM) model to improve estimation performance in the presence of noise and ensure estimation robustness. This model enables robust estimations in the presence of noise by positioning an STFT layer before feeding the data into the LSTM layer. The output transformed into the time‐frequency domain by the STFT layer is learned by the LSTM model. The robustness of the proposed model was validated using a numerical model with three degrees of freedom at various signal‐to‐noise ratio levels, and its robustness against impulse and periodic noise was verified. Experimental validation assessed the estimation robustness under impact load and verified the robustness against environmental noise in the acquired acceleration response.

Investigating Model Form Error Estimation for Sparse Data

Computational simulations of dynamical systems involve the use of mathematical models and algorithms to mimic and analyze complex real-world phenomena. By leveraging computational power, simulations enable researchers to explore and understand systems that are otherwise challenging to study experimentally. They offer a cost-effective and efficient means to predict and analyze the behavior of physical, biological, and social systems. However, model form error arises in computational simulations from simplifications, assumptions, and limitations inherent in the mathematical model formulation. Several methods for addressing model form error have been proposed in the literature, but their robustness in the face of challenges inherent to real-world systems has not been thoroughly investigated. In this work, a data assimilation-based approach for model form error estimation is investigated in the presence of sparse observation data. Including physics-based domain knowledge to improve estimation performance is also explored. The Lotka-Volterra equations are employed as a simple computational simulation for demonstration.

A review of Bayes filters with machine learning techniques and their applications

A Bayes filter is a widely used estimation algorithm, but it has inherent limitations. Performance can degrade when the dynamics are highly nonlinear or when the probability distribution of the state is unknown. To mitigate these issues, machine learning (ML) techniques have been incorporated into many Bayes filters, due to their advantage of being able to map between the input and the output without explicit instructions. In this review, we reviewed 90 papers that proposed the use of ML techniques with Bayes filters to improve estimation performance. This review provides an overview of Bayes filters with ML techniques, categorised according to the role of ML, remaining challenges and research gaps. In the concluding section of this review, we point out directions for future research.

FusionNetV2: Explicit Enhancement of Edge Features for 6D Object Pose Estimation

FusionNet is a hybrid model that incorporates convolutional neural networks and Transformers, achieving state-of-the-art performance in 6D object pose estimation while significantly reducing the number of model parameters. Our study reveals that FusionNet has local and global attention mechanisms for enhancing deep features in two paths and the attention mechanisms play a role in implicitly enhancing features around object edges. We found that enhancing the features around object edges was the main reason for the performance improvement in 6D object pose estimation. Therefore, in this study, we attempt to enhance the features around object edges explicitly and intuitively. To this end, an edge boosting block (EBB) is introduced that replaces the attention blocks responsible for local attention in FusionNet. EBB is lightweight and can be directly applied to FusionNet with minimal modifications. EBB significantly improved the performance of FusionNet in 6D object pose estimation in experiments on the LINEMOD dataset.

A Multi-Mechanism Fusion Method for Robust State of Charge Estimation via Bidirectional Long Short-Term Memory Model with Mixture Kernel Mean p-Power Error Loss Optimized by Golden Jackal Optimization Algorithm

Abstract Accurate state of charge (SOC) estimation is crucial for effective battery management in various applications. The bidirectional long short-term memory (BiLSTM) as an outstanding nonlinear regression model can be used for SOC estimation. This work develops a novel multi-mechanism fusion method based on BiLSTM to further enhance its estimation performance for SOC, in which the convolutional neural network (CNN), attention mechanism, and mixture kernel mean p-power error (MKMPE) loss are introduced into the BiLSTM framework for addressing different issues. First, the introduction of CNN components aims to extract essential features from battery data, enhancing the model's comprehension of complex information. Then, the attention mechanism is used to further refine the model's perceptual ability and a robust MKMPE loss is introduced into the BiLSTM framework to replace its original mean squared error loss, and a novel robust model is developed to suppress non-Gaussian noise interference. Finally, some key hyperparameters of the proposed model are fine-tuned using the golden jackal optimization algorithm, resulting in improved estimation performance. Comparative numerical experiments are meticulously conducted in various cases to evaluate the performance of the proposed method, and the experiment results demonstrate that it can perform outstanding effectiveness in handling non-Gaussian noise scenarios.

A Computation Model to Estimate Interaction Intensity through Non-Verbal Behavioral Cues: A Case Study of Intimate Couples under the Impact of Acute Alcohol Consumption

This work introduced a novel analysis method to estimate interaction intensity, i.e., the level of positivity/negativity of interaction, for intimate couples (married and heterosexual) under the impact of alcohol, which has great influences on behavioral health. Non-verbal behaviors are critical in interpersonal interactions. However, whether computer vision-detected non-verbal behaviors can effectively estimate the interaction intensity of intimate couples is still unexplored. In this work, we proposed novel measurements and investigated their feasibility to estimate interaction intensities through machine learning regression models. Analyses were conducted based on a conflict-resolution conversation video dataset of intimate couples before and after acute alcohol consumption. Results showed the estimation error was at the lowest in the no-alcohol state but significantly increased if the model trained using no-alcohol data was applied to after-alcohol data, indicating that alcohol altered the interaction data in the feature space. While training a model using rich after-alcohol data is ideal to address the performance decrease, data collection in such a risky state is challenging in real life. Thus, we proposed a new State-Induced Domain Adaptation (SIDA) framework, which allows for improving estimation performance using only a small after-alcohol training dataset, pointing to a future direction of addressing data scarcity issues.

Generalized Moment Estimators Based on Stein Identities

For parameter estimation of continuous and discrete distributions, we propose a generalization of the method of moments (MM), where Stein identities are utilized for improved estimation performance. The construction of these Stein-type MM-estimators makes use of a weight function as implied by an appropriate form of the Stein identity. Our general approach as well as potential benefits thereof are first illustrated by the simple example of the exponential distribution. Afterward, we investigate the more sophisticated two-parameter inverse Gaussian distribution and the two-parameter negative-binomial distribution in great detail, together with illustrative real-world data examples. Given an appropriate choice of the respective weight functions, their Stein-MM estimators, which are defined by simple closed-form formulas and allow for closed-form asymptotic computations, exhibit a better performance regarding bias and mean squared error than competing estimators.

A Soft Sensor for Flow Estimation and Uncertainty Analysis Based on Artificial Intelligence: A Case Study of Water Supply Systems

The fourth industrial revolution has transformed the industry, with information technology playing a crucial role in this shift. The increasing digitization of industrial systems demands efficient sensing and control methods, giving rise to soft sensors that have the potential to replace traditional physical sensors in order to reduce costs and enhance efficiency. This study explores the implementation of an artificial neural network (ANN) based soft sensor model in a water supply system to predict flow rates within the system. The soft sensor is centered on a Long Short-Term Memory (LSTM) artificial neural network model using Monte Carlo dropout to reduce uncertainty and improve estimation performance. Based on the results of this work, it is concluded that the proposed soft sensor (with Monte Carlo dropout) can predict flow rates more precisely, contributing to the reduction in water losses, as well as cost savings. This approach offers a valuable solution for minimizing water losses and ensuring the efficient use of this vital resource. Regarding the use of soft sensors based on LSTM neural networks with a careful choice of Monte Carlo dropout parameters, when compared to the multilayer perceptron model, the LSTM model with Monte Carlo dropout showed better mean absolute error, root mean square error, and coefficient of determination: 0.2450, 0.3121, and 0.996437 versus 0.2556, 0.3522, and 0.9954. Furthermore, this choice of Monte Carlo dropout parameters allowed us to achieve an LSTM network model capable of reducing uncertainty to 1.8290, keeping the error metrics also at low levels.

A statistical sensing method by utilizing Wi-Fi CSI subcarriers: Empirical study and performance enhancement

In modern Wi-Fi systems, channel state information (CSI) serves as a foundational support for various sensing applications. Currently, existing CSI-based techniques exhibit limitations in terms of environmental adaptability. As such, optimizing the utilization of subcarrier CSI stands as a critical avenue for enhancing sensing performance. Within the OFDM communication framework, this work derives sensing outcomes for both detection and estimation by harnessing the CSI from every individual measured subcarrier, subsequently consolidating these outcomes. When contrasted against results derived from CSI based on specific extraction protocols or those obtained through weighted summation, the methodology introduced in this study offers substantial improvements in CSI-based detection and estimation performance. This approach not only underscores the significance but also serves as a robust exemplar for the comprehensive application of CSI.

Using transformer and a reweighting technique to develop a remaining useful life estimation method for turbofan engines

Estimating the Remaining Useful Life (RUL) of industrial machinery is an important task in Prognostics and Health Management (PHM). Accurate RUL prediction based on current health status and real-time sensor measurements can help establish efficient maintenance strategies in advance, thus improving the productivity and efficiency of machinery operations. However, considering the nature of industrial machinery, a data imbalance problem, which exists in current RUL estimation datasets like the Commercial Modular Aero-Propulsion System Simulation (C-MAPSS) dataset, hinders data-driven approaches. To address the detrimental data imbalance problem and improve estimation performance, this research proposes an Adaptive RUL-wise Reweighting (ARR) technique. ARR uses kernel smoothing to maintain the inherent continuity of RUL and a reweighting method to rebalance the effects of samples with different ground-truth RULs. In addition, this research proposes a novel RUL estimation method that uses a transformer architecture, which has not yet been widely applied for PHM tasks. Using a transformer-based estimation model with high expressive power capable of handling multivariate time-series data has proven effective in RUL estimation. Comprehensive experiments using the C-MAPSS dataset based on cross validation are conducted to find optimal hyperparameter configurations and to validate the proposed method’s effectiveness in RUL estimation. For subsets FD001 and FD003 of the C-MAPSS dataset, the proposed method provides state-of-the-art estimation performance with the lowest Root Mean Squared Error (RMSE) values of 11.36 and 11.28 and score values of 192.22 and 133.41, respectively. The proposed method also shows highly competitive performance on subsets FD002 and FD004 of the C-MAPSS dataset.

A Network-Group Target State and Network Topology Estimation Method Based on Signals Containing Delay, Doppler and Address

Network-group targets are a set of objectives that adhere to a shared communication protocol, perform common tasks, and exhibit relatively coordinated movements. Typically, network-group targets emit radar and communication signals. However, they often employ a silent mode to evade our perception. Despite this, they continue to exchange data through their communication networks. By intercepting the communication signals of these targets, this paper proposes a method for estimating the state and network topology of network-group targets based on random finite set (RFS) theory. This method is based on the modeling of network-group targets using a labeled multi-Bernoulli (LMB) RFS. In state estimation, the usual method involves extracting the localization parameters from the signals in the first step and use these parameters for target tracking in the second step. This study focused on estimating the kinematic states of network-group targets using communication signals containing delay and Doppler information received by multiple mobile sensor platforms. The proposed method considers the coherency between frequency-hopping pulses in the signals, resulting in an improved estimation performance. In addition, considering that network-group targets require network access for information exchange, graph theory concepts were utilized to estimate the network topology of network-group targets by address measurement. The simulation results validated the effectiveness of the proposed method and demonstrated its superior performance.

Augmenting sparse arrays in ocean acoustics: A Gaussian Process approach

Source localization and geoacoustic inversion performance depends on the location of receiving phones, their spacing, their number, and the array aperture. Performance suffers when practical design issues limit the capabilities of the array. In this work, sparse arrays are augmented leading to the generation of virtual arrays for better sampling of the ocean and, thus, improved estimation performance. To that effect, Gaussian Processes are employed, which are shown to create high-fidelity field “measurements” at virtual, densely spaced hydrophones. Kernel functions are key building blocks in the implementation of Gaussian Processes, as they quantify the field coherence at neighboring spatial points. Functions of interest are the squared exponential, Matern, and plane wave kernels. We validate our method with application to synthetic data as well as data collected during the Seabed Characterization Experiment conducted in 2022. [Work supported by ONR.]

Maximum likelihood based direction estimation for noncircular signals

In the direction estimation for the signals incident on a sensor array, maximum likelihood (ML) based methods can provide superior performance than subspace based ones such as the multiple signal classification (MUSIC). When the incoming signals are noncircular the exploitation of the property can allow us to improve estimation performance. Based on the deterministic ML criterion, a direction estimation method for strictly noncircular signals is proposed that can outperform the conventional alternating maximization (AM) with no use of noncircularity. Using a generalized Gaussian probability density function, we approach the ML estimation, which is formulated as a nonlinear multidimensional problem. The application of the alternating projection to the multidimensional problem leads to the maximization of an objective function of two variables associated with the direction and the initial phase of a noncircular signal. Theoretically optimizing the phase variable, we obtain the maximum through one-dimensional search with respect to the direction variable only. The complexity of the proposed noncircular AM (NC-AM) is far less than that of the existing ML based method, noncircular decoupled maximization (NC-DM). Moreover, simulation results demonstrate that NC-AM outperforms NC-DM as well as the noncircular MUSIC.

The use of crowdsourced mobile data in estimating pedestrian and bicycle traffic: A systematic review

To address the need for better non-motorized traffic data, policymakers and researchers are collaborating to develop new approaches and methods for estimating pedestrian and bicyclist traffic volumes. Crowdsourced mobile data, which has higher spatial and temporal coverage and lower collection costs than data collected through traditional approaches, may help improve pedestrian and bicyclist traffic estimation despite their limitations or biases. This systemic literature review documents how researchers have used crowdsourced mobile data to estimate pedestrian and bicyclist traffic volumes. We find that one source of commercial fitness application data (i.e., Strava) has been used much more frequently than other crowdsourced mobile data, and that most studies have used crowdsourced mobile data to estimate bicyclist volumes. Comparatively few studies have estimated pedestrian volumes. The most common approach to the use of crowdsourced counts is as independent variables in direct demand models. Variables constructed from crowdsourced mobile data not only have significant correlations with observed counts in statistical models but also have larger relative importance than other factors in machine learning models. Studies also show that including crowdsourced mobile data can significantly improve estimation performance. Future research directions include application of crowdsourced mobile data in more pedestrian traffic estimations, comparison of the performance of different crowdsourced mobile data, incorporation of multiple data sources, and expansion of the methods using crowdsourced mobile data for non-motorized traffic estimation.

Improvement of real-time state estimation performance in HVDC systems using an adaptive nonlinear observer

This paper proposes a novel adaptive nonlinear observer design for a voltage source converter based on high-voltage direct current (VSC-HVDC) transmission systems. We consider a system consisting of a power grid, and a converter station connected to an unknown load through a long HVDC cable. The primary contribution of this work is the development of a global high-gain observer that facilitates the estimation of all system states. Specifically, it encompasses the estimation of power grid parameters, such as the angular frequency and the voltage at the point of common coupling (PCC), as well as the states of the HVDC cable and the current absorbed by the load. The performance of the proposed observer is assessed through theoretical analysis and simulations. Additionally, we implemented our observer on a digital signal processor (DSP) eZdsp in a processor-in-the-loop (PIL) quasi-real-time setting. Experimental results, coupled with numerical simulations, showcase the outstanding performance of our proposed observer.

A new sensor configuration design method for source term estimation in urban neighborhood with complex conditions under different wind directions

When hazardous substances leak into the atmosphere, Source Term Estimation (STE) can quickly determine the information of the leaked substance to control the threat of the leakage event to the environment and public safety. Reasonable sensor configuration is the foundation of STE. Existing sensor configuration schemes have limited applicability in STE processes. When urban neighborhood has complex building structures and airflow conditions, there is a lack of a method that can provide reasonable sensor configurations to ensure good STE performance. In this study, a new sensor configuration design method is proposed, which weights the joint entropy of adjoint concentration information of multi wind directions to obtain the cost function, and then uses Simulated Annealing (SA) algorithm to minimize the cost function. This article selects a real complex urban neighborhood for simplified modeling, and designs the optimal sensor array for it. STE results of several leakage scenarios are compared with the optimal and uniform sensor configurations at different leakage locations and wind directions. Compared to traditional uniform sensor configurations, the optimal sensor configuration has significantly improved estimation performance in both source location and strength, with a reduction of 27 %–51 % in the bias of source location identification and 17 %–39 % in the bias of source strength estimation.

Learning Generative Models Using Denoising Density Estimators.

Learning probabilistic models that can estimate the density of a given set of samples, and generate samples from that density, is one of the fundamental challenges in unsupervised machine learning. We introduce a new generative model based on denoising density estimators (DDEs), which are scalar functions parametrized by neural networks, that are efficiently trained to represent kernel density estimators of the data. Leveraging DDEs, our main contribution is a novel technique to obtain generative models by minimizing the Kullback-Leibler (KL)-divergence directly. We prove that our algorithm for obtaining generative models is guaranteed to converge consistently to the correct solution. Our approach does not require specific network architecture as in normalizing flows (NFs), nor use ordinary differential equation (ODE) solvers as in continuous NFs. Experimental results demonstrate substantial improvement in density estimation and competitive performance in generative model training.

IMPACTS OF USING POROUS CORNER PARTITIONS AND BLADE SHAPED NANOPARTICLES IN BASE FLUID ON THE PERFORMANCE IMPROVEMENT OF THERMOELECTRIC GENERATOR MOUNTED VENTED CAVITIES AND INTERFACE TEMPERATURE ESTIMATION WITH PROPER ORTHOGONAL DECOMPOSITION

Numerous technical applications, such as solar power, refrigeration, waste heat recovery, thermal management, and many more, employ thermoelectric devices because of the numerous advantages they offer. Techniques are required to improve their effectiveness in usage. In this study, a unique technique for improving the performance of a thermoelectric generator (TEG) positioned between vented cavities is proposed. The method combines the effects of blade-shaped nanoparticles in the base fluid and corner porous partitions. A numerical study using the finite element method is conducted for different values of Darcy number of upper and lower cavity (10<sup>-6</sup> ≤ Da<sub>1</sub> ≤ 10<sup>-2</sup>, 10<sup>-6</sup> ≤ Da<sub>2</sub> ≤ 10<sup>-2</sup>), opening ratio (SR between 0.5 and 2.5), and nanoparticle loading amount (SVF between 0 and 0.03). By adjusting the permeability of the partitions and opening ratio of the cavities, significant changes in the flow field can be obtained. TEG power increases with greater opening ratios and lower partition permeabilities. By changing the permeability, the TEG power can increase by up to 27.5 percent, while increasing the opening ratio from SR = 0.5 to SR = 1 can increase the TEG power by up to 140 percent. Using blade-shaped nanoparticles results in additional improvements, with values of 32.7% at SR = 0.5 and 20.26% at SR = 2.5. Using 180 parametric computational fluid dynamics cases, a POD-based low-cost reconstruction model is created for the interface temperatures at the hot and cold sides using five-POD modes. The method may be used for other complicated geometries where 3-D computations are costly and allows for the rapid computation of coupled TEG-vented cavity systems with corner partitions.

Efficient Sink Node Position Estimation using Harris Hawks Optimization Algorithm in Wireless Sensor Networks

Wireless sensor network (WSN) was utilized widely in numerous areas owing to their accessibility in data collection, processing, and transmission, and the strength and reliability of data processing and transmission are based on the accuracy of the positions of sensor nodes (SNs) in the WSN. Sink node location estimation in WSN is a vital task intended to define the geographical position of the sink node in the network area of coverage. This procedure normally includes using numerous localization techniques that trust data like received signal strength, arrival time, time variance of arrival, or angle of arrival from adjacent SNs. The accuracy of sink node localization directly influences the efficiency of data aggregation, routing procedures, and complete performance of the network in tasks like environmental monitoring, target tracking, and event recognition. As WSNs are frequently used in remote environments where physical involvement is unusable, an effective and accurate sink node localization model plays a vital part in certifying the network's longevity and reliability. This study develops an Efficient Sink Node Position Estimation using the Harris Hawks Optimization (SNPE-HHO) Algorithm in WSN. The main intention of the SNPE-HHO technique is to recognize the optimal position of the sink node in the network. To achieve this, the SNPE-HHO technique employs the HHO system which gets inspiration from the hunting tactics of Harris Hawk. Moreover, the SNPE-HHO technique computes a fitness function that can drive the searching direction of the HHO algorithm and enhance the node estimation performance. The performance analysis of the SNPE-HHO method is verified by utilizing distinct metrics. The experimentation values confirmed the improved estimation performance of the SNPE-HHO technique over other existing methods

Estimation of Activity and Molar Excess Gibbs Energy of Binary Liquid Alloys Al-Cu, Al-Ni, and Al-Fe from the Partial Radial Distribution Function Simulated by Ab Initio Molecular Dynamics

To accurately and conveniently obtain the thermodynamic data of binary liquid alloys, a new method is proposed in this study. It combines ab initio molecular dynamics (AIMD) simulation with a thermodynamic model to estimate the activity and molar excess Gibbs energy of binary liquid alloys. Additionally, two methods of grouping the partial radial distribution function (PRDF) of 5000 steps obtained by simulation are proposed for the first time. The PRDF of Al50Cu50, Al50Ni50, and Al50Fe50 is obtained by AIMD simulation. These PRDF are combined with four thermodynamic models to estimate the activity and molar excess Gibbs energy. Furthermore, the estimation results of the four models are compared with those of the Miedema model. The results show that when the first peak of the PRDF is obtained by the symmetric method, the average relative deviation (ARD) of the activity and molar excess Gibbs energy of the four models are, respectively: 28% and 32% for Molecular Interaction Volume Model (MIVM); 162% and 38% for Regular Solution Model (RSM); 508% and 65% for Wilson model; 562% and 67% for Non-Random Two-Liquid (NRTL). When the first peak of PRDF is obtained by non-symmetric method, the average ARD of the activity and molar excess Gibbs energy of the four models are, respectively: 64% and 20% for MIVM; 115% and 26% for RSM; 661% and 70% for Wilson; 727% and 72% for NRTL. In addition, the average ARD of the activity and molar excess Gibbs energy of the Miedema model are 113% and 33%. These data indicate that the estimation performance of the MIVM model is superior to the other four models, and the symmetric method performs better than the non-symmetric method. The grouping treatment of PRDF data effectively improves estimation performance.