Feedforward Method Research Articles

Road surface semantic segmentation for autonomous driving

Although semantic segmentation is widely employed in autonomous driving, its performance in segmenting road surfaces falls short in complex traffic environments. This study proposes a frequency-based semantic segmentation with a transformer (FSSFormer) based on the sensitivity of semantic segmentation to frequency information. Specifically, we propose a weight-sharing factorized attention to select important frequency features that can improve the segmentation performance of overlapping targets. Moreover, to address boundary information loss, we used a cross-attention method combining spatial and frequency features to obtain further detailed pixel information. To improve the segmentation accuracy in complex road scenarios, we adopted a parallel-gated feedforward network segmentation method to encode the position information. Extensive experiments demonstrate that the mIoU of FSSFormer increased by 2% compared with existing segmentation methods on the Cityscapes dataset.

Identification of Distorted Gamma-Ray Signature Patterns Using Digital Filtering and Auto-Associative Memory Implemented with a Hopfield Neural Network

The detection and identification of radioactive sources in search applications involve analyzing passive gamma-ray emissions from high-level radioactive materials. This process uses a mobile detector-spectrometer in a complex field test environment. Recently, the use of artificial intelligence for gamma-ray spectrum analysis has shown promising results. However, challenges persist in identifying isotopic signatures from spectral measurements that may be distorted due to source shielding, random variations in natural radioactive background, or insufficient measurement time to obtain clear spectral lines. This paper presents a novel intelligent signature recognition method that combines digital filtering techniques with an artificial Hopfield Neural Network (HNN). The HNN leverages auto-associative memory to store training sample patterns and match them with incoming gamma spectra from distorted sources. It restores the testing sources’ measurements by finding the closest matching signature patterns in the spectral library. Before HNN recognition, the measured spectrum undergoes preprocessing with a digital image filter to reduce fluctuations. Performance of the proposed method is evaluated using a set of gamma-ray spectra measured with a sodium iodide detector. The data collected include measurements from six pure samples: 241Am, 60Co, 137Cs, 192Ir, 239Pu, and 235U, which are used for training and validation (i.e. six cases). Additionally, the data set contains 24 distorted synthesized sources with various fluctuating backgrounds. Test results demonstrate the potential of the proposed method to accurately recognize the correct isotope with high precision, achieving an accuracy rate exceeding 85%. Furthermore, the proposed method exhibits superior performance compared to the conventional multiple regression fitting and simple feedforward neural network methods.

Novel multi-degree-of-freedom force–displacement mixed control framework for low loading rate substructure hybrid testing

In substructure hybrid tests, imposing realistic multi-degree-of-freedom (MDOF) displacement and force boundary conditions on a specimen involves controlling several actuators to emulate the MDOF boundary conditions of the experimental substructure. This setup presents significant challenges for decoupling control in MDOF systems. Due to the typically high stiffness of the specimen in the axial direction, force control is preferred over displacement control, which has limited accuracy in this direction. Typically, force control in the axial direction is combined with simultaneous displacement control in the softer direction, forming an MDOF force-displacement mixed control scheme. This scheme transforms force control into displacement control in the inner loop in the axial direction to ensure safe loading. The scheme involves addressing two major challenges: the nonlinear force–displacement transformation and the nonlinear coordinate transformation from the local actuator space (LAS) to the global Cartesian coordinate (GCC) system. To address these challenges, the Modified Newton–Raphson iteration integrated with proportional integral plus feed-forward (MNR-PIFF) and incremental kinematic transformation integrated with proportional integral plus feed-forward (IKT-PIFF) methods are proposed. Based on these methods, a novel MDOF force–displacement mixed control (NMFDC) framework is developed for substructure hybrid testing. A 6-DOF cyclic test simulation using the NMFDC framework was conducted to demonstrate its superiority. Furthermore, a 6-DOF cyclic test and substructure hybrid test at low loading rates were performed to validate the NMFDC framework because of the loading equipment limit. Experimental results and numerical simulations indicate that the NMFDC framework achieves swifter response with lower error levels. For applications in fast and real-time hybrid testing, dynamic loading equipment need to be used, the PIFF controller in the NMFDC framework should be replaced by adaptive time delay compensation in future developments.

Hydrodynamic response on trajectory tracking of a tethered underwater robot system under hybrid control algorithms of umbilical cable and propellers

A hydrodynamic control model is established in attention to coupling relationships among cable, propellers and robot body of a tethered underwater robot system. In this model the governing equations of umbilical cable and the robot are firstly introduced, then supplementary conditions are coupled into the equations to forming the dynamic mathematical model; finally a hybrid control strategy based on feed-forward negative feedback method for the cable and PID rule for the propellers are integrated in the mathematical model for composing the whole hydrodynamic control model. Both the mathematical model and the control algorithms are proved to be effective and reliable through comparing simulation with the experimental data in existed references. Based on the numerical model constructed in this paper, trajectory tracking of a tethered underwater robot system in different motion combinations are numerically simulated through computational fluid dynamics method. In the numerical simulations, finite difference method is used to solving the kinematic parameters of the mathematical model, while finite volume method is applied on calculating the hydrodynamic forces under a hybrid control manipulations. The robot motion in vertical direction is determined primary by feed-forward negative feedback strategy of adjusting the cable length, while the horizontal movement of the robot is controlled mainly through PID algorithm; The hydrodynamic loading on the robot body are influenced by the flow fields around the robot.

Speed regulation system based on proportional resonance self-coupling PI control

In this paper, an improved self-coupled PI control dual-loop control strategy is proposed for the vector control of permanent magnet synchronous electric speed control system, to improve the poor motor control effect caused by the dead zone of the inverter, the presence of harmonics in the motor magnetic field distribution, and the error in the sampling process of the current. In order to reduce the tracking error of the signal, an acceleration feed-forward method is adopted, and the self-coupled PI controller with an additional Levant differentiator is introduced into the speed loop of the permanent magnet synchronous motor to ensure the accurate derivation of the input signal containing noise to improve the robustness of the system. In addition, this strategy combines the proportional term of the autotuned PI control structure with the standard resonator to form a proportional resonant autotuned PI in the current loop, so that the permanent magnet synchronous motor can better compensate the current in the current loop control to achieve the effect of current harmonic suppression. Finally, the control mechanism of the servo system analyzed, and based on the mathematical model of the controlled object PMSM, the design method of the improved self-coupling PI control structure is given. The experimental results show that the proposed control strategy can reduce the impact of the motor's own parameter changes on the system performance and has a good current harmonic suppression in the low-speed domain, which verifies the effectiveness of the designed improved self-coupled PI controller.

Linearizing AFM raster scans acquired using dual-stage lateral scanners

Dual-stage scanners can successfully increase the range of atomic force microscope (AFM) scanners in space and/or frequency by designing and constructing two nanopositioners with significant differences in static and dynamic characteristics. In such cases, the positioning signal has to be split to meet the displacement and frequency limitations of each stage. Without closed-loop displacement measurement sensors, linearizing images acquired using signals split in time and frequency can be challenging. A common method, often cumbersome and time consuming, uses inverse model-based feedforward compensation. In this work, we show that image-based linearization can be used to achieve acceptable results for raster scans acquired using dual-stage scanners. In particular, we apply the feedforward and image-based methods to our homemade dual-stage lateral scanner and high-speed AFM (HS-AFM) system. The acquired scans compare well for the two methods, indicating that image-based raster scan linearization can be used in place of inverse model-based feedforward approaches.

The Southern Ocean carbon sink has been overestimated in the past three decades

Employing machine learning methods for mapping surface ocean pCO2 has reduced the uncertainty in estimating sea-air CO2 flux. However, a general discrepancy exists between the Southern Ocean carbon sinks derived from pCO2 products and those from biogeochemistry models. Here, by performing a boosting ensemble learning feed-forward neural networks method, we have identified an underestimation of the surface Southern Ocean pCO2 due to notably uneven density of pCO2 measurements between summer and winter, which resulted in about 16% overestimating of Southern Ocean carbon sink over the past three decades. In particular, the Southern Ocean carbon sink since 2010 was notably overestimated by approximately 29%. This overestimation can be mitigated by a winter correction in algorithms, with the average Southern Ocean carbon sink during 1992-2021 corrected to −0.87 PgC yr−1 from the original −1.01 PgC yr−1. Furthermore, the most notable underestimation of surface ocean pCO2 mainly occurred in regions south of 60°S and was hiding under ice cover. As the surface ocean pCO2 under sea ice coverage in the winter is much higher than the atmosphere, if sea ice melts completely, there could be a further reduction of about 0.14 PgC yr−1 in the Southern Ocean carbon sink.

Challenges and Solutions in High-Speed SerDes Data Path Design

Modern high-performance communication systems require high-speed Serializer/Deserializer (SerDes) data path design to meet the growing demand for data throughput and reliability due to technological advancements and data-intensive applications. This research article examines the complex problems of designing high-speed SerDes data routes and suggests novel solutions. Data fidelity across high-speed channels requires signal integrity, the first key difficulty. Attenuation, crosstalk, and EMI may degrade SerDes signal performance. Advanced methods like equalization, impedance matching, and effective shielding mitigate these impacts. Equalization, including feedforward and feedback methods, reduces channel losses and distortions, enhancing signal quality. Power consumption is another major issue as data rates and system complexity rise. Power-hungry high-speed SerDes circuits cause thermal management and energy efficiency difficulties. Adaptive power management, low-power design, and new materials with higher thermal performance and lower power dissipation are discussed in the study. Design complexity is another issue in high-speed SerDes data route design. High-precision timing and synchronization and many component integration in a small area need advanced design tools and methods. The study emphasizes simulation and modeling for complexity management and design for testability (DFT) for full verification and debugging.

An artificial neural network and SCS–CN-based model for runoff estimation: a case study of the Peddavagu watershed

ABSTRACT The accurate prediction of runoff from rainfall events is crucial for effective water resource management, especially in regions with diverse climatic patterns like India. This study proposes a novel approach by integrating the soil conservation service (SCS)–curve number (CN) method with artificial neural networks (ANNs) to model rainfall–runoff relationships. In this research, an SCS–CN method is utilized to estimate initial runoff volumes, accounting for local soil and land-cover characteristics. Subsequently, an ANN-based model is developed to capture the complex nonlinear relationships between rainfall inputs and runoff outputs. Taking the rainfall of the watershed as the inputs, an ANN model was developed in MATLAB for runoff simulation at the main outlet of the Peddavagu watershed. A feed-forward back-propagation method is employed in the ANN model. The architecture with a 1-10-1 configuration based on tan–sig transfer function performed well in terms of MSE 0.019. The model efficiency was satisfactory with the coefficient of correlation (R), 0.96 for training, 0.98 for validation, and 0.98 for testing period. The overall value of R (0.96) indicates the utility of this ANN–SCS-based coupled model for rainfall–runoff simulation.

PCa-RadHop: A transparent and lightweight feed-forward method for clinically significant prostate cancer segmentation

Prostate Cancer is one of the most frequently occurring cancers in men, with a low survival rate if not early diagnosed. PI-RADS reading has a high false positive rate, thus increasing the diagnostic incurred costs and patient discomfort. Deep learning (DL) models achieve a high segmentation performance, although require a large model size and complexity. Also, DL models lack of feature interpretability and are perceived as “black-boxes” in the medical field. PCa-RadHop pipeline is proposed in this work, aiming to provide a more transparent feature extraction process using a linear model. It adopts the recently introduced Green Learning (GL) paradigm, which offers a small model size and low complexity. PCa-RadHop consists of two stages: Stage-1 extracts data-driven radiomics features from the bi-parametric Magnetic Resonance Imaging (bp-MRI) input and predicts an initial heatmap. To reduce the false positive rate, a subsequent stage-2 is introduced to refine the predictions by including more contextual information and radiomics features from each already detected Region of Interest (ROI). Experiments on the largest publicly available dataset, PI-CAI, show a competitive performance standing of the proposed method among other deep DL models, achieving an area under the curve (AUC) of 0.807 among a cohort of 1,000 patients. Moreover, PCa-RadHop maintains orders of magnitude smaller model size and complexity.

Intelligent PID Controller Based on Neural Network for AI-Driven Control Quadcopter UAV

Unmanned Aerial Vehicle (UAV), specifically a quadcopter is publicly popular which it provides services in different applications such as aerial delivery, aerial photography, military, weather forecasting and more examples to date. A Proportional-Integral-Derivative (PID) controller is one of the control techniques that can provide stabilization and reliable trajectory tracking. However, proper PID gains are needed to ensure a stable flight and it should be hybridized or improved to increase the robustness, reliability, and stabilization during flight. In this paper, an intelligent PID controller using neural network is proposed based on Levenberg-Marquardt feedforward neural network training method. The PID gains are initialized using different ranges according to the optimal gains generated by Particle Swarm Optimization, and this contributes towards a good training performance using Mean Square Error (MSE) evaluation. The trained network takes desired output and references as input data to calculate the required combination of PID gains as the output. The including of the response characteristics as the input data for the network, together with reference, error, and control input is the significance of the work. The performance of this work is presented using MSE performances, attitudes and altitude stabilization, and trajectory tracking reliability through error index performances. The simulation results graphically prove that the proposed controller provides better stability with reduced overshoot and settling times. Disturbance rejection is also enhanced by 1.7% compared to manual tuned PID controller. The reliability of the proposed controller highlights avenues for further exploration in AI-driven control strategies for quadcopter systems.

This is my story: An early career professor’s experience using a feedforward teaching method

This is my story: An early career professor’s experience using a feedforward teaching method

FPRF: Feed-Forward Photorealistic Style Transfer of Large-Scale 3D Neural Radiance Fields

We present FPRF, a feed-forward photorealistic style transfer method for large-scale 3D neural radiance fields. FPRF stylizes large-scale 3D scenes with arbitrary, multiple style reference images without additional optimization while preserving multi-view appearance consistency. Prior arts required tedious per-style/-scene optimization and were limited to small-scale 3D scenes. FPRF efficiently stylizes large-scale 3D scenes by introducing a style-decomposed 3D neural radiance field, which inherits AdaIN’s feed-forward stylization machinery, supporting arbitrary style reference images. Furthermore, FPRF supports multi-reference stylization with the semantic correspondence matching and local AdaIN, which adds diverse user control for 3D scene styles. FPRF also preserves multi-view consistency by applying semantic matching and style transfer processes directly onto queried features in 3D space. In experiments, we demonstrate that FPRF achieves favorable photorealistic quality 3D scene stylization for large-scale scenes with diverse reference images.

Development of a novel nonlinear model and control strategy for soft continuum robots featuring hard magnetoactive elastomers

Magnetoactive soft continuum robots (MSCRs), capable of controllable steering and navigation, hold substantial promise for healthcare applications. However, advancements in MSCRs have been hindered by a limited understanding of MSCR dynamics and a lack of effective control methods. Addressing these gaps, this study presents a novel, time-dependent, and computationally efficient analytical model of MSCR, alongside a new optimal closed-loop control strategy for precise high-frequency trajectory tracking. A finite element (FE) model of the MSCR is initially developed, with its validity confirmed through rigorous laboratory measurements. Using the formulated FE model, a new and computationally efficient analytical model is subsequently developed to accurately predict the highly nonlinear response of MSCR. This model operates as a system of switched linear models, each of which is a reduced-order version of its corresponding high-order linear model extracted from the FE analysis. This innovative approach not only maintains the predictive accuracy of the FE model but also significantly reduces computational demands, operating in just a few seconds. The results highlight that the developed model can accurately predict the dynamic responses of the MSCR while significantly reducing the computational load by almost 80 orders of magnitude compared with the FE model on the same simulation platform. The proposed model has been effectively utilized to develop a novel optimal control strategy using the feedforward interval type-2 fractional-order fuzzy-PID method. A hardware-in-the-loop experimental test has been finally designed to demonstrate the superior performance of the MSCR under the proposed controller.

Prediksi Harga Kelapa Sawit Menggunakan Metode Extreme Learning Machine

Palm oil is one of the keys to the Indonesian economy and the main commodity for attracting foreign investment. The palm oil and palm kernel industry generates most of the foreign currency from palm oil. The price of palm oil often goes up and down every month resulting in instability in the income received by people who own oil palm plantations. The aim of predicting palm oil prices is to carry out appropriate planning or steps for palm oil business actors. One way to overcome this problem is to make predictions. One method that can make predictions is the Extreme Learning Machine (ELM). ELM is an artificial neural network method used to predict palm oil prices. The ELM method is a feedforward method with a single hidden layer which is better known as a single hidden layer feedforward neural network (SLFNs). In this research, the best implementation was 5 inputs with 20 neurons in the hidden layer with output in the form of palm oil price predictions. Based on the tests carried out, the research produced the smallest error rate of 0.0027111424247658633 using 20 neurons in the hidden layer so that the latest data prediction test results for 5 price rotations in September rotation 1 were 1400.314191, September rotation 2 were 1846.798921, September rotation 3 amounted to 1505.430419, September rotation 4 amounted to 2301.853412, September rotation 5 amounted to 2645.082489 in palm oil price predictions.

Stability Enhancement Method of Standalone Modular Multilevel Converters Based on Impedance Reshaping

Modular multilevel converters (MMCs) are susceptible to subsynchronous oscillations (SSOs) caused by impedance interactions in the power line. Current research into the stability of MMCs focuses mainly on voltage feed-forward control, while the effect of current feed-forward control is neglected. This paper proposes a current feed-forward compensation method based on impedance reshaping for standalone MMCs. Initially, an impedance model was developed to identify the stability risks caused by the interaction between the MMC and power line impedance. The proposed method feeds the current compensation signal into the modulation circuit, thereby improving the control signal and suppressing the impedance interaction between the MMC and the power line. The analysis of the harmonic state space (HSS) method verifies that the proposed approach effectively reduces the negative damping region in the frequency band where the SSO is located. Additionally, the impedance frequency scanning method confirms the accuracy of impedance modeling. Using the MATLAB/Simulink platform and StarSim HIL hardware-in-the-loop experimental platform, the SSO phenomenon of the MMC is simulated, and the results show that the proposed method can effectively suppress harmonic currents during SSO, which verifies the accuracy of the stability analysis and the feasibility of the proposed method.

A novel sea surface pCO2-product for the global coastal ocean resolving trends over 1982–2020

Abstract. In recent years, advancements in machine learning based interpolation methods have enabled the production of high-resolution maps of sea surface partial pressure of CO2 (pCO2) derived from observations extracted from databases such as the Surface Ocean CO2 Atlas (SOCAT). These pCO2-products now allow quantifying the oceanic air–sea CO2 exchange based on observations. However, most of them do not yet explicitly include the coastal ocean. Instead, they simply extend the open ocean values onto the nearshore shallow waters, or their spatial resolution is simply so coarse that they do not accurately capture the highly heterogeneous spatiotemporal pCO2 dynamics of coastal zones. Until today, only one global pCO2-product has been specifically designed for the coastal ocean (Laruelle et al., 2017). This product, however, has shortcomings because it only provides a climatology covering a relatively short period (1998–2015), thus hindering its application to the evaluation of the interannual variability, decadal changes and the long-term trends of the coastal air–sea CO2 exchange, a temporal evolution that is still poorly understood and highly debated. Here we aim at closing this knowledge gap and update the coastal product of Laruelle et al. (2017) to investigate the longest global monthly time series available for the coastal ocean from 1982 to 2020. The method remains based on a two-step Self-Organizing Maps and Feed-Forward Network method adapted for coastal regions, but we include additional environmental predictors and use a larger pool of training and validation data with ∼18 million direct observations extracted from the latest release of the SOCAT database. Our study reveals that the coastal ocean has been acting as an atmospheric CO2 sink of −0.40 Pg C yr−1 (−0.18 Pg C yr−1 with a narrower coastal domain) on average since 1982, and the intensity of this sink has increased at a rate of 0.06 Pg C yr−1 decade−1 (0.02 Pg C yr−1 decade−1 with a narrower coastal domain) over time. Our results also show that the temporal changes in the air–sea pCO2 gradient plays a significant role in the long-term evolution of the coastal CO2 sink, along with wind speed and sea-ice coverage changes that can also play an important role in some regions, particularly at high latitudes. This new reconstructed coastal pCO2-product (https://doi.org/10.25921/4sde-p068; Roobaert et al., 2023) allows us to establish regional carbon budgets requiring high-resolution coastal flux estimates and provides new constraints for closing the global carbon cycle.

NEURAL NETWORK BASED ESTIMATION OF ECOLOGICAL FOOTPRINT IN TURKEY WITH ECONOMIC AND ENVIRONMENTAL FACTORS

The globe is currently confronted with escalating ecological challenges as a result of the exponential growth of the population, the expansion of industrial activities, the rapid urbanization process, and the heightened levels of consumption. Due to the prevailing ecological issues, there has been an unbridled surge in the demand for natural resources. The emergence of the Ecological Footprint was driven by the observable escalation of environmental degradation, with the primary objective of promoting sustainability. In this study, an Artificial Neural Network (ANN) model was developed for the estimation of ecological footprint. In the first stage, the variables affecting the ecological footprint were determined. It was concluded that there are strong relationships between the independent variables of Gross Domestic Product (GDP), KOF Globalization Index (KOFGI), and Natural Resource Rent (NRR) and the dependent variable of ecological footprint. In the light of this information, the data of GDP (% annual increase), NRR (% GDP) and KOF Globalization Index and ecological footprint data of Turkey between 1970 and 2016 were used to present the ANN model. Feed-Forward Backprop Method, which is one of the multi-layer network models, was applied in the modeling of the ANN. Levenberg-Marquardt optimization, which updates the weight and bias values of the network, was used as the network training function. The study's results suggest that there was a strong correlation between the whole dataset and the model's accuracy, with a close match of 99.316%. Based on the findings, it can be deduced that the developed artificial neural network (ANN) model has a significant level of precision in forecasting the Ecological Footprint.

Application of machine learning approach (artificial neural network) and shrinking core model in cobalt (II) and copper (II) leaching process

The leaching laboratory experiment uses the artificial neural network (ANN) to predict and evaluate copper and cobalt recovery. This study aimed to evaluate the efficacy of using the shrinking core model in conjunction with an artificial neural network (ANN) as part of a machine learning strategy to improve the leaching process of cobalt (II) and copper (II). The numerous factors in the leaching process, such as acid concentration, leaching time, temperature, soil-to-solution ratio, and stirring speed, are adjusted using an ANN with several layers, feed-forward, and back-propagation learning methods. These variables are in charge of the high cobalt recovery during the reduced sulfuric acid leaching procedure. The ANN algorithm has 10 hidden layers, 5 input variables describing the leaching parameters, and two neurons as output layers corresponding to copper and cobalt leaching recovery. The optimum conditions were found to be acid concentration of 100 g/L, leaching duration 120 min, temperature 55 °C, soil-to-solution ratio of 1:40 g/mL, and stirring speed 300 rpm. The optimized trained neural networks tested, trained, and validated steps are represented by R 2 values of 0.94, 0.99, 0.97, and 0.97, respectively, equating to 97.5% copper recovery and 95.4% cobalt recovery.

Feed-forward networks using logistic regression and support vector machine for whole-slide breast cancer histopathology image classification

The performance of an image classification depends on the efficiency of the feature learning process. This process is a challenging task that traditionally requires prior knowledge from domain experts. Recently, representation learning was introduced to extract features directly from the raw images without any prior knowledge. Deep learning using a Convolutional Neural Network (CNN) has gained massive attention for performing image classification, as it achieves remarkable accuracy that sometimes exceeds human performance. But this type of network learns features by using a back-propagation approach. This approach requires a huge amount of training data and suffers from the vanishing gradient problem that deteriorates the feature learning. The forward-propagation approach uses predefined filters or filters learned outside the model and applied in a feed-forward manner. This approach is proven to achieve good results with small size labeled datasets. In this work, we investigate the suitability of using two feed-forward methods such as Convolutional Logistic Regression Network (CLR), and Convolutional Support Vector Machine Network for Histopathology Images (CSVM-H). The experiments we have conducted on two small breast cancer datasets (Sultan Qaboos University Hospital (SQUH) and BreaKHis dataset) demonstrate the advantage of using feed-forward approaches over the traditional back-propagation ones. On those datasets, the proposed models CLR and CSVM-H were faster to train and achieved better classification performance than the traditional back-propagation methods (VggNet-16 and ResNet-50) on the SQUH dataset. Importantly, our proposed approach CLR and CSVM-H efficiently learn representations from small amounts of breast cancer whole-slide images and achieve an AUC of 0.83 and 0.84, respectively, on the SQUH dataset. Moreover, the proposed models reduce memory footprint in the classification of Whole-Slide histopathology images since their training time is significantly reduced compared to the traditional CNN on the SQUH and BreaKHis datasets.