Error Accumulation Problem Research Articles

A novel hybrid machine learning coulomb counting technique for state of charge estimation of lithium-ion batteries

Accurate state of charge (SOC) estimation is crucial to ensure safe and reliable operation of battery systems. However, the capacity of an operating battery is difficult to measure, and some existing SOC estimation methods either rely too much on the established battery physical model, such as the Kalman filter and particle filter algorithm, or the number of training samples required is too large, and the online estimation performance is poor, such as neural network and Gaussian regression algorithm, cannot achieve good online estimation results. To solve these problems, this paper proposes a novel hybrid SOC estimation method, movIRVM-Coulomb, which integrates moving mean, incremental learning, relevance vector machine (movIRVM), and coulomb counting method. The hybrid algorithm uses the movIRVM algorithm and Coulomb counting method to estimate SOC alternately so that the estimated value is kept in the accuracy of the first estimation, which avoids the error accumulation problem when using a single algorithm to predict multiple times and improves the generalization performance of the algorithm. Firstly, the offline estimation model based on the RVM algorithm and the online estimation model based on the incremental learning RVM(IRVM) algorithm are established by using only 10 % to 30 % training data approximately. Secondly, aiming at the problem that the estimation result of the RVM algorithm fluctuates violently, this paper combines the moving average strategy to make the geometric average of SOC value by setting a reasonable moving window. This method further improves the estimation accuracy, fitting ability and generalization performance of the RVM algorithm. Finally, jointing the coulomb counting method to further improve the online estimation accuracy. The effectiveness of the proposed algorithm is verified by a single-cell public data set, battery pack experimental test data set and Advisor simulation data set. The experimental results highlight that the estimation accuracy of the movIRVM-Coulomb algorithm is >30 % improved compared with that of the movIRVM algorithm and its root means square error can be restricted within 2 % in a wide temperature range and variable dynamic conditions, including operating conditions of US06, UDDS, NYCC and 1015. Overall, it is manifesting that the movIRVM-Coulomb algorithm requires only a few training samples to have high estimation accuracy, strong generalization ability, and excellent robustness.

Accumulating Errors in Tests of General Relativity with Gravitational Waves: Overlapping Signals and Inaccurate Waveforms

Observations of gravitational waves (GWs) from compact binary coalescences provide powerful tests of general relativity (GR), but systematic errors in data analysis could lead to incorrect scientific conclusions. This issue is especially serious in the third-generation GW detectors in which the signal-to-noise ratio (S/N) is high and the number of detections is large. In this work, we investigate the impacts of overlapping signals and inaccurate waveform models on tests of GR. We simulate mock catalogs for Einstein Telescope and Cosmic Explorer and perform parametric tests of GR using waveform models with different levels of inaccuracy. We find that the systematic error in non-GR parameter estimates could accumulate toward a false deviation from GR when combining results from multiple events, although a Bayesian model selection analysis may not favor a deviation. Waveform inaccuracies contribute most to the systematic errors, but multiple overlapping signals could magnify the effects of systematics owing to the incorrect removal of signals. We also point out that testing GR using selected “golden binaries” with high S/N is even more vulnerable to false deviations from GR. The problem of error accumulation is universal; we emphasize that it must be addressed to fully exploit the data from third-generation GW detectors and that further investigations, particularly in waveform accuracy, will be essential.

Bidirectional Sentence Ordering with Interactive Decoding

Sentence ordering aims at restoring orders of shuffled sentences in a paragraph. Previous methods usually predict orders in a single direction, i.e., from head to tail. However, unidirectional prediction inevitably causes error accumulation, which restricts performance. In this article, we propose a bidirectional ordering method, which predicts orders in both head-to-tail and tail-to-head directions at the same time. In our bidirectional ordering method, two directions can interact with each other and help alleviate the error accumulation problem of ordering. Experiments demonstrate that our method can effectively improve performance of previous models.

Weighted error-output recurrent echo kernel state network for multi-step water level prediction

With development of information techniques in navigation and shipping, machine learning algorithms are applied in enhancing navigation safety. One of critical areas, which attracts lots of attention from scientists and researchers, is water level prediction. Although randomization-based algorithms obtain good performance in water level prediction, these algorithms still have their own limitations. For example, unstable prediction problem caused by random weights selection, and the error accumulation problem caused by the conventional recurrent algorithm. In this study, we combine three proposed approach with the conventional echo state network. Firstly, the Gaussian kernel method transforms the input features into a high-dimensional features, which in some extent improves the forecasting accuracy. Secondly, kernel reservoir states are proposed. It not only abandons the random selected weights, but it also makes hidden neurons to connect closely. Lastly, a novel weighted error-output recurrent multi-step algorithm is proposed. It uses previous forecasting errors to update the current output weights in order to overcome the error accumulation problem. Based on above three approaches, a multi-step water level prediction model is proposed called Weighted Error-output Recurrent Echo Kernel State Network (WER-EKSN). The experimental results and statistical analysis represent that our proposed model has better forecasting performance than other compared models. It not only has the superior ability in water level prediction, but it also provides the evidences for the management of transportation in water-land, such as flood protection, and management of ship route.

Enhancement of Underwater Images with Retinex Transmission Map and Adaptive Color Correction

Underwater images often suffer from low contrast, low visibility, and color deviation. In this work, we propose a hybrid underwater enhancement method consisting of addressing an inverse problem with novel Retinex transmission map estimation and adaptive color correction. Retinex transmission map estimation does not rely on channel priors and aims to decouple from the unknown background light, thus avoiding error accumulation problem. To this end, global white balance is performed before estimating the transmission map using multi-scale Retinex. To further improve the enhancement performance, we design the adaptive color correction which cleverly chooses between two color correction procedures and prevents channel stretching imbalance. Quantitative and qualitative comparisons of our method and state-of-the-art underwater image enhancement methods demonstrate superiority of the proposed method. It achieves the best performance in terms of full-reference image quality assessment. In addition, it also achieves superior performance in the non-reference evaluation.

A Parallel Principal Skewness Analysis and Its Application in Radar Target Detection

Radar is often affected by various clutter backgrounds in complex environments, so clutter suppression has important practical significance for radar target detection. The clutter suppression process conforms to the blind source separation (BSS) model. The principal skewness analysis (PSA) algorithm is a BSS algorithm with third-order statistics as the objective function, and its running speed is faster than the conventional BSS algorithm. Still, the PSA algorithm has the problem of error accumulation. This paper improves the PSA algorithm and proposes a parallel PSA (PPSA) algorithm. PPSA can estimate the directions corresponding to each independent component simultaneously and avoid the problem of error accumulation. PPSA uses parallel instead of serial computing, significantly improving the running speed. In this paper, the PPSA algorithm is applied to radar target detection. The simulation data and real data experiments verify the effectiveness and superiority of the PPSA algorithm in suppressing clutter.

Storage Life Prediction of Rubber Products Based on Step Stress Accelerated Aging and Intelligent Algorithm.

Compared with the constant stress accelerated aging test, the step stress accelerated aging test reduces the accelerated aging test time by increasing the aging temperature step by step to obtain the aging failure life of rubber in a shorter time, but its data processing method is not mature enough. In this paper, a simplified step is proposed to process the step stress accelerated aging data. The identification of the acceleration factor is transformed into an optimization problem to avoid the error accumulation problem caused by fitting the data at each temperature. Considering the non-Arrhenius phenomenon in the rubber aging process, a modified Arrhenius equation was used to extrapolate the acceleration factor at low temperatures to calculate the prediction curves for the degradation of polyurethane rubber properties at low temperatures. The life prediction results of the constant stress accelerated aging test and step stress accelerated aging test were compared, and the dispersion coefficient between the two results was between 0.9 and 1. The results obtained by the two methods were in good agreement, which proved the correctness and feasibility of the method used in this paper.

A hierarchical structure built on physical and data-based information for intelligent aero-engine gas path diagnostics

Future trends in engine health management (EHM) systems are information fusion, advanced analytical methods, and the concept of the Intelligent Engines. Machine Learning (ML)-based aero-engine gas path diagnostic methods are promising under the motivation of these trends. However, previous ML-based diagnostic structures are rarely applied in actual engineering practice because they are purely mathematical and lack physical insight or are limited by the error accumulation problem. Developing an accurate, flexible and interpretable intelligent diagnostic method has always posed a challenge, especially when physical knowledge is also available for more diagnostic information. Instead of modifying and applying existing ML methods for classification or regression, this study proposes a novel hierarchical diagnostic method to get insight into the physical systems, build hierarchies automatically, and recommend the classification structures. The proposed hierarchical diagnostic method is evaluated against a NASA model high-bypass two-spool turbofan engine. NASA's blind test case results show that Kappa Coefficient of the proposed hierarchical diagnostic method is 0.693 and is at least 0.008 higher than the other diagnostic methods in the open literature. It has been proved that the proposed method can quantify the dependence relationships between the fault classes for enhanced diagnostic information, recommend the best diagnostic structure for reduced complexity, and solve the error accumulation problem for improved diagnostic accuracy. The proposed method could support intelligent condition monitoring systems by effectively exploiting physical and data-based information for improved model interpretability, model flexibility, diagnostic visibility, diagnostic accuracy, and diagnostic reliability.

Application of Knowledge Graph in Water Conservancy Education Resource Organization under the Background of Big Data

The key to improving the readability and usage of educational resources is their orderly arrangement and integration. Knowledge graphs, which are a large-scale form of knowledge engineering, are an effective tool for managing and organizing educational resources. The water conservancy’s educational big data is separated into three tiers of objectives–courses–knowledge units based on the connotation level of self-directed learning. Combined with the idea of Outcome-based Education(OBE), the goal-oriented knowledge graph structure of water conservancy disciplines and graph creation method is proposed. The focus is the error accumulation problem brought about by the traditional relational extraction method of Named Entity Recognition based on rules or sequence labeling. We first complete this objective, and then the relationship classification is performed according to the water conservancy disciplines entity and relations joint extraction (WDERJE) model, on which the prompt mechanism design is based. Think of the entity-relationship extraction task as a sequence-to-sequence generation task, and take the structured extraction language to unify the coding entity extraction and relationship extraction structures. The evaluation results of the WDERJE model show that the F_0.5 value of each entity extraction is above 0.76, and the cumulative extraction relationship triple is nearly 180,000. The graph fully optimizes the organization and management of water conservancy education resources and effectively improves the readability and utilization rate of water conservancy teaching resources.

Inertial Indoor Pedestrian Navigation Based on Cascade Filtering Integrated INS/Map Information.

Indoor pedestrian positioning has been widely used in many scenarios, such as fire rescue and indoor path planning. Compared with other technologies, inertial measurement unit (IMU)-based indoor positioning requires no additional equipment and has a lower cost. However, IMU-based indoor positioning has the problem of error accumulation, resulting in inaccurate positioning. Therefore, this paper proposes a cascade filtering algorithm to correct the accumulated error using only a small amount of map information. In the lower filter, the zero-velocity correction and the attitude-extended complementary filtering (ECF) algorithm are utilized to initially solve the pedestrian's trajectory. In the upper filter, a particle filter (PF) combined with the map information is adopted to correct the accumulated error of the heading and stride length. In the 2D positioning process, the root mean square error (RMSE) of the proposed algorithm is only 1.35 m. In the altitude correction, this paper proposes a method of clustering floor discrimination to deal with the instability of the barometer resulting from an uneven pressure and temperature. In the final 3D positioning experiment, with a total length of 536.5 m and including the process of going up and down the stairs, the end-point error is only 2.45 m by the proposed algorithm.

Efficient Distributed Particle Filter for Robust Range-Only SLAM

Compared with simultaneous localization and mapping (SLAM) problems based on Lidar or visual sensors, range-only SLAM (RO-SLAM) is lacking bearing information. It brings challenges to particle sampling and SLAM estimation. This article proposes an efficient distributed particle filter (EDPF) for RO-SLAM problems. To overcome the difficulties of sampling in a high-dimensional state space, EDPF is decomposed into a set of subparticle filters (sub-PFs) with low dimensionality. Each sub-PF corresponds with an observed beacon, which directly reduces the sampling complexity. A joint weight update method is proposed to exploit the correlation among sub-PFs. It reweighs the particles via an auxiliary distribution after each sub-PF’s local filtering and embodies a better proposal distribution for distributed filtering. In addition, a beacon diagnosis method is proposed, it can detect and reinitialize the wrong converged beacon position estimates, which further reduces the SLAM error accumulation problem. We consider a RO-SLAM system with an odometer and ultrawideband (UWB) to verify the proposed EDPF. Results show that EDPF outperforms many existing RO-SLAM methods, which obtains the best performance with the acceptable computational load.

A Universal Framework of Spatiotemporal Bias Block for Long-Term Traffic Forecasting

Recent studies have demonstrated the great success of graph convolutional networks in short-term traffic forecasting (e.g., 15–30 min ahead) tasks by capturing dependencies in road network structure. Based on these models, long-term forecasting can be achieved by two approaches: (1) recursively generating a one-step-ahead prediction and (2) adapting the models to sequence-to-sequence (seq2seq) learning. However, in practice, these two approaches often show poor performance in long-term forecasting tasks. The recursive approach suffers from the error accumulation problem, as the model is trained based on one-step-ahead loss. On the other hand, seq2seq shows convergence issues that limit its application. To address the issues for long-term forecasting, in this paper, we propose a universal framework that directly transforms any existing state-of-the-art models for one-step-ahead prediction to achieve more accurate long-term forecasting. The proposed framework consists of two components—a base model and a bias block. The base model is assumed to be a well-trained state-of-the-art one-step-ahead forecasting model, and the bias block is constructed by a spatiotemporal graph neural network composed of gated temporal convolution layers and graph convolution layers. The base model and the bias block are residually-connected so that we can substantially reduce the training complexity. Extensive experiments are conducted on existing benchmark datasets. We experiment with several state-of-the-art models in the literature as base models, and our results demonstrate the ability of the proposed universal framework to greatly improve the long-term prediction accuracy for all models.

A Dual Fuzzy-Enhanced Neurodynamic Scheme for Model-Less Kinematic Control of Redundant and Hyperredundant Robots

Tracking control of redundant and hyperredundant manipulators is a fundamental and critical problem in practical applications. In order to effectively decrease the end-effector position errors, a novel dual fuzzy-enhanced neurodynamic (DFEN) scheme is put forward for solving the position error accumulation problem followed by achieving accurate tracking control results. The proposed scheme is established based on a zeroing neurodynamic approach in conjunction with two fuzzy adjustment units that are capable of tuning the control parameters by monitoring the tracking error. Moreover, the DFEN scheme can effectively solve the tracking problem without requiring knowing <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">a priori</i> knowledge of the kinematic model of the robot. The convergence and the stability of the proposed approach are demonstrated by theoretical analysis. The effectiveness, accuracy, and robustness of the proposed DFEN scheme are verified on the simulative redundant manipulator, continuum robot, and hybrid robot (integrating the redundant manipulator and the continuum robot). A practical experiment is provided to validate the proposed scheme as well.

Time Series Forecasting of Motor Bearing Vibration Based on Informer.

Electric energy, as an economical and clean energy, plays a significant role in the development of science and technology and the economy. The motor is the core equipment of the power station; therefore, monitoring the motor vibration and predicting time series of the bearing vibration can effectively avoid hazards such as bearing heating and reduce energy consumption. Time series forecasting methods of motor bearing vibration based on sliding window forecasting, such as CNN, LSTM, etc., have the problem of error accumulation, and the longer the time-series forecasting, the larger the error. In order to solve the problem of error accumulation caused by the conventional methods of time series forecasting of motor bearing vibration, this paper innovatively introduces Informer into time series forecasting of motor bearing vibration. Based on Transformer, Informer introduces ProbSparse self-attention and self-attention distilling, and applies random search to optimize the model parameters to reduce the error accumulation in forecasting, achieve the optimization of time and space complexity and improve the model forecasting. Comparing the forecasting results of Informer and those of other forecasting models in three publicly available datasets, it is verified that Informer has excellent performance in time series forecasting of motor bearing vibration and the forecasting results reach ∼.

Dynamic Response Calculation Algorithm for Floating Offshore Wind Turbines Based on a State-Space Method of Transfer Function

The Cummins equation is commonly applied to describe the motion of floating structures. Because of the convolution term, the efficiency of calculation is low and the error accumulation problems in numerical integration calculation are serious. For those reasons, a dynamic response analysis method of floating structures based on a state-space model is proposed. The method uses the state-space method to replace the force-to-motion relationship of the Cummins equation, that is, transfer function, which transforms the problem of calculating the dynamic response into obtaining the output of the state-space model. The method is easy to implement and avoids the problems caused by the presence of the convolution term.

A Target Deblurring Method for Frequency Diverse Array Imaging

The frequency diversity array (FDA) is subject to its own property, which will cause the problem of target blurring in imaging. The existing FDA deblurring target imaging method generally adopts the design of nonlinear frequency increments, but it also brings new problems associated with sidelobe increases. Aiming at the problem of frequency diversity array image deblurring, this letter proposes an FDA deblurring target imaging method. Starting from the mechanism through which the blur is caused by the mutual influence of frequency increments, a mathematical model of image domain blur is established. A deconvolution method is proposed to achieve high-quality imaging without blur under arbitrary frequency increments of FDA. What`s more, a low-error imaging method based on ripple transfer is proposed to deal with the problem of error accumulation in deblurring. Finally, the effectiveness of the algorithm is verified by experimental simulation.

Added Objective Guided Optimization of Adversarial Text Generation Method

The problem of error accumulation is caused by the supervision of deep neural network text generation model. In &#x0D; order to solve this problem, a text generation model based on the reinforcement of antagonistic thought training is proposed.The &#x0D; adversarial network can be generated by the proposed model, and then the adversarial network can be used for identification, &#x0D; the learning reward function can be optimized, and the generated model can be optimized to reduce the probability of error &#x0D; accumulation.More text structure knowledge can be added into the generated text model by integrating the target guidance feature &#x0D; into the actual generation process to make the generated text model have higher authenticity. In this paper, the author optimizes &#x0D; the adversarial text generation method on the basis of target-guided optimization, which can be used for reference by practitioners.

Multi-step solar irradiation prediction based on weather forecast and generative deep learning model

With the rapid development of computer technology, more and more deep learning models are used in solar radiation (irradiation) prediction. There have been a lot of studies discussing the research of this type of model. However, how to better apply the deep learning model in the optimization method of building energy system, such as multi-step solar radiation (irradiation) prediction model in model predictive control (MPC), is still a challenging issue due to the complexity of the time series and the accumulation of errors in multi-step forecasts.In this research, a deep generative model based on LSTM is developed for multi-step solar irradiation prediction at least 24 h in the future. Measured data and temperature forecast data from the Tokyo Meteorological Agency were used for training and testing in this experiment. The results show that generating the model first can effectively avoid the problem of error accumulation. The generative model can produce an accuracy improvement of 7.7% against traditional regression LSTM model. Secondly, the introduction of the temperature forecast data from the previous one day can increase the forecast accuracy by about 18% points. When the earlier temperature forecast is used, the forecast accuracy will gradually decrease, and the use of the temperature forecast released 3 days before can hardly improve the forecast effect. In the end, using hourly temperature forecasts will result in better forecast accuracy than using daily temperature forecasts.

MLTF: Model less time-series forecasting

Multi-step ahead long term forecasting remains a pertinent challenge in time series literature due to non-stationary behaviour of real-world data. Predominantly most traditional time series models are parametric in nature and they use the predicted values to generate forecast for future time steps. This leads to error accumulation in each step of the forecasting horizon which causes increasingly poorer forecast in long-term. Other than the problem of error accumulation, most parametric algorithms also require significant pre-processing, hyper-parameter tuning, training and post-processing which can often put high computational burden on the system. Therefore, this paper proposes, Model Less Time-series Forecasting (MLTF), a non-parametric approach for forecasting which does not require any pre-processing or traditional training (i.e. Backpropagation). MLTF is a non-parametric method which uses statistical representations such as trend, linearity, entropy etc. to cluster series from a pre-defined repository and the series from same cluster are tagged as similar series. The trajectory of the target series is extracted from these similar series after applying an adaptive re-sampling technique. There is minimal training involved in MLTF, therefore this framework is computationally very efficient. The model-less nature also enables it to not suffer from error accumulation in long-horizon forecast. MLTF is validated empirically with a rich set of experiments involving M1, M3 competition dataset, Electricity, Volatility and COVID-19 data (over 4500 independent uni-variate series of different frequencies i.e. Hourly, Daily, Monthly, Quarterly and Yearly). The experiments demonstrate that, MLTF is significantly faster while being similar (or better) in terms of forecasting accuracy than the state-of-the-art DL methods and other non-parametric time series model.

Pole-residue method for dynamic response analysis of floating structures depending on decoupled Cummins equation

This study proposes a new dynamic response estimation algorithm for floating structures based on the pole-residue method, which can be used to calculate the dynamic response induced by the wave force and the transient response created by initial conditions. The proposed algorithm deals with the Cummins equation in the Laplace domain, which can avoid the integral computation of the convolution items and the error accumulation problem in the process of numerical integration. One theoretical improvement is that the transfer function obtained by the Laplace transform is decoupled and the corresponding poles and residues are calculated. Consequently, the coupling system of the floating body and fluid is replaced with the calculated poles and residues, which greatly simplifies the procedure for dynamic response analysis. Furthermore, the method avoids the integral errors of the time-domain method and the application limits of the frequency-domain method, such as energy leakage, periodic hypothesis, and allowing only the stable response calculation. Three test cases were employed to evaluate the performance of the proposed algorithm. The first case is a single degree of freedom (SDOF) system in the form of the Cummins equation. The example shows the implementation procedure of the proposed algorithm in detail. The results demonstrate that the proposed algorithm is able to simulate the dynamic response of the SDOF system accurately, and that its calculation efficiency is better than that of the Newmark-β method. The second case is a spar-type floating offshore wind turbine. Studies show that the dynamic responses estimated by the proposed algorithm match well with those of the time-domain method both in the stable and transient response analysis, and that the proposed algorithm has a better calculation accuracy. The final case is a Floating Production Storage and Offloading unit (FPSO) model. The studies of the FPSO model indicate that the estimated dynamic response obtained applying the proposed algorithm agrees well with that of the WASIM codes of SESAM, which further proves the validity of the proposed algorithm.