Real-time Correction Research Articles

Enhancing Yoga Learning with a Low-Complexity, Mobilebased Deep Learning System for Real-Time Posture Identification and Correction

Enhancing Yoga Learning with a Low-Complexity, Mobilebased Deep Learning System for Real-Time Posture Identification and Correction

Efficient Framework for Real-Time Color Cast Correction and Dehazing Using Online Algorithms to Approximate Image Statistics

Efficient Framework for Real-Time Color Cast Correction and Dehazing Using Online Algorithms to Approximate Image Statistics

Voice-Based Sinhala Document Maker Application

This research paper introduces "Word Sri," an inventive voice-based web application designed to enhance written Sinhala communication and originality. The application incorporates features such as a plagiarism checker, grammar checker, voice-activated punctuation, and intuitive voice commands in an effort to accommodate a diverse user base, including individuals with disabilities, language learners, writers, journalists, students, and professionals. The first aspect of the paper explores the creation of a user-friendly Sinhala Grammar Checker with Fix Sentences & Punctuation, addressing the dearth of effective language tools for grammar, punctuation, and orthography correction. Using a comprehensive analysis of Sinhala language components, literature, and a corpus of text data, the model provides students, academicians, and professionals with real-time error feedback and correction. The second aspect highlights the distinctive characteristics of the "Word Sri" application, emphasizing its usefulness for voice-based Sinhala content creation. Notably, the application includes a cutting-edge Sinhala plagiarism analyzer that uses machine learning algorithms, synonym and paraphrase detection to improve the originality of content. This research represents a significant advancement in Sinhala language technology, as it provides a unified platform for efficient voice-based typing, accurate grammar checks, and plagiarism detection, thereby facilitating effective written communication and originality in Sinhala for global users.

A Real-Time Radar QPE Error Correction for Convective Precipitation Using Long-Term TRMM-PR and GPM-DPR Observations

A Real-Time Radar QPE Error Correction for Convective Precipitation Using Long-Term TRMM-PR and GPM-DPR Observations

Dual-layer operation strategy of integrated energy system based on multivariate predictive control

Scheduling and control are two key dimensions to ensure the smooth operation of the integrated energy system, but there still exists the problem of insufficient accuracy of scheduling instructions due to modeling deviation between the scheduling model and the actual equipment. In this paper, a dual-layer operation strategy for integrated energy system based on multivariate predictive control is proposed, so as to make real-time corrections to the scheduling power instructions. The results show that the dual-layer operation strategy based on multivariate predictive control can better deal with the uncertainties of renewable energy output and electric load, and effectively reduce the impact of prediction deviation and scheduling model deviation on system operation. In the example calculation, compared with the operation strategy with only the scheduling layer, the total supply energy deviation with the dual-layer operation strategy decreases from 61.16 kWh to 22.96 kWh, which realizes the smooth and efficient operation of the system.

An Enhanced Multi-Sensor Simultaneous Localization and Mapping (SLAM) Framework with Coarse-to-Fine Loop Closure Detection Based on a Tightly Coupled Error State Iterative Kalman Filter

In order to attain precise and robust transformation estimation in simultaneous localization and mapping (SLAM) tasks, the integration of multiple sensors has demonstrated effectiveness and significant potential in robotics applications. Our work emerges as a rapid tightly coupled LIDAR-inertial-visual SLAM system, comprising three tightly coupled components: the LIO module, the VIO module, and the loop closure detection module. The LIO module directly constructs raw scanning point increments into a point cloud map for matching. The VIO component performs image alignment by aligning the observed points and the loop closure detection module imparts real-time cumulative error correction through factor graph optimization using the iSAM2 optimizer. The three components are integrated via an error state iterative Kalman filter (ESIKF). To alleviate computational efforts in loop closure detection, a coarse-to-fine point cloud matching approach is employed, leverging Quatro for deriving a priori state for keyframe point clouds and NanoGICP for detailed transformation computation. Experimental evaluations conducted on both open and private datasets substantiate the superior performance of the proposed method compared to similar approaches. The results indicate the adaptability of this method to various challenging situations.

Cross-attention learning enables real-time nonuniform rotational distortion correction in OCT.

Nonuniform rotational distortion (NURD) correction is vital for endoscopic optical coherence tomography (OCT) imaging and its functional extensions, such as angiography and elastography. Current NURD correction methods require time-consuming feature tracking/registration or cross-correlation calculations and thus sacrifice temporal resolution. Here we propose a cross-attention learning method for the NURD correction in OCT. Our method is inspired by the recent success of the self-attention mechanism in natural language processing and computer vision. By leveraging its ability to model long-range dependencies, we can directly obtain the spatial correlation between OCT A-lines at any distance, thus accelerating the NURD correction. We develop an end-to-end stacked cross-attention network and design three types of optimization constraints. We compare our method with two traditional feature-based methods and a CNN-based method on two publicly-available endoscopic OCT datasets. We further verify the NURD correction performance of our method on 3D stent reconstruction using a home-built endoscopic OCT system. Our method achieves a ∼3 × speedup to real time (26 ± 3 fps), and superior correction performance.

Machine Learning-Based Optimization of Wind-PV Solution for Grid Demand

This research addresses the imperative need for innovative approaches to harness renewable resources while ensuring grid stability amidst global electricity demand. The proposed optimization methodology, centered on machine learning, harmonizes the intermittent output of Wind-PV systems with dynamic grid requirements. Leveraging historical meteorological data, energy trends, and real-time demand, the study employs ensemble learning, specifically the CNN (Convolutional Neural Network) algorithm, for accurate forecasting of PMSG-based (Permanent Magnet Synchronous Generator) Wind-PV system output. The two-tiered optimization approach involves short-term prediction using the CNN model and real-time correction through a genetic algorithm, enhancing the Capacity Utilization factor from 73% to 92% and reducing mean absolute error from 8% to 4.5%. This method not only anticipates energy fluctuations efficiently but also contributes to grid stability and decreased reliance on nonrenewable sources. Comparative analyses demonstrate the superiority of the machine learning-based approach in terms of accuracy and adaptability over traditional control methods.

Self-calibrating dual-sensing electrochemical sensors for accurate detection of carbon dioxide in blood.

A paper-based electrochemical dual-function biosensor capable of determining pH and TCO2 was synthesized for the first time using an iridium oxide pH electrode and an all-solid-state ion electrode (ASIE). In the study, to obtain highly reliable results, the biosensor was equipped with a real-time pH correction function before TCO2 measurements. Compared to traditional liquid-filling carbon dioxide detection sensors, the utilization of ferrocene endows our novel sensor with abundant positive sites, and thus greatly improves its performance. Conversely, the introduction of MXene with conductivity close to that of metals reduces electrode resistance, which is beneficial for accelerating the electrochemical reaction of the sensor and reducing LOD. After optimization, the detection range of TCO2 is 0.095 nM-0.66 M, with a detection limit of as low as 0.023 nM. In addition, the sensor was used in real serum sample-spiked recovery experiments and comparison experiments with existing clinical blood gas analyzers, which confirmed the effectiveness of its clinical application. This study provides a method for the rational design of paper-based electrochemical biosensors and a new approach for the clinical detection of blood carbon dioxide.

Robust real-time estimation of non-uniform angular velocity and sub-pixel jitter in images captured with resonant scanners.

Images captured with resonant scanners are affected by angular velocity fluctuations that result in image distortion and by poor synchronization between scanning and light detection that creates jitter between image rows. We previously demonstrated that both problems can be mitigated in post-processing by recording the scanner orientation in synchrony with the image capture, followed by data resampling [Opt. Express30, 112 (2022)10.1364/OE.446162]. Here we introduce more robust algorithms for estimation of both angular velocity fluctuation and jitter in the presence of random and deterministic noise. We also show linearization of the scanner oscillation model to reduce calculation times by two orders of magnitude, reaching 65,000 jitter estimations per second when using 2,800 samples per image row, and 500,000 when using only 500 samples, easily supporting real-time generation of jitter-corrected images.

FAST CONVERGING LIDAR-AIDED PRECISE POINT POSITIONING: A CASE STUDY WITH LOW-COST GNSS

Abstract. With the growing interest in autonomous driving, accurate vehicle positioning remains an open problem, especially in urban environments. According to regulatory organisations, the vehicle positioning accuracy is required to be centimetre-level. As the most used positioning technique which provides globally referenced positioning solutions, GNSS is the fundamental component for realising real-time vehicle positioning, usually through the RTK approach. However, RTK requires a nearby reference station to enable integer ambiguity resolution for the ultra-precise carrier phase observations. In comparison, PPP makes use of State-Space Representation (SSR) corrections produced by global networks for satellite orbits and clocks to facilitate phase-based positioning. Moreover, IGS now offers Real-Time Service (RTS) to transmit such corrections. Notably, the major drawback of PPP is that it takes a long time to converge to precise solutions due to the carrier phase ambiguities being real-valued, which can be severely elongated when real-time corrections and low-cost GNSS receivers are used. In this paper, a tightly coupled positioning method is proposed, which shortens RT-PPP convergence to seconds by using lidar measurements referenced from an HD map through deep learning. The lidar measurements are generated by point cloud registration and weighted by their intensity values and geometric distributions, and are then combined with RT-PPP in an Extended Kalman-Filter (EKF), thus achieving fast convergence. Experimental results show that the proposed method achieves and maintains centimetre-level accuracy within 2 seconds using a low-cost UBLOX F9P receiver, which is a significant improvement as compared to the decimetre-level accuracy obtained from standalone RT-PPP.

Real-time pose correction and registration of complex workpiece in robotic machining considering normal consistency constrains

Efficient and precision model registration of complex workpiece with arbitrary pose is a challenging task that determines the effects of intelligent robotic machining. Aiming at the problems of limited field of view of 3D vision sensors such as surface structured light and the complexity of scanning-splicing-registration algorithm, in this paper a real-time pose correction and registration algorithm of complex workpiece in robotic machining considering normal consistency constrains is proposed. Specifically, the operations on ellipse feature detection and center point extraction are performed based on the arc-support line segment, and then the actual three-dimensional coordinate of the ellipse is obtained by the camera conversion matrix. Furthermore, considering the constrained normal consistency of the center points set of ellipses on the model surface, the pose of the target workpiece is corrected in real-time by the improved ICP registration algorithm. Taking the automotive flywheel shell as a typical application object, the experimental results demonstrate that the registration correction time of the proposed algorithm is 0.5 s, and the registration accuracy can reach 0.35 mm. The repetitive correction time of the algorithm after arbitrarily changing the workpiece pose is less than 1 s, and the positioning accuracy of the guiding robot is less than 1.5 mm. On the premise of satisfying the robotic machining quality requirements, the pose of the complex workpiece model can be corrected in real-time for subsequent robotic machining.

Multi-time-scale Optimal Scheduling of Park-specific Integrated Energy System Considering Stackelberg Game

With the reform of the electricity market, the coupling interaction between different stakeholders becomes increasingly more obvious. This paper proposes a multi-time-scale optimal scheduling model for park-specific integrated energy system operator (IESO) considering the Stackelberg game given the different interest demands of different subjects in the park-specific integrated energy system (IES) and the accuracy of source-load forecasting. Firstly, a day-ahead Stackelberg game model is established with the park-specific IESO as the leader and the park users as followers. Afterward, taking into account the influence of large errors in source-load forecasts, an intraday dispatch model and a real-time correction model based on model predictive control were established. Through the analysis of calculation examples, it is verified that the multi-time scale optimal scheduling of park-specific IESO which considers the Stackelberg game can effectively reduce system operating costs, improve system economy, and stabilize fluctuations caused by source-load forecast errors.

Deep learning-based attenuation correction method in 99mTc-GSA SPECT/CT hepatic imaging: a phantom study.

This study aimed to evaluate a deep learning-based attenuation correction (AC) method to generate pseudo-computed tomography (CT) images from non-AC single-photon emission computed tomography images (SPECTNC) for AC in 99mTc-galactosyl human albumin diethylenetriamine pentaacetic acid (GSA) scintigraphy and to reduce patient dosage. A cycle-consistent generative network (CycleGAN) model was used to generate pseudo-CT images. The training datasets comprised approximately 850 liver phantom images obtained from SPECTNC and real CT images. The training datasets were then input to CycleGAN, and pseudo-CT images were output. SPECT images with real-time CT attenuation correction (SPECTCTAC) and pseudo-CT attenuation correction (SPECTGAN) were acquired. The difference in liver volume between real CT and pseudo-CT images was evaluated. Total counts and uniformity were then used to evaluate the effects of AC. Additionally, the similarity coefficients of SPECTCTAC and SPECTGAN were assessed using a structural similarity (SSIM) index. The pseudo-CT images produced a lower liver volume than the real CT images. SPECTCTAC exhibited a higher total count than SPECTNC and SPECTGAN, which were approximately 60% and 7% lower, respectively. The uniformities of SPECTCTAC and SPECTGAN were better than those of SPECTNC. The mean SSIM value for SPECTCTAC and SPECTGAN was 0.97. We proposed a deep learning-based AC approach to generate pseudo-CT images from SPECTNC images in 99mTc-GSA scintigraphy. SPECTGAN with AC using pseudo-CT images was similar to SPECTCTAC, demonstrating the possibility of SPECT/CT examination with reduced exposure to radiation.

A deep-learning real-time bias correction method for significant wave height forecasts in the Western North Pacific

Significant wave height (SWH) is one of the most important parameters characterizing ocean waves, and accurate numerical ocean wave forecasting is crucial for coastal protection and shipping. However, due to the randomness and nonlinearity of the wind fields that generate ocean waves and the complex interaction between wave and wind fields, current forecasts of numerical ocean waves have biases. In this study, a spatiotemporal deep-learning method was employed to correct gridded SWH forecasts from the European Centre for Medium-Range Weather Forecasting System Integrated Forecast System Global Model (ECMWF-IFS). This method was built on the trajectory gated recurrent unit deep neural network, and it conducts real-time rolling correction for the 0–240-h SWH forecasts from ECMWF-IFS. The correction model is co-driven by wave and wind fields, providing better results than those based on wave fields alone. A novel pixel-switch loss function was developed. The pixel-switch loss function can dynamically fine-tune the pre-trained correction model, focusing on pixels with large biases in SWH forecasts. According to the seasonal characteristics of SWH, four correction models were constructed separately, for spring, summer, autumn, and winter. The experimental results show that, compared with the original ECMWF SWH predictions, the correction was most effective in spring, when the mean absolute error decreased by 12.972–46.237%. Although winter had the worst performance, the mean absolute error decreased by 13.794–38.953%. The corrected results improved the original ECMWF SWH forecasts under both normal and extreme weather conditions, indicating that our SWH correction model is robust and generalizable.

Real-time correction of channel-bed roughness and water level in river network hydrodynamic modeling for accurate forecasting

The accuracy and reliability of hydrodynamic models are sensitive to both hydraulic state variables and model parameters, particularly the bed roughness, while their simultaneous real-time corrections and corresponding effects still need to be well-established and understood. This paper presents a real-time data assimilation model that corrects channel-bed roughness and water level in a river network hydrodynamic model, ensuring its accuracy and reliability. Experiments and parameter analysis evaluated the effect of initial roughness and observation noise level on model performance. Correcting both roughness and water level improved filtering time and forecasting accuracy by up to 63% and 80%, respectively, compared to methods only correcting water level. The filtering time was reduced by 44–63%, and the water level forecasting RMSE decreased by up to 80%. Both models experienced increased filtering time and forecasting error as observation noise increased, but the proposed model had a lower increase. With accurate hydraulic state measurement (e.g., 0.005 m error), the model achieved negligible water level forecasting error after 7 h of data assimilation. The model's accuracy depended on the initial channel-bed roughness, and the algorithm enables real-time roughness correction, making it useful for flood forecasting.

Characterizing the fault performance of real-time precise satellite orbit and clock correction products

The a priori fault probability of the real-time precise satellite orbit and clock correction products is the critical parameter for integrity monitoring of precise point positioning (PPP). The traditional fault probability evaluation methods use the worst-case instantaneous user ranging error (IURE) as the conservative test statistic. However, the systematic biases of IURE contained in the worst-case IURE barely affect the PPP accuracy, which will undermine the statistical distribution of test statistic and reduce the sensitivity of fault detection. The fault probability will be estimated over-conservatively for the traditional methods. By clarifying the sources of the systematic biases, a new test statistic is constructed by deliberately removing the systematic biases of IURE originated from satellite orbit and clock errors. One-year Global Positioning System correction products evaluation results have demonstrated that the constructed test statistic follows the Gaussian distribution with the decreased uncertainty and the improved fault detection sensitivity. The real-world data experiments have shown that the a priori probabilities of the satellite fault and the constellation fault are at the order of 10−4 and 10−5 levels, respectively.

Integrated adaptive iterative learning control based on inter-trial iteration and real-time correction for nonlinear systems with external morphologically-similar disturbances

This paper proposes an adaptive iterative learning control (ILC) scheme for nonlinear systems with iteration-varying trial lengths under external morphologically-similar disturbances, utilising inter-trial iteration and real-time correction. To attenuate external quasi-periodic disturbances and non-repetitive uncertainties, and further achieve better tracking performance along iteration and time axes, the proposed scheme combines the iteration-to-iteration proportional-type (P-type) ILC with the within-iteration P-type scheme. The tracking error with dead zone property and zero filling treatment is constructed. In addition, as opposed to the existing two-dimensional (2D) ILC works, the integrated framework is formed through the connection of adaptive weights, which are calculated by the adaptive weight determination method based on the ideology of the Kalman filter. The convergence of the algorithm is proved based on the contraction mapping principle. Compared with the traditional ILC schemes, illustrative and applicational simulations are provided to demonstrate the effectiveness and the superiority of the proposed framework.

A Real-Time Energy Response Correction Method for Cs3Cu2I5:Tl Scintillating Dosimeter.

The uneven energy response of radiation detectors severely limits the accuracy of the dose rate meter used for radiation protection. Currently widely used in dose rate meters as a physical method of setting shielding compensation, the energy response correction error of the detector at different energies is mostly between 15 and 25%. This work designs a real-time correction method for energy response based on a novel Cs3Cu2I5:Tl scintillation detector to improve the accuracy of the dose rate meter used for radiation protection. The technique utilizes the idea of pulse amplitude weighting (PAW) to segment the pulse amplitude histogram. This detector achieves an almost constant energy response after our correction. The experimental results show that compared to 137Cs γ rays, the maximum error of the response is 8.26% in the photon energy ranging from 33 keV to 1.25 MeV, which is much better than ±30% of the recommended IEC 61526:2010, verifying the feasibility of PAW.

Comparative Analysis of Image-shift Measurement Algorithms for Solar Shack–Hartmann Wavefront Sensors

Observations of the Sun provide unique insights into its structure, evolution, and activity, with significant implications for space weather forecasting and solar energy technologies. Ground-based telescopes offer cost-effective and flexible solutions for high-resolution solar observations, but image quality can be affected by atmospheric turbulence. Adaptive optics (AO) systems equipped with Shack–Hartmann wave front sensors (SH-WFS) enable real-time image correction to mitigate these effects. The accuracy of SH-WFS relies on correlation algorithms that measure wave front shifts, but reaching consistent conclusions regarding their accuracy remains challenging. In this study, we conducted an evaluation and comparison of standard correlation algorithms (the Square Difference Function, Normalized Cross-Correlation, Absolute Difference Function, Absolute Difference Function-Squared, and the Covariance Function in the frequency domain (CFF)) using simulated and authentic solar images. We optimized the algorithms through pre-processing techniques and carefully selected the most suitable window function for the CFF algorithm. Additionally, we analyzed the influence of various factors, such as shift ranges, bias, and the size of live images on the accuracy of algorithms. The consistent findings revealed that the CFF algorithm demonstrates superior measurement accuracy and robustness compared to the others. Choosing the CFF algorithm for solar observations can significantly enhance measurement accuracy, AO system performance, and the overall quality of solar research findings, thereby providing crucial support for space weather forecasting and other related scientific fields.