Limited Number Of Measurements Research Articles

Prime number-based multistep measurement for separation of roundness errors

Roundness measurement is critical in manufacturing, as it ensures that products conform to precise design specifications. However, traditional multistep measurements for roundness error separation are time-consuming and limited in their ability to separate specific Fourier components. In this study, we propose a novel combined multistep measurement method with prime numbers that overcomes these limitations. We demonstrate this method through three experimental cases, achieving high levels of Fourier components in error separation with a limited number of measurements. Our method combines two ( p and q) or three ( p, q, and r) steps of prime numbers to achieve high levels of Fourier components for error separation, compared to traditional multistep measurements that require more steps. In the first experimental case, we use a 2-step and 5-step measurement to achieve traditional multistep measurement in ten steps. In the second case, we use 3-step and 5-step measurements, and in the third, we combine the 2-step, 3-step, and 5-step measurements. We achieve roundness deviations (RONt) of 12.7, 7.8, and 9.9 nm, respectively, and maximum En-values of 0.8, 0.8, and 0.7, respectively. Our proposed combined multistep measurement method using prime numbers has practical applications in manufacturing, as it reduces the time and resources required for roundness error separation while achieving a higher level of Fourier components. Our results demonstrate the effectiveness of our method and its potential to revolutionize roundness measurement in industry.

Source Tracing of PM2.5 in a Metropolitan Area Using a Low-Cost Air Quality Monitoring Network: Case Study of Denver, Colorado, USA

Air quality assessments often require source apportioning of the air pollutants observed at the receptor site. Conventional source apportionment models are subject to high uncertainties due to the lack of accurate emission profiles of all the contributing sources and a limited number of measurements at the receptor sites. Recent advances in the development and application of low-cost PM2.5 sensors have facilitated the formation of a more robust database with greater numbers of measurements per location and time. The main objective of this study is to combine a large database of PM2.5 concentration records to records from low-cost sensors in Denver, Colorado, during January 2021. Using wind speed and wind direction at the receptors, we developed a visualization tool for source tracing of PM2.5 with resulting statistical analyses and back-trajectory modeling. For this purpose, a combination of in-house and existing packages of R scripts along with National Oceanic and Atmospheric Administration (NOAA)’s trajectory model and climate and weather toolkits were used. In general, the results show that the PM2.5 measurements obtained from such a network of PM2.5 sensors incorporated with hourly wind field data, which are publicly available, can provide a powerful screening tool to discover the transport pathways of PM2.5 before requiring costly source apportionment approaches. The fraction of PM2.5 concentration detected by each sensor in regard to wind direction and speed bins were quantified using this method. The results of cluster analysis identified the area groups in respect to wind speed and wind direction bins, which shines a light on how far and in which direction polluting sources are. Finally, the back-trajectory modeling outputs illustrated the exact travel path of the PM2.5-laden air parcels of each day to each sensor.

Surface velocity reconstruction of a vibrating structure based on dictionary learning and sparse sampling

The sparse regularization has been successfully applied to near-field acoustic holography to provide the reconstruction accuracy of sound field with limited number of measurements. However, most of the applications are concentrated on the reconstruction of the sound pressure and the corresponding sparse bases are designed for the sound pressure. In this study, dictionary learning is introduced and K-SVD is utilized to generate a sparse basis for the velocity. Then the reconstruction of surface velocity of a vibrating structure can be realized in a sparse framework to improve the reconstruction accuracy with limited number of measurements. In the process of data sample selection, the equivalent source method is used to generate the velocity samples according to the feature of the sound field and samples can be obtained by numerical simulations. The results of numerical simulations and experiment demonstrate the validity of the learned dictionary and the advantage of the proposed method.

Back-and-forth nudging moving horizon estimation for discrete-time linear systems

In this paper, we propose a novel moving horizon estimation algorithm for discrete-time linear systems with a limited number of measurements. Motivated by the idea of the back-and-forth nudging algorithm, we design the back-and-forth nudging in time moving-horizon estimation method by deploying two moving horizon estimators that move backward and forward iteratively. Based on a finite number of measurements, the proposed algorithm can be used for moving-horizon state estimation with guaranteed convergence. By using the proposed method, we show that the norm of the state estimation error is upper bounded by a sequence that converges to its steady-state in finite-time (i.e., using a finite number of measurements), provided that suitable parameters are selected. The unbiasedness properties and comparison results with the conventional (forward-in-time) moving horizon estimation are discussed. The effectiveness of the proposed approach is validated through a numerical example.

Global Modeling of Heat-Integrated Distillation Column Based on Limited Local Measurements

The heat-integrated distillation column (HIDiC) has more energy-saving potential than conventional distillation columns. However, its nonlinearity and coupling effects pose significant challenges for the online operation of the HIDiC. To overcome these challenges, it becomes necessary to utilize accurate nonlinear models for design optimization or control schemes. Traditional modeling methods require extensive tray information, implying the impractical use of numerous sensors in real-world applications. This paper proposes a modeling approach for the HIDiC based on a limited number of measurements. It only requires the measurement of a finite amount of tray information to construct a global model of the HIDiC. This method serves as an online observer, providing real-time information about the entire column, and also enables the prediction of tray concentration changes. The proposed model forms the basis for developing model-based online monitoring and control schemes. Experimental simulation results demonstrate that the proposed method achieves high accuracy in global observation and prediction for the HIDiC.

Muscle-Tendon Unit Length Measurement Using 3D Ultrasound in Passive Conditions: OpenSim Validation and Development of Personalized Models.

This study investigated the validity of using OpenSim to measure muscle-tendon unit (MTU) length of the bi-articular lower limb muscles in several postures (shortened, lengthened, a combination of shortened and lengthened involving both joints, neutral and standing) using 3D freehand ultrasound (US), and to propose new personalized models. MTU length was measured on 14 participants and 6 bi-articular muscles (semimembranosus SM, semitendinosus ST, biceps femoris BF, rectus femoris RF, gastrocnemius medialis GM and gastrocnemius lateralis GL), considering 5 to 6 postures. MTU length was computed using OpenSim with three different models: OS (the generic OpenSim scaled model), OS + INSER (OS with personalized 3D US MTU insertions), OS + INSER + PATH (OS with personalized 3D US MTU insertions and path obtained from one posture). Significant differences in MTU length were found between OS and 3D US models for RF, GM and GL (from - 6.3 to 10.9%). Non-significant effects were reported for the hamstrings, notably for the ST (- 1.5%) and BF (- 1.9%), while the SM just crossed the alpha level (- 3.4%, p = 0.049). The OS + INSER model reduced the magnitude of bias by an average of 4% for RF, GM and GL. The OS + INSER + PATH model showed the smallest biases in length estimates, which made them negligible and non-significant for all the MTU (i.e. ≤ 2.2%). A 3D US pipeline was developed and validated to estimate the MTU length from a limited number of measurements. This opens up new perspectives for personalizing musculoskeletal models using low-cost user-friendly devices.

COMPARATIVE STUDY OF TRANSFORMER- AND LSTM-BASED MACHINE LEARNING METHODS FOR TRANSIENT THERMAL FIELD RECONSTRUCTION

Solution reconstruction from limited number of measurements is useful in many areas of heat transfer applications. Unlike the standard problems, such reconstruction problems are ill-posed; thus, the nonuniqueness of solution and inherent instability severely complicates the modeling process. Consequently, more conventional inverse analysis methods to reconstruct solutions remain computationally intractable and lacking sufficient flexibility, especially when dealing with time-dependent problems. Aided by powerful graphical processing units (GPUs), machine learning (ML) methods rose in popularity due to their flexibility and ability to efficiently process large amounts of data. In recent years, the transformer-based ML models have gained recognition for their remarkable performance in natural language processing (NLP) tasks as well as time-series analysis, overshadowing the performance of the ML models conventionally used for sequence processing, such as the long short-term memory (LSTM) models. These achievements make transformer-based models seemingly ideal candidates for reconstructing full solutions from a few measurements. This article compares the performance of these novel transformer-based models with a simple LSTM model in reconstructing transient one-dimensional (1-D) and two-dimensional (2-D) thermal fields using sparse spatial measurements. Counterintuitively, the simple LSTM model achieves higher or comparable prediction accuracy compared to the complex transformer-based models while also exhibiting shorter or comparable training times, which may render transformerbased models a suboptimal choice for reconstructing transient solutions. Instead, more traditional sequence processing ML models, such as LSTM, might be preferred for this purpose.

Assimilation of airborne gamma observations provides utility for snow estimation in forested environments

Abstract. An airborne gamma-ray remote-sensing technique provides a strong potential to estimate a reliable snow water equivalent (SWE) in forested environments where typical remote-sensing techniques have large uncertainties. This study explores the utility of assimilating the temporally (up to four measurements during a winter period) and spatially sparse airborne gamma SWE observations into a land surface model (LSM) to improve SWE estimates in forested areas in the northeastern US. Here, we demonstrate that the airborne gamma SWE observations add value to the SWE estimates from the Noah LSM with multiple parameterization options (Noah-MP) via assimilation despite the limited number of measurements. Improvements are witnessed during the snow accumulation period, while reduced skills are seen during the snowmelt period. The efficacy of the gamma data is greater for areas with lower vegetation cover fraction and topographic heterogeneity ranges, and it is still effective at reducing the SWE estimation errors for areas with higher topographic heterogeneity. The gamma SWE data assimilation (DA) also shows a potential to extend the impact of flight-line-based measurements to adjacent areas without observations by employing a localization approach. The localized DA reduces the modeled SWE estimation errors for adjacent grid cells up to 32 km distance from the flight lines. The enhanced performance of the gamma SWE DA is evident when the results are compared to those from assimilating the existing satellite-based SWE retrievals from the Advanced Microwave Scanning Radiometer 2 (AMSR2) for the same locations and time periods. Although there is still room for improvement, particularly for the melting period, this study shows that the gamma SWE DA is a promising method to improve the SWE estimates in forested areas.

Fast Radio Map Construction with Domain Disentangled Learning for Wireless Localization

The accuracy of wireless fingerprint-based indoor localization largely depends on the precision and density of radio maps. Although many research efforts have been devoted to incremental updating of radio maps, few consider the laborious initial construction of a new site. In this work, we propose an accurate and generalizable framework for efficient radio map construction, which takes advantage of readily-available fine-grained radio maps and constructs fine-grained radio maps of a new site with a small proportion of measurements in it. Specifically, we regard radio maps as domains and propose a Radio Map construction approach based on Domain Adaptation (RMDA). We first employ the domain disentanglement feature extractor to learn domain-invariant features for aligning the source domains (available radio maps) with the target domain (initial radio map) in the domain-invariant latent space. Furthermore, we propose a dynamic weighting strategy, which learns the relevancy of the source and target domain in the domain adaptation. Then, we extract the domain-specific features based on the site's floorplan and use them to constrain the super-resolution of the domain-invariant features. Experimental results demonstrate that RMDA constructs a fine-grained initial radio map of a target site efficiently with a limited number of measurements. Meanwhile, the localization accuracy of the refined radio map with RMDA significantly improved by about 41.35% after construction and is comparable with the dense surveyed radio map (the reduction is less than 8%).

Development of Suspended Sediment Monitoring of the Tisza Using an Indirect Measurement Method

The Tisza River experienced successive flood peaks between 1998 and 2010. The reasons of the increased flood height are various, one of them is overbank floodplain accumulation. The aim of the present study is to apply and test an alternative sediment measurement method to evaluate the sediment transport of the Tisza River, Hungary. The new method could help practitioners to better understand the fluvial processes. A significant problem with the current sediment measurement practice is that it does not consider or just to a limited extent, the hydrological conditions of a river. The data measured at the turbidity measuring station installed in the Middle Tisza at Szolnok are evaluated to determine whether or not this measurement procedure can be applied for the highly various sediment transport conditions of the Tisza. In measurement campaigns and based on continuous data from the turbidity measuring probe, a close relationship between near-bank turbidity and suspended sediment concentration was established. The suspended sediment concentrations calculated from the near-bank turbidity were compared with the results of suspended sediment yield from a few cross-section measurements. The results are encouraging; despite the limited number of measurements, the relationship between the parameters is close. In order to make the method more precise, additional series of measurements are needed, which also cover the high water range.

Generative adversarial networks with physical sound field priors.

This paper presents a deep learning-based approach for the spatiotemporal reconstruction of sound fields using generative adversarial networks. The method utilises a plane wave basis and learns the underlying statistical distributions of pressure in rooms to accurately reconstruct sound fields from a limited number of measurements. The performance of the method is evaluated using two established datasets and compared to state-of-the-art methods. The results show that the model is able to achieve an improved reconstruction performance in terms of accuracy and energy retention, particularly in the high-frequency range and when extrapolating beyond the measurement region. Furthermore, the proposed method can handle a varying number of measurement positions and configurations without sacrificing performance. The results suggest that this approach provides a promising approach to sound field reconstruction using generative models that allow for a physically informed prior to acoustics problems.

Efficient Fourier single-pixel imaging based on weighted sorting.

Fourier single-pixel imaging (FSI) has attracted increased attention in recent years with the advantages of a wide spectrum range and low cost. FSI reconstructs a scene by directly measuring the Fourier coefficients with a single-pixel detector. However, the existing sampling method is difficult to balance the noise suppression and image details within a limited number of measurements. Here we propose a new sampling strategy for FSI to solve this problem. Both the generality of the spectral distribution of natural images in the Fourier domain and the uniqueness of the spectral distribution of the target images in the Fourier domain are considered in the proposed method. These two distributions are summed with certain weights to determine the importance of the Fourier coefficients. Then these coefficients are sampled in order of decreasing importance. Both the simulations and experiments demonstrate that the proposed method can capture more key Fourier coefficients and retain more details with lower noise. The proposed method provides an efficient way for Fourier coefficient acquisition.

Physics-informed variational inference for uncertainty quantification of stochastic differential equations

We propose a physics-informed learning based on variational autoencoder (VAE) to solve data-driven stochastic differential equations when the governing equation is known and a limited number of measurements are available. Our model integrates VAE with the given physical laws expressed by stochastic partial differential equations, allowing the encoder to infer the randomness of the solution. The decoder employs a separate structure of two neural networks, where one network learns the spatial behavior and the other network learns the random behavior of the solution, making both training and inference computationally efficient. We use an evidence lower bound (ELBO) as the loss function, which incorporates the given physical laws by using automatic differentiation to compute the differential operators. The proposed model can be used to solve data-driven forward and inverse stochastic differential equations in a unified framework. We demonstrate the efficiency of the proposed model for learning stochastic processes and solving various types of stochastic partial differential equations.

Heat Transfer Optimization of an Electronic Control Unit Immersed in Forced Liquid Coolant

The current paper aims to present a cooling concept for future centralized platforms of ECUs (Electronic Control Units) from the automotive industry that involves grouping multiple electronic devices into a single system and cooling them with forced convection dielectric coolant. The enhancement consists of replacing the inside air of the module with a dielectric coolant that has a higher thermal conductivity than air and employing an additional prototype system that aids in forced liquid cooling. To meet automotive requirements, the experiments were exposed to an ambient temperature of 85 °C. Temperature measurements on these solutions’ hot spots were compared to those on a thermal paste-only reference electronic module. This study used DFSS (Design for Six Sigma) techniques to determine the ideal pump flow rate, fan air flow rate, and liquid volume in the housing, leading to an optimization in heat dissipation. Finding a trustworthy transfer function that could forecast the impact of the crucial design parameters that had been found was the main goal. The electronics cooled by forced convection coolant improved heat dissipation by up to 60% when compared to the reference module. This demonstrates that the DoE (Design of Experiments) method, which is based on a limited number of measurements, can estimate the behavior of the ECU without the need for a more involved theoretical framework.

Phase Space Reconstruction from Accelerator Beam Measurements Using Neural Networks and Differentiable Simulations.

Characterizing the phase space distribution of particle beams in accelerators is a central part of understanding beam dynamics and improving accelerator performance. However, conventional analysis methods either use simplifying assumptions or require specialized diagnostics to infer high-dimensional (>2D) beam properties. In this Letter, we introduce a general-purpose algorithm that combines neural networks with differentiable particle tracking to efficiently reconstruct high-dimensional phase space distributions without using specialized beam diagnostics or beam manipulations. We demonstrate that our algorithm accurately reconstructs detailed 4D phase space distributions with corresponding confidence intervals in both simulation and experiment using a limited number of measurements from a single focusing quadrupole and diagnostic screen. This technique allows for the measurement of multiple correlated phase spaces simultaneously, which will enable simplified 6D phase space distribution reconstructions in the future.

Federated End-to-End Unrolled Models for Magnetic Resonance Image Reconstruction.

Image reconstruction is the process of recovering an image from raw, under-sampled signal measurements, and is a critical step in diagnostic medical imaging, such as magnetic resonance imaging (MRI). Recently, data-driven methods have led to improved image quality in MRI reconstruction using a limited number of measurements, but these methods typically rely on the existence of a large, centralized database of fully sampled scans for training. In this work, we investigate federated learning for MRI reconstruction using end-to-end unrolled deep learning models as a means of training global models across multiple clients (data sites), while keeping individual scans local. We empirically identify a low-data regime across a large number of heterogeneous scans, where a small number of training samples per client are available and non-collaborative models lead to performance drops. In this regime, we investigate the performance of adaptive federated optimization algorithms as a function of client data distribution and communication budget. Experimental results show that adaptive optimization algorithms are well suited for the federated learning of unrolled models, even in a limited-data regime (50 slices per data site), and that client-sided personalization can improve reconstruction quality for clients that did not participate in training.

Slam and wave load response reconstruction in high speed catamarans using transmissibility on full scale sea trials

In-service measurement of slam and wave load responses provide valuable information in the structural design and weight optimisation of high-speed catamarans. However, a limited number of measurements may be possible due to cost, and data may sometimes be lost or noise contaminated. This paper uses the transmissibility concept to reconstruct a set of unknown responses using a different set of known responses. Transmissibility functions and a transmissibility matrix based on linear response theory are derived from less than 12% of available data. The method is applied to full scale sea trials data on a 98 m high speed catamaran HSV 2 Swift, on which strain, acceleration, and angular velocity were measured. Responses are reconstructed for different vessel headings (head, port bow and starboard seas), speeds (15–35 knots), and sea states. The transmissibility matrix is tested on data from unseen sea states. Excellent agreement is achieved, particularly for larger loading events that would be of interest to structural designers. The study demonstrates the possibility to significantly reduce or build redundancy into the instrumentation.

A Hybrid Traffic Forecasting Model for Urban Environments Based on Convolutional and Recurrent Neural Networks

This article presents a hybrid model for the continuous monitoring and prediction of the spatio-temporal dependencies of road traffic. This model is built on the basis of training a YOLOv4 convolutional neural network (CNN) and two-level long-short-term memory (LSTM). The initial module of the hybrid model ensures the continuous collection of data and the filling of the database on road traffic parameters, taking into account transport infrastructure and meteorological factors. Taking the interpreted and aggregated data as input, a recurrent neural network (RNN) generates more accurate predictions. Experimental results show that the proposed model provides prediction accuracy (for a 20-minute time interval) in the range of 82–97 % using a limited number of measurements.

An Improved Data-Driven Calibration Method with High Efficiency for a 6-DOF Hybrid Robot

This paper proposes an improved data-driven calibration method for a six degrees of freedom (DOF) hybrid robot. It focuses mainly on improving the measurement efficiency and practicability of existing data-driven calibration methods through the following approaches. (1) The arbitrary motion of the hybrid robot is equivalently decomposed into three independent sub-motions by motion decomposition. Sequentially, the sub-motions are combined according to specific motion rules. Then, a large number of robot poses can be acquired in the whole workspace via a limited number of measurements, effectively solving the curse of dimensionality in measurement. (2) A mapping between the nominal joint variables and joint compensation values is established using a back propagation neural network (BPNN), which is trained directly using the measurement data through a unique algorithm involving inverse kinematics. Thus, the practicability of data-driven calibration is significantly improved. The validation experiments are carried out on a TriMule-200 robot. The results show that the robot’s maximal position/orientation errors are reduced by 91.16%/88.17% to 0.085 mm/0.022 deg, respectively, after calibration.

Ratio of non-growing season to growing season N2O emissions in Canadian croplands: an update to national inventory methodology

National inventory reporting of agricultural nitrous oxide (N2O) emissions in Canada is based primarily on measurements obtained using static chambers. In regions with cold winters and an accumulated snowpack (including Canada), these measurements tend to focus on the growing season (typically May–October). However, research has shown that emissions continue throughout the non-growing season (NGS) and that these account for a significant proportion of annual emissions. In the Canadian National Inventory NGS emissions currently are assumed to be adequately captured in western Canada, while they are accounted for in eastern Canada by multiplying the growing season emissions by a correction factor of 1.4, a value that was derived based on a limited number of measurements. Here we use recent Canadian studies to validate this correction factor. We collected data from available Canadian studies that measured soil N2O emissions from agricultural systems for the entire year and determined the proportion of these emissions that occurred during the NGS. The proportion of annual N2O emissions that occurred during the NGS varied widely, ranging from −4% to 119% with a mean of 35.5%, compared to the previous estimate of 30%. Due to high variability, few differences were observed between means associated with climatic, soil, and management variables. To correct for NGS N2O emissions from Canadian agricultural soils, we suggest that the current correction factor for converting growing season to total annual emissions be changed from 1.4 to 1.55 and that this be used for all agricultural soils in Canada rather than just eastern Canada.