Lower Estimation Error Research Articles

The computational age‐at‐death estimation from 3D surface models of the adult pubic symphysis using data mining methods

Age-at-death estimation of adult skeletal remains is a key part of biological profile estimation, yet it remains problematic for several reasons. One of them may be the subjective nature of the evaluation of age-related changes, or the fact that the human eye is unable to detect all the relevant surface changes. We have several aims: (1) to validate already existing computer models for age estimation; (2) to propose our own expert system based on computational approaches to eliminate the factor of subjectivity and to use the full potential of surface changes on an articulation area; and (3) to determine what age range the pubic symphysis is useful for age estimation. A sample of 483 3D representations of the pubic symphyseal surfaces from the ossa coxae of adult individuals coming from four European (two from Portugal, one from Switzerland and Greece) and one Asian (Thailand) identified skeletal collections was used. A validation of published algorithms showed very high error in our dataset—the Mean Absolute Error (MAE) ranged from 16.2 and 25.1 years. Two completely new approaches were proposed in this paper: SASS (Simple Automated Symphyseal Surface-based) and AANNESS (Advanced Automated Neural Network-grounded Extended Symphyseal Surface-based), whose MAE values are 11.7 and 10.6 years, respectively. Lastly, it was demonstrated that our models could estimate the age-at-death using the pubic symphysis over the entire adult age range. The proposed models offer objective age estimates with low estimation error (compared to traditional visual methods) and are able to estimate age using the pubic symphysis across the entire adult age range.

Model of Series-Parallel Photovoltaic Arrays Using Double-Diode Model and Parallel Computing

Several applications require to estimate the power production of photovoltaic (PV) systems under partial shading conditions. For example, dynamic reconfiguration of the array connections is needed to maximize the power production under partial shading conditions, which requires estimating the power generated by the PV array in several possible configurations. Therefore, a fast and accurate modeling technique is needed to perform those calculations in practical times and with low estimation errors. To address those kinds of problems, this paper proposes a modeling approach based on the double-diode model to provide high accuracy at low voltage and low irradiance conditions, which are important for partial-shading analysis. Moreover, the proposed modeling technique is designed to be implemented in parallel processing devices; thus, the calculation time is much shorter in comparison with classical serial solutions. The proposed model is tested in terms of accuracy and speed, obtaining satisfactory results. Finally, the applicability of the parallel model in reconfiguration applications is demonstrated using an application example.

Advanced sliding mode speed observer with compensated stray‐load and iron losses for a stand‐alone wind turbine‐driven induction generator

AbstractModel‐based speed estimation schemes in sensorless field‐oriented control (FOC) of induction machines (IMs) are designed on the basis of a corresponding mathematical model. Thus, a reliable and robust sensorless controller requires using a more detailed IM model. The omission of nonlinear phenomena in real‐time application, such as magnetic saturation, iron losses, and stray‐load losses, certainly leads to inaccuracy in the estimated speed due to the existence of the mentioned phenomena in the IM. In this context, we address here the problem of neglecting the stray‐load and iron losses in a model‐based speed observer, synthesized by means of sliding mode theory, and we evaluate the impact of this omission on the execution of the utilized FOC and maximum power point tracking (MPPT) algorithm applied for a stand‐alone wind energy conversion system (WECS). The performance of the proposed sliding mode speed observer (SMSO), derived from the IM model including the above mentioned nonlinear phenomena, is experimentally evaluated over wide ranges of wind speed and IM magnetizing flux. It is also compared with the model reference adaptive system (MRAS) speed estimator, which is derived from the same IM model, as well as to the simplified SMSO, designed based on the IM model without stray‐load and iron losses. The obtained results show that the proposed SMSO has greater robustness with respect to abrupt variations of the rotor flux in comparison with the MRAS speed estimator, whereas it ensures significantly more stable operation and lower speed estimation error in comparison with the simplified SMSO.

Channel estimation algorithm based on attention mechanism

As the key to wireless communication, channel estimation has become a hot research topic in recent years. In this paper, we propose a deep learning method based on the channel estimation of inverse convolutional network and expanded convolutional network to address the problems that the performance of traditional channel estimation algorithms in orthogonal frequency division multiplexing (OFDM) systems can hardly meet the communication requirements of complex scenarios and are greatly affected by noise. The method constructs a lightweight deconvolutional network using the correlation of channels, and achieves channel interpolation and estimation step by step with a few layers of deconvolutional operations, which achieves channel estimation with low complexity. To improve the estimation performance, an expanded convolutional network is further constructed to suppress the channel noise and improve the accuracy of channel estimation. The simulation results show that the channel estimation can be performed at different signal levels. The simulation results show that the proposed deep learning method based on deconvolution and dilation convolution has lower estimation error and lower complexity than the traditional methods under different signal-to-noise ratios (SNR).

Federated Learning for Channel Estimation in Conventional and RIS-Assisted Massive MIMO

Machine learning (ML) has attracted a great research interest for physical layer design problems, such as channel estimation, thanks to its low complexity and robustness. Channel estimation via ML requires model training on a dataset, which usually includes the received pilot signals as input and channel data as output. In previous works, model training is mostly done via centralized learning (CL), where the whole training dataset is collected from the users at the base station (BS). This approach introduces huge communication overhead for data collection. In this paper, to address this challenge, we propose a federated learning (FL) framework for channel estimation. We design a convolutional neural network (CNN) trained on the local datasets of the users without sending them to the BS. We develop FL-based channel estimation schemes for both conventional and RIS (intelligent reflecting surface) assisted massive MIMO (multiple-input multiple-output) systems, where a single CNN is trained for two different datasets for both scenarios. We evaluate the performance for noisy and quantized model transmission and show that the proposed approach provides approximately 16 times lower overhead than CL, while maintaining satisfactory performance close to CL. Furthermore, the proposed architecture exhibits lower estimation error than the state-of-the-art ML-based schemes.

MIMO FMCW Radar with Doppler-Insensitive Polyphase Codes

Co-located MIMO is used to enlarge the antenna aperture virtually and increase the angular resolution. This paper shows FMCW radar using MIMO VAA. Polyphase codes are designed for modulating the successive chirps of transmitting signals. The codes are optimized to have low cross-correlations regardless of the Doppler filter mismatch. Compared with orthogonal codes, the designed codes show robust performance for Doppler mismatch and lower angle estimation errors. The entire procedure is explained, and the simulation results are provided.

Calibration and validation of phenological models for Biome-BGCMuSo in the grasslands of Tibetan Plateau using remote sensing data

Modeling vegetation phenology is crucial to assessing how climate change impacts carbon cycles in terrestrial ecosystems. The process-based biogeochemical model Biome-BGCMuSo is widely used for simulating carbon and water storages and fluxes of grassland ecosystems. However, the lack of accurate phenological information, such as the start of the growing season (SOS), impedes better simulations of the biogeochemical processes in the Tibetan Plateau (TP). Here, based on the snow-free satellite-derived SOS and the end of the growing season (EOS) in the TP during 1982–2018, we calibrated and validated three phenological models for SOS (i.e., the Biome-BGC phenological (BBGC) model, the heatsum growing season index (HSGSI) model, and the alpine meadow prognostic phenological (AMPP) model) and five phenological models for EOS (i.e., BBGC, HSGSI, AMPP, the low temperature and photoperiod multiplicative model induced by photoperiod (TPMP) and temperature (TPMT)) using particle swarm optimization (PSO) algorithm. For SOS, all three phenological models with calibrated parameters performed similarly and all captured the change in SOS well along gradients of aridity. The performance of BBGC, HSGSI, and AMPP models were largely improved with the calibration. The AMPP model simulated SOS with the lowest estimation errors with the mean absolute error (MAE) of 18.67 days and the Kling-Gupta efficiency (KGE) of 0.47 in validation. For EOS, the calibrated HSGSI and AMPP models, with mean MAEs of 9.85 and 9.29 days, respectively, captured the change in EOS well along the gradients of aridity and performed better than other models. The calibration significantly improved the simulation performance of all five models. Therefore, the phenological models can be calibrated and validated at a large scale with snow-free satellite-derived phenological data. Our study recommends that calibration and validation for the phenological model play a vital role in accurately simulating SOS and EOS in the regional carbon cycle simulation.

Time‐Varying Channel Estimation Based on Air‐Ground Channel Modelling and Modulated Learning Networks

To improve the time-varying channel estimation accuracy of orthogonal frequency division multiplexing air-ground datalink in complex environment, this paper proposes a time-varying air-ground channel estimation algorithm based on the modulated learning networks, termed as MB-ChanEst-TV. The algorithm integrates the modulated convolutional neural networks (MCNN) with the bidirectional long short term memory (Bi-LSTM), where the MCNN subnetworks accomplish channel interpolation in frequency domain and compress the network model while the Bi-LSTM subnetworks achieve channel prediction in time domain. Considering the unique characteristics of airframe shadowing for unmanned aircraft systems, we propose to combine the classical 2-ray channel model with the tapped delay line model and present a more realistic channel impulse response samples generation approach, whose code and dataset have been made publicly available. We demonstrate the effectiveness of our proposed approach on the generated dataset, where experimental results indicate that the MB-ChanEst-TV model outperforms existing state-of-the-art methods with a lower estimation error and better bit error ratio performance under different signal to noise ratio conditions. We also analyze the effect of roll angle of the aircraft and the duration percentage of the airframe shadow on the channel estimation.

A GCICA Grant-Free Random Access Scheme for M2M Communications in Crowded Massive MIMO Systems

A novel grant-free random access scheme with a high success rate is proposed to support massive access for machine-to-machine communications in massive multiple-input–multiple-output (MIMO) systems. This scheme allows active user equipments (UEs) to transmit their modulated uplink messages and super pilots consisting of multiple subpilots to a base station (BS). Then, the BS performs channel state information (CSI) estimation and uplink message decoding by utilizing a proposed graph combined clustering independent component analysis (GCICA) decoding algorithm and then employs the estimated CSIs to detect active UEs by using the characteristic of asymptotic favorable propagation of massive MIMO channel. We call this proposed scheme as the GCICA-based random access (GCICA-RA) scheme. We analyze the successful access probability, missed detection probability, and uplink throughput of the GCICA-RA scheme. Numerical results show that the GCICA-RA scheme significantly improves the successful access probability and uplink throughput, decreases missed detection probability, and provides low CSI estimation error at the same time.

Magnetic and electric antennas calibration for partial discharge charge estimation in gas-insulated substations

There are no accepted procedures that quantify the apparent charge of partial discharge (PD) in gas-insulated substations (GIS). This paper proposes a calibration method for PD charge estimation using unconventional electromagnetic sensors: a magnetic loop antenna (inductive coupler) and an electric antenna (capacitive coupler.) The calibration procedure is intended for the voltage double integral method, which is reviewed for magnetic antennas and extended for electric antennas. By injecting low-frequency sinusoidal signals, the calibration constants are determined for two different test setups: the first one being a testbench where the characteristic impedance is matched and the second one a full-scale 420 kV GIS. The calibration method is validated in three ways: with a calibrated pulse in the testbench, a calibrated pulse in a full-scale GIS, and PD defects in the full-scale GIS. The calibration procedure revealed a frequency limit range dependent on the GIS length and the sensor’s signal-to-noise ratio. The three validation methods showed low charge estimation errors for the magnetic and electric antennas, demonstrating that the PD calibration method applies to any electric/magnetic detector with a low-frequency derivative response. This research paves the way for better GIS insulation monitoring and PD sensor harmonization.© 2017 Elsevier Inc. All rights reserved.

Design and Evaluation of a Robust Observer Using Dead-Zone Lyapunov Functions—Application to Reaction Rate Estimation in Bioprocesses

This paper addresses the design and evaluation of a robust observer for second order bioprocesses considering unknown bounded disturbance terms and uncertainty in the dynamics of the unknown and known states. The observer design and the stability analysis are based on dead-zone Lyapunov functions, and a detailed procedure is provided. The transient response bounds and the convergence region of the unknown observer error are determined in terms of the disturbance bounds, considering persistent but bounded disturbances in the dynamics of both the known and unknown observer errors. This is a significant contribution to closely related observer design studies, in which the transient response bounds are determined, but persistent and bounded disturbances are not considered in the dynamics of the known observer error. Other important contributions are: (i) the procedure for defining the observer parameters is significantly simpler than common observer designs, since a solution to the Ricatti equation, solution to LMI constraints, or the accomplishment of eigenvalue inequality conditions are not required; (ii) discontinuous signals are not used in the observer; and (iii) the effect of the gain sign associated with the unknown state in the dynamics of the known state is explicitly and clearly considered in the observer design and in the convergence study. In addition, the guidelines for selecting the observer parameters are provided. Numerical simulation confirms the stability analysis results: the observer errors converge within a short time, with a low estimation error, if observer-parameters are properly defined.

Three-dimensional continuous gait trajectory estimation using single Shank-Worn inertial measurement units and clinical walk test application

State-of-the-art estimation methods using inertial measurement units (IMUs) for global continuous gait path and local stepwise gait trajectory during walking have been developed. However, estimation methods for continuous gait trajectory integrating both these aspects with high accuracy are almost lacking. Thus, continuous gait trajectory estimation using a single shank-worn IMU with high accuracy is proposed in this study. This method calculates three-dimensional local stepwise gait trajectory based on IMU measurement data extracted between adjacent middle points of stance phases during walking. Continuous gait trajectory is estimated by concatenating adjacent local stepwise gait trajectories based on relative angles determined according to stride vectors and shank orientations. Evaluation experiments results obtained using the optical motion capture system with 12 healthy participants demonstrated estimation errors in the stride length (− 0.027 (− 0.054 to − 0.006) m) and turning angle (0.7 (− 0.2–1.7)°), and normalized endpoint position error (0.029 (0.019–0.04) m). Comparing with previous reports, the proposed method integrally achieves a continuous gait trajectory with a low estimation error level in both local and global aspects despite the continuous measurement of multiple gait cycles. The proposed simple and low-cost method can be applied in the medical field and contribute to expansion of the application of precise gait information in daily life.

A dataset of eye gaze images for calibration-free eye tracking augmented reality headset

Eye tracking is a widely used technique. To enhance eye gaze estimation in different contexts, many eye tracking datasets have been proposed. However, these datasets depend on calibrations in data capture and its applications. We seek to construct a dataset that enables the design of a calibration-free eye tracking device irrespective of users and scenes. To reach this goal, we present ARGaze, a dataset with 1,321,968 pairs of eye gaze images at 32 × 32 pixel resolution and 50 corresponding videos of world views based on a replicable augmented reality headset. The dataset was captured from 25 participants who completed eye gaze tasks for 30 min in both real-world and augmented reality scenes. To validate the dataset, we compared it against state-of-the-art eye gaze datasets in terms of effectiveness and accuracy and report that the ARGaze dataset achieved record low gaze estimation error by 3.70 degrees on average and 1.56 degrees on specific participants without calibrations to the two scenes. Implications for generalising the use of the dataset are discussed.

Preliminary study on estimation of flow velocity vectors using focused transmit beams

High-frame-rate ultrasound imaging with plane wave transmissions is a predominant method of blood flow imaging, and methods for estimation of blood flow velocity vectors have been developed based on high-frame-rate imaging. On the other hand, in imaging of soft tissues, such as arterial walls and atherosclerotic plaques, high-frame-rate imaging sometimes suffers from high-level clutters. Even in observation of the arterial wall with a focused transmit beam, it would be highly beneficial if blood flow velocity vectors could be estimated simultaneously. We conducted a preliminary study on the estimation of blood flow velocity vectors based on a multi-angle Doppler method with focused transmit beam and parallel receive beamforming. It was shown that the lowest estimation error was achieved at a steering angle of 25° by simulation. Moreover, velocity vectors with typical velocity magnitudes and directions could be obtained by the proposed method in in vivo measurement of a carotid artery.

Optimal Planning of Air Quality-Monitoring Sites for Better Depiction of PM2.5 Pollution across China.

A myriad of studies have attempted to use ground-level observations to obtain gap-free spatiotemporal variations of PM2.5, in support of air quality management and impact studies. Statistical methods (machine learning, etc.) or numerical methods by combining chemical transport modeling and observations with data assimilation techniques have been typically applied, yet the significance of site placement has not been well recognized. In this study, we apply five proper orthogonal decomposition (POD)-based sensor placement algorithms to identify optimal site locations and systematically evaluate their reconstruction ability. We demonstrate that the QR pivot is relatively more reliable in deciding optimal monitoring site locations. When the number of planned sites (sensors) is limited, using a lower number of modes would yield lower estimation errors. However, the dimension of POD modes has little impact on reconstruction quality when sufficient sensors are available. The locations of sites guided by the QR pivot algorithm are mainly located in regions where PM2.5 pollution is severe. We compare reconstructed PM2.5 pollution based on QR pivot-guided sites and existing China National Environmental Monitoring Center (CNEMC) sites and find that the QR pivot-guided sites are superior to existing sites with respect to reconstruction accuracy. The current planning of monitoring stations is likely to miss sources of pollution in less-populated regions, while our QR pivot-guided sites are planned based on the severity of PM2.5 pollution. This planning methodology has additional potentials in chemical data assimilation studies as duplicate information from current CNEMC-concentrated stations is not likely to boost performance.

A Spatially Confined, Platform-Based HF Direction Finding Array

A spatially-confined, platform-based HF antenna array with enhanced bandwidth and direction-finding (DF) accuracy is presented. As the HF band is widely used in many important civilian and military applications, HF DF systems are of great interest in electronic warfare support applications. However, due to the large wavelength, practical DF arrays mounted on moving platforms tend to be electrically small, resulting in fundamental limitations on antenna bandwidth and DF accuracy. Furthermore, many applications require mounting the array at a single location on the platform due to limited space and mobility considerations. To alleviate these issues, we employ the platform as the major radiator by exciting several of its significant characteristic modes (CMs) using a spatially-confined antenna array. Specifically, we examined four DF array designs mounted on a representative, mid-size passenger airplane operating at 10 MHz in simulation, each of which is confined within a virtual sphere with a 3-m diameter. A scaled model of one design was fabricated and characterized. Cramer-Rao bounds and Monte-Carlo test results are also reported. Our results demonstrate that the proposed DF array can provide a 15% lower DoA estimation error than the theoretical limit of a standalone array possessing the same volume and having the same bandwidth.

A Real-Time PPG Peak Detection Method for Accurate Determination of Heart Rate during Sinus Rhythm and Cardiac Arrhythmia.

Objective: We have developed a peak detection algorithm for accurate determination of heart rate, using photoplethysmographic (PPG) signals from a smartwatch, even in the presence of various cardiac rhythms, including normal sinus rhythm (NSR), premature atrial contraction (PAC), premature ventricle contraction (PVC), and atrial fibrillation (AF). Given the clinical need for accurate heart rate estimation in patients with AF, we developed a novel approach that reduces heart rate estimation errors when compared to peak detection algorithms designed for NSR. Methods: Our peak detection method is composed of a sequential series of algorithms that are combined to discriminate the various arrhythmias described above. Moreover, a novel Poincaré plot scheme is used to discriminate between basal heart rate AF and rapid ventricular response (RVR) AF, and to differentiate PAC/PVC from NSR and AF. Training of the algorithm was performed only with Samsung Simband smartwatch data, whereas independent testing data which had more samples than did the training data were obtained from Samsung’s Gear S3 and Galaxy Watch 3. Results: The new PPG peak detection algorithm provides significantly lower average heart rate and interbeat interval beat-to-beat estimation errors—30% and 66% lower—and mean heart rate and mean interbeat interval estimation errors—60% and 77% lower—when compared to the best of the seven other traditional peak detection algorithms that are known to be accurate for NSR. Our new PPG peak detection algorithm was the overall best performers for other arrhythmias. Conclusion: The proposed method for PPG peak detection automatically detects and discriminates between various arrhythmias among different waveforms of PPG data, delivers significantly lower heart rate estimation errors for participants with AF, and reduces the number of false negative peaks. Significance: By enabling accurate determination of heart rate despite the presence of AF with rapid ventricular response or PAC/PVCs, we enable clinicians to make more accurate recommendations for heart rate control from PPG data.

Data‐Driven–Based Internal Temperature Estimation for Lithium‐Ion Battery Under Variant State‐of‐Charge via Electrochemical Impedance Spectroscopy

Internal temperature estimation is critical to the safe operation of lithium‐ion batteries (LIBs), and electrochemical impedance spectroscopy (EIS)‐based methods have been demonstrated to be promising. However, accurate internal temperature estimation under variant state‐of‐charge (SoC) is still challenging due to the combined impact of temperature and SoC on the EIS. Accordingly, this work proposes a novel EIS‐based internal temperature estimation approach, for which SoC‐insensitive EIS features are quantitatively selected and utilized for temperature estimation using support vector regression (SVR) with unknown SoC. First, the EIS feature selection is performed to select SoC‐insensitive features from the imaginary of impedance spectrum. Subsequently, an SVR‐based framework and a well‐trained SVR model are created to estimate the internal temperature of LIBs. The performance of the proposed model is validated by its lowest estimation error (0.57 °C) under known and unknown SoCs compared to that of the existing methodologies. The results confirmed that the proposed method holds the advantage in estimating the internal battery temperature with different SOCs.

Hybrid Approach Based on Machine Learning for Hand Shape and Key Point’s Estimation

In human-computer interaction and virtual truth, hand pose estimation is essential. Public dataset experimental analysis Different biometric shows that a particular system creates low manual estimation errors and has a more significant opportunity for new hand pose estimation activity. Due to the fluctuations, self-occlusion, and specific modulations, the structure of hand photographs is quite tricky. Hence, this paper proposes a Hybrid approach based on machine learning (HABoML) to enhance the current competitiveness, performance experience, experimental hand shape, and key point estimation analysis. In terms of strengthening the ability to make better self-occlusion adjustments and special handshake and poses estimations, the machine learning algorithm is combined with a hybrid approach. The experiment results helped define a set of follow-up experiments for the proposed systems in this field, which had a high efficiency and performance level. The HABoML strategy decreased analysis precision by 9.33% and is a better solution.

STSR: Spectro-Temporal Super-Resolution Analysis of a Reference Signal Less Photoplethysmogram for Heart Rate Estimation During Physical Activity

The increasing demand for intelligent, wearable health monitoring devices with a Photoplethysmogram (PPG) sensor to monitor real-time heart rate (HR) in a subject attracts great attention from researchers. However, the PPG-enabled wearable devices are sensitive to motion artifacts, which results in inaccurate HR estimation. The accuracy of HR estimation is possible only by eliminating motion artifacts. A method based on multi-resolution Spectro-Temporal Super Resolution (STSR) and superlet transform (SLT) is proposed to improve the estimation accuracy of a motion artifact corrupted PPG signal. This proposed methodology does not require a reference accelerometer signal to suppress motion artifacts to estimate HR in real-time during physical activities. Hence, the proposed method is a computationally efficient algorithm. In the first step, a multi-order SLT separates the motion artifacts from the acquired PPG signal in the proposed method. In the next step, a high-intensity lobe in the spectro-temporal representation of the SLT spectrogram is detected to compute the HR. In the final step, an HR smoothing algorithm that uses HR from the preceding windows is proposed to accurately estimate HR in a motion artifact effected signal. The performance of the SLT-based HR estimation algorithm is evaluated using the publicly available IEEE SPC dataset. The average HR estimation error is 1.01 beats per minute, and the Pearson correlation is 0.997. Low estimation error, short computational time, and fast HR tracking make the proposed method an ideal choice to implement in wearable devices for real-time HR estimation.