Search

This paper advances real-time cursor control for individuals with motor impairments through a novel brain–computer interface (BCI) system based solely on motor imagery. We introduce an enhanced deep neural network (DNN) classifier integrated with a Four-Class Iterative Filtering (FCIF) technique for efficient preprocessing of neural signals. The underlying approach is the Four-Class Filter Bank Common Spatial Pattern (FCFBCSP) and it utilizes a customized filter bank for robust feature extraction, thereby significantly improving signal quality and cursor control responsiveness. Extensive testing under varied conditions demonstrates that our system achieves an average classification accuracy of 89.1% and response times of 663 milliseconds, illustrating high precision in feature discrimination. Evaluations using metrics such as Recall, Precision, and F1-Score confirm the system's effectiveness and accuracy in practical applications, making it a valuable tool for enhancing accessibility for individuals with motor disabilities. [ABSTRACT FROM AUTHOR]

We introduce EFIT-Prime, a novel machine learning surrogate model for EFIT (Equilibrium FIT) that integrates probabilistic and physics-informed methodologies to overcome typical limitations associated with deterministic and ad hoc neural network architectures. EFIT-Prime utilizes a neural architecture search-based deep ensemble for robust uncertainty quantification, providing scalable and efficient neural architectures that comprehensively quantify both data and model uncertainties. Physically informed by the Grad–Shafranov equation, EFIT-Prime applies a constraint on the current density J tor and a smoothness constraint on the first derivative of the poloidal flux, ensuring physically plausible solutions. Furthermore, the spatial location of the diagnostics is explicitly incorporated in the inputs to account for their spatial correlation. Extensive evaluations demonstrate EFIT-Prime's accuracy and robustness across diverse scenarios, most notably showing good generalization on negative-triangularity discharges that were excluded from training. Timing studies indicate an ensemble inference time of 15 ms for predicting a new equilibrium, offering the possibility of plasma control in real-time, if the model is optimized for speed. [ABSTRACT FROM AUTHOR]

This study harnesses EEG signals to enable the real-time control of servo motors, utilizing the OpenBCI Community Dataset to identify and assess brainwave patterns related to motor imagery tasks. Specifically, the dataset includes EEG data from 52 subjects, capturing electrical brain activity while participants imagined executing specific motor tasks. Each participant underwent multiple trials for each motor imagery task, ensuring a diverse and comprehensive dataset for model training and evaluation. A deep neural network model comprising convolutional and bidirectional long short-term memory (LSTM) layers was developed and trained using k-fold cross-validation, achieving a notable accuracy of 98%. The model's performance was further compared against recurrent neural networks (RNNs), multilayer perceptrons (MLPs), and Τransformer algorithms, demonstrating that the CNN-LSTM model provided the best performance due to its effective capture of both spatial and temporal features. The model was deployed on a Python script interfacing with an Arduino board, enabling communication with two servo motors. The Python script predicts actions from preprocessed EEG data to control the servo motors in real-time. Real-time performance metrics, including classification reports and confusion matrices, demonstrate the seamless integration of the LSTM model with the Arduino board for precise and responsive control. An Arduino program was implemented to receive commands from the Python script via serial communication and control the servo motors, enabling accurate and responsive control based on EEG predictions. Overall, this study presents a comprehensive approach that combines machine learning, real-time implementation, and hardware interfacing to enable the precise and real-time control of servo motors using EEG signals, with potential applications in the human–robot interaction and assistive technology domains. [ABSTRACT FROM AUTHOR]

The engine main shaft forgings have high requirements for product consistency and reliability, which are difficult to be guaranteed by traditional manual free forging. On the basis of the already-in-place machinery, technical optimization was done in order to realize intelligent forging of the engine main shaft forgings. A rail trailer, holding robot, inspection robot, and expert system were installed as well as other hardware and software. Additionally, the procedure and specifications for robot intelligent free forging were revised. Based on the artificial neural network (ANN) model, an optimization model and a prediction model were created, and the process parameters can be controlled during forging. The verification result shows that the intelligent free forging production line can achieve real-time control of the shaft forging process, and obtain the forgings whose shape, size, microstructure, performance and consistency meet the requirements. With the help of this production line, free forging can be produced more quickly and efficiently, which is crucial for realizing the automation, digitization, and intelligence of shaft forging free forging. [ABSTRACT FROM AUTHOR]

In a previous research effort by this group, pseudo engine dynamometer data in multi-dimensional arrays were combined with dynamic equations to form a crank angle resolved engine model compatible with a real-time simulator. The combination of the real-time simulator and external targets enabled the development of a software-in-the-loop (SIL) environment that enabled near-real-time development of AI/ML and real-time deployment of the resulting AI/ML. Military applications, in particular, are unlikely to possess large quantities of non-sparse operational data that span the full operational range of the system, stove-piping the ability to develop and deploy AI/ML which is sufficient for near-real-time or real-time control. AI/ML has been shown to be well suited for predicting highly non-linear mathematical phenomena and thus military systems could potentially benefit from the development and deployment of AI/ML acting as soft sensors. Given the non-sparse nature of the data, it becomes exceedingly important that AI/ML be developed and deployed in near-real-time or real-time in parallel to a real-time system to overcome the inadequacy of applicable data. This research effort used parallel processing to reduce the training duration of the shallow artificial neural networks (SANN) and forest algorithms forming ensemble models. This research is novel in that the SIL environment enables pre-developed AI/ML to be adapted in near-real-time or develop AI/ML in response to changes within the operation of the applied system, different load torques, engine speeds, and atmospheric conditions to name a few. Over time it is expected that the continued adaptation of the algorithms will lead to the development of AI/ML that is suitable for real-time control and energy management. [ABSTRACT FROM AUTHOR]

Controlled de-orbiting is crucial for a low earth orbiting (LEO) satellite or capsule intending to land in a desired location and to prevent damage to people and property on the ground caused by debris. Drag modulation is one possible mechanism to control de-orbiting, exploiting atmospheric drag variation to reduce the necessary orbital energy from the initial conditions to the re-entry interface. In this context, this paper proposes a novel targeted de-orbiting Artificial Neural Network (ANN) and drag-based guidance algorithm for a LEO artificial satellite. It relies on the combination of empirical and analytical relations with Deep Neural Networks (DNNs) to estimate the main parameters of the optimal control law in real time. The training set is composed of several optimal control solutions generated with a previously developed optimal control algorithm, exploiting the formulation in equinoctial modified orbital parameters to manage the large time scale of the problem and the computational cost. An innovative procedure for the training set generation led to the achievement of general results with a limited number of samples. The successful outcome of the guidance algorithm on over 1000 cases demonstrates its robustness and generalization of results. To conclude, a Linear Quadratic Regulator (LQR) feedback control is applied to one case to deal with a more realistic and uncertain density model. • Analytical optimal guidance for a satellite de-orbiting from Low Earth Orbit. • A novel approach based on Deep Neural Networks. • Prediction of the key parameters of the optimal control law in real time. • A Monte Carlo analysis showed the algorithm's robustness and generalization. • Successful feedback control thanks to a Linear Quadratic Regulator (LQR). [ABSTRACT FROM AUTHOR]

The meaningful problem of improving crane safety, reliability, and efficiency is extensively studied in the literature and targeted via various model-based control approaches. In recent years, crane data-driven modeling has attracted much attention compared to physics-based models, particularly due to its potential in real-time crane control applications, specifically in model predictive control. This paper proposes grammar-guided genetic programming with sparse regression (G3P-SR) to identify the nonlinear dynamics of an underactuated crane system. G3P-SR uses grammars to bias the search space and produces a fixed number of candidate model terms, while a local search method based on an l0-regularized regression results in a sparse solution, thereby also reducing model complexity as well as reducing the probability of overfitting. Identification is performed on experimental data obtained from a laboratory-scale overhead crane. The proposed method is compared with multi-gene genetic programming (MGGP), NARX neural network, and Takagi-Sugeno fuzzy (TSF) ARX models in terms of model complexity, prediction accuracy, and sensitivity. The G3P-SR algorithm evolved a model with a maximum mean square error (MSE) of crane velocity and sway prediction of 1.1860 × 10−4 and 4.8531 × 10−4, respectively, in simulations for different testing data sets, showing better accuracy than the TSF ARX and MGGP models. Only the NARX neural network model with velocity and sway maximum MSEs of 1.4595 × 10−4 and 4.8571 × 10−4 achieves a similar accuracy or an even better one in some testing scenarios, but at the cost of increasing the total number of parameters to be estimated by over 300% and the number of output lags compared to the G3P-SR model. Moreover, the G3P-SR model is proven to be notably less sensitive, exhibiting the least deviation from the nominal trajectory for deviations in the payload mass by approximately a factor of 10. [ABSTRACT FROM AUTHOR]

The accurate online detection of laser welding penetration depth has been a critical problem to which the industry has paid the most attention. Aiming at the laser welding process of TC4 titanium alloy, a multi-sensor monitoring system that obtained the keyhole/molten pool images and laser-induced plasma spectrum was built. The influences of laser power on the keyhole/molten pool morphologies and plasma thermo-mechanical characteristics were investigated. The results showed that there were significant correlations among the variations of the keyhole–molten pool, plasma spectrum, and penetration depth. The image features and spectral features were extracted by image processing and dimension-reduction methods, respectively. Moreover, several penetration depth prediction models based on single-sensor features and multi-sensor features were established. The mean square error of the neural network model built by multi-sensor features was 0.0162, which was smaller than that of the model built by single-sensor features. The established high-precision model provided a theoretical basis for real-time feedback control of the penetration depth in the laser welding process. [ABSTRACT FROM AUTHOR]

The 'ideal' optimal control of a radiant floor heating system (RFHS) can be achieved when a central controller is fully aware of the interrelationship between the different thermal dynamics of a boiler, radiant floor, and room. This study proposed integrated control strategies that can consider the interaction of heterogeneous RFHSs using a minimalistic approach with minimal data easily available from residential buildings. Three individual artificial neural networks were developed to choreograph the heterogeneous dynamics of the RFHSs for each household. Then, model predictive control was applied to realize the real-time integrated control of the RFHSs in the residential buildings. Approximately 14–46% of heating energy could be saved by integrated control. Moreover, the potential energy saving rate from integrated control varied depending on the degree of human intervention. [ABSTRACT FROM AUTHOR]

With high efficiency and low energy loss, high-temperature superconductors (HTS) have demonstrated their profound applications in various fields, such as medical imaging, transportation, accelerators, microwave devices, and power systems. The high-field applications of HTS tapes have raised the demand for producing cost-effective tapes with long lengths in superconductor manufacturing. However, achieving the uniform and enhanced performance of a long HTS tape is challenging due to the unstable growth conditions in the manufacturing process. Although it is confirmed that the process parameters during the advanced metal organic chemical vapor deposition (A-MOCVD) process influence the uniformity of the produced HTS tapes, the high-dimensional process parameter signals and their complicated interactions make it difficult to develop an effective control policy. In this paper, we propose a local measure for the uniformity of HTS tapes to provide instant feedback for our control policy. Then, we model the manufacturing of HTS tapes as a Markov decision process (MDP) with continuous state and action spaces to assess the instant reward in real time in our feedback control model. As our MDP involves continuous and high-dimensional state and action spaces, a neural fitted Q-iteration (NFQ) algorithm is adopted to solve the MDP with artificial neural network (ANN) function approximation. The collinearity of process parameters can restrict our capability of adjusting the process parameters, which is addressed by the principal component analysis (PCA) in our method. The control policy adjusts the PCA of process parameters using the NFQ algorithm. Based on our case studies on real A-MOCVD dataset, the obtained control policy increases the average uniformity of tapes by 5.6% and performs especially well on sample HTS tapes with a low uniformity. [ABSTRACT FROM AUTHOR]

A rendezvous mission with a tumbling space object has been studied by developing a novel three-phase mission design architecture which includes model and state uncertainties by utilizing Ensemble Optimal Control and warm-start Bellman Pseudospectral Optimal Control. Ensemble optimal control problem is solved via in-house developed Stochastic-Collocation based Ensemble Pseudospectral Optimal Control Software (SC-EPOCS) by modeling the uncertainties via samples generated by Conjugate Unscented Transformation based on the initial statistical distributions. Then, a new concept called checkpoints has been introduced to decide a feasible and optimal transition trajectory from uncertainty-aware trajectories to a deterministic optimal trajectory. This is achieved by conducting a reachability analysis for all ensembles and uncertainties. As a result, two Deep Neural Networks (DNNs), one for translational and other for the rotational motion, have been trained by utilizing initial costate values that correspond to optimal trajectories, starting from the checkpoints. By deriving costate dynamics for optimal trajectories, a warm-start algorithm is developed by applying Covector Mapping Theorem. Finally, Bellman Pseudospectral Optimal Control is applied in a Caratheodory- Π sense by recursively solving an optimal control problem, subjected to state estimation errors. Results prove the feasibility, robustness, and optimality of the approach. • A novel robust trajectory optimization framework for rendezvous is proposed. • A robust trajectory database is generated via Conjugate Unscented Transformation. • Two Deep Neural Networks are trained for optimal transition trajectories. • Framework has demonstrated optimality, safety, robustness, and flexibility. [ABSTRACT FROM AUTHOR]

The product quality of injection‐molded plastic is closely related to the injection flow velocity of molten plastics. In this article, an optimal tracking control problem for the injection flow front position arising in the filling process in the injection molding machine (IMM) is considered, and an intelligent real‐time optimal control method based on deep neural networks (DNNs) is developed for the online tracking of the flow front position to improve the efficient production process of the plastics. To this end, a highly nonlinear dynamic model describing the filling process is first proposed and then an optimal control problem is formulated. A high‐fidelity numerical optimization approach known as the Gauss pseudospectral method is used to numerically obtain the optimal state‐control solutions, which can then be saved as a huge training data set for the DNNs. The DNNs architecture is then developed and offline trained using the obtained data set to determine the best state–control relationship. As a consequence, the well‐trained DNN may be utilized to produce the matching optimum feedback action on‐board. Numerical studies are also carried out to verify the performance of our proposed approach. [ABSTRACT FROM AUTHOR]

High latency in teleoperation has a significant negative impact on operator performance. While deep learning has revolutionized many domains recently, it has not previously been applied to teleoperation enhancement. We propose a novel approach to predict video frames deep into the future using neural networks informed by synthetically generated optical flow information. This can be employed in teleoperated robotic systems that rely on video feeds for operator situational awareness. We have used the image‐to‐image translation technique as a basis for the prediction of future frames. The Pix2Pix conditional generative adversarial network (cGAN) has been selected as a base network. Optical flow components reflecting real‐time control inputs are added to the standard RGB channels of the input image. We have experimented with three data sets of 20,000 input images each that were generated using our custom‐designed teleoperation simulator with a 500‐ms delay added between the input and target frames. Structural Similarity Index Measures (SSIMs) of 0.60 and Multi‐SSIMs of 0.68 were achieved when training the cGAN with three‐channel RGB image data. With the five‐channel input data (incorporating optical flow) these values improved to 0.67 and 0.74, respectively. Applying Fleiss' κ gave a score of 0.40 for three‐channel RGB data, and 0.55 for five‐channel optical flow‐added data. We are confident the predicted synthetic frames are of sufficient quality and reliability to be presented to teleoperators as a video feed that will enhance teleoperation. To the best of our knowledge, we are the first to attempt to reduce the impacts of latency through future frame prediction using deep neural networks. [ABSTRACT FROM AUTHOR]

A delicate balance between safety, efficiency, and fluidity needs to be carefully maintained in vehicle following, in strict accordance with real-time control imperatives. Achieving efficient vehicle-following operations under safe driving conditions, through smooth behavioral adjustments, presents a significant challenge for data-driven vehicle-following models. In response to this challenge, we have developed a deep neural network based on gated recurrent unit (GRU) with attention mechanism, named AGRUNet model, for artificial intelligence (AI) control of vehicle following behavior. Through training and testing on diverse datasets, the AGRUNet model not only establishes a nonlinear mapping relationship between the following vehicle's acceleration and the speeds of the leading and following vehicles, their distance, and the control strategy of the leading vehicles but also accurately forecasts the future behaviors of following vehicles in complex vehicle-following scenarios in real-time. This capability enables the following vehicle to optimize its behavior based on the current vehicle-following situation and control requirements, thereby improving safety, efficiency, and smoothness. Rigorous simulations of AGRUNet on the Highway Drone(HighD), Next Generation Simulation(NGSIM), Waymo, and Lyft Level 5(Lyft) datasets demonstrate its superior performance in prediction accuracy and vehicle-following control. Compared to the widely adopted, high-performance Long Short-Term Memory (LSTM) model, AGRUNet achieves prediction accuracy gains of approximately 2%, 7%, 22%, and 3% across these datasets. Extensive testing further indicates that AGRUNet significantly reduces collision rates during sudden emergency braking by the leading vehicle, enhancing safety, and improving the efficiency and smoothness of behavior adjustments, all while ensuring vehicle-following safety. • An AGRUNet model is developed to optimize vehicle-following control under complex scenarios. • The model integrates attention mechanism and GRU to capture crucial features for vehicle following. • Extensive evaluations on public driving datasets demonstrate superior performance. • Improving vehicle-following behavior in terms of safety, efficiency, and smoothness. [ABSTRACT FROM AUTHOR]

Patient-based real-time quality control (PBRTQC) utilizes patient test results to continuously monitor laboratory test quality, addressing issues like discontinuities and matrix effects of traditional internal quality control. However, its clinical performance still requires enhancement. This study combined neural networks (NN) and joint probability analysis (NN-PBRTQC) to improve the clinical performance of PBRTQC. Data were collected from Peking University Shenzhen Hospital and Nanfang Hospital Southern Medical University, which included a series of analytes. A neural network model was trained to predict the test results by integrating patient demographics. Residuals between the expected and actual test results were inputs for statistical process control algorithms to monitor analytical errors. Additionally, an intelligent alarm system using joint probability analysis was developed to reduce the false alarm rate (FAR). The performance of NN-PBRTQC was evaluated using FAR, and the number of patients until error detection was compared to traditional PBRTQC. NN-PBRTQC significantly enhanced the clinical performance of PBRTQC. Under the same desired FAR (DFAR) of 0.1 %, NN-PBRTQC required 64 % fewer samples for error detection than traditional PBRTQC for the analytes, which improved the sensitivity of error detection. NN-PBRTQC provides a novel method for PBRTQC, effectively addressing sample variations and false alarms. It significantly reduces the false alarm rate and the sample size required for error detection, accelerating the implementation of PBRTQC in laboratories. [ABSTRACT FROM AUTHOR]

Rapid changes in the wind field of tropical cyclones can cause excessive loads and threaten the safety of offshore wind turbines. This paper designs a wind turbine yaw optimization strategy based on the deep neural network to reduce the structural loads of wind turbines caused by tropical cyclones. Firstly, the high-precision tropical cyclone data is used to analyze the characteristics of the wind field. Then, a pseudo-Monte Carlo experiment is designed to compensate for the incompleteness of the observed data. Furthermore, a robust nonlinear coupling model between the wind characteristics and the loads of the wind turbine is constructed by a deep neural network, and the optimal yaw angle is searched in real time based on this model. The simulation results show that the proposed deep neural network model based on pseudo-Monte Carlo scenario generation can robustly calculate the structural loads of wind turbines with a calculation error of less than 8 %. After applying this model to the real-time optimization control loop, the corresponding optimized yaw angle can be obtained according to the operating data of the wind turbine under tropical cyclone conditions so that the structural loads of the wind turbine are reduced. Compared with the no-yaw and traditional yaw strategies, the load suppression effect is 1 %–9 % under different working conditions. The proposed data-driven structural load model and load suppression algorithm will effectively improve the operating safety of wind turbines under tropical cyclone conditions. • Insufficient tropical cyclone observation data has been made up; • Deep neural network load models can be calculated quickly and accurately; • Optimizing yaw strategy reduces wind turbine structural loads in tropical cyclones. [ABSTRACT FROM AUTHOR]

To cope with the weak grid stability issue of grid-tied voltage source converters (VSCs), this article proposes an artificial neural network (ANN) based approach for online stabilization control of the grid-tied VSC with the pole-tracking feature. First, an ANN is adopted to establish the mapping between the control parameters and the closed-loop poles of the grid-VSC system, serving as a computationally light model surrogate that is favorable for real-time control applications. Then, an online parameter search algorithm enabling simultaneous tuning of multiple controllers and parameters is developed, by which the system's poles under a new grid condition can be pulled to the reference ones, i.e., achieving the pole-tracking-based stabilization control of this article. Finally, the efficacy of the proposed method along with its stabilization effect is verified by MATLAB simulations and experimental results. [ABSTRACT FROM AUTHOR]

Solving high-dimensional optimal control problems in real-time is an important but challenging problem, with applications to multiagent path planning problems, which have drawn increased attention given the growing popularity of drones in recent years. In this paper, we propose a novel neural network method called SympOCnet that applies the symplectic network to solve high-dimensional optimal control problems with state constraints. We present several numerical results on path planning problems in two- and three-dimensional spaces. Specifically, we demonstrate that our SympOCnet can solve a problem with more than 500 dimensions in 1.5 hours on a single GPU, which shows the effectiveness and efficiency of SympOCnet. The proposed method is scalable and has the potential to solve truly high-dimensional path planning problems in real-time. [ABSTRACT FROM AUTHOR]

We present formulation and open-source tools to achieve in-material model predictive control of sensor/actuator systems using learned forward kinematics and on-device computation. Microcontroller units that compute the prediction and control task while colocated with the sensors and actuators enable in-material untethered behaviors. In this approach, small parameter size neural network models learn forward kinematics offline. Our open-source compiler, nn4mc, generates code to offload these predictions onto MCUs. A Newton-Raphson solver then computes the control input in real time. We first benchmark this nonlinear control approach against a PID controller on a mass-spring-damper simulation. We then study experimental results on two experimental rigs with different sensing, actuation and computational hardware: a tendon-based platform with embedded LightLace sensors and a HASEL-based platform with magnetic sensors. Experimental results indicate effective high-bandwidth tracking of reference paths (≥120 Hz) with a small memory footprint (≤6.4% of flash memory). The measured path following error does not exceed 2mm in the tendon-based platform. The simulated path following error does not exceed 1mm in the HASEL-based platform. The mean power consumption of this approach in an ARM Cortex-M4f device is 45.4 mW. This control approach is also compatible with Tensorflow Lite models and equivalent on-device code. In-material intelligence enables a new class of composites that infuse autonomy into structures and systems with refined artificial proprioception. Graphical Abstract [ABSTRACT FROM AUTHOR]

Portable devices play an essential role where edge computing is necessary and mobility is required (e.g., robots in agriculture within remote-sensing applications). With the increasing applications of deep neural networks (DNNs) and accelerators for edge devices, several methods and applications have been proposed for simultaneous crop and weed detection. Although preliminary studies have investigated the performance of inference time for semantic segmentation of crops and weeds in edge devices, performance degradation has not been evaluated in detail when the required optimization is applied to the model for operation in such edge devices. This paper investigates the relationship between model tuning hyperparameters to improve inference time and its effect on segmentation performance. The study was conducted using semantic segmentation model DeeplabV3 with a MobileNet backbone. Different datasets (Cityscapes, PASCAL and ADE20K) were analyzed for a transfer learning strategy. The results show that, when using a model hyperparameter depth multiplier (DM) of 0.5 and the TensorRT framework, segmentation performance mean intersection over union (mIOU) decreased by 14.7% compared to that of a DM of 1.0 and no TensorRT. However, inference time accelerated dramatically by a factor of 14.8. At an image resolution of 1296 × 966 , segmentation performance of 64% mIOU and inference of 5.9 frames per second (FPS) was achieved in Jetson Nano's device. With an input image resolution of 513 × 513 , and hyperparameters output stride OS = 32 and DM = 0.5, an inference time of 0.04 s was achieved resulting in 25 FPS. The results presented in this paper provide a deeper insight into how the performance of the semantic segmentation model of crops and weeds degrades when optimization is applied to adapt the model to run on edge devices. Lastly, an application is described for the semantic segmentation of weeds embedded in the edge device (Jetson Nano) and integrated with the robotic orchard. The results show good spraying accuracy and feasibility of the method. [ABSTRACT FROM AUTHOR]

A drift-driving maneuver is a control technique used by an expert driver to control a vehicle along a sharply curved path or slippery road. This study develops a nonlinear model predictive control (NMPC) method for the autonomous vehicle to perform a drift maneuver and generate the datasets necessary for training the deep neural network(DNN)-based drift controller. In general, the NMPC method is based on numerical optimization which is difficult to run in real-time. By replacing the previously designed NMPC method with the proposed DNN-based controller, we avoid the need for complex numerical optimization of the vehicle control, thereby reducing the computational load. The performance of the developed data-driven drift controller is verified through realistic simulations that included drift scenarios. Based on the results of the simulations, the DNN-based controller showed similar tracking performance to the original nonlinear model predictive controller; moreover, the DNN-based controller can demonstrate stable computation time, which is very important for the safety critical control objective such as drift maneuver. [ABSTRACT FROM AUTHOR]

Significant research efforts have focused on advancing upper limb prostheses, addressing issues such as dexterity, embodiment, weight, and resistance. These key features require enhancements to mitigate prosthesis abandonment. Sensory feedback is a vital feature for improving the dexterity and embodiment of prostheses for patients. Current techniques utilize electronic signals for feedback to control systems and to partially restore tactile sensations to amputees. However, it has already been demonstrated that, in order to elicit responses from biological neurons signals must be converted into neuromorphic tactile information. Most of those demonstrations relied on algorithmic implementations of mechanoreceptor and neuron models running on classic processor cores. However, integrating real-time encoding and control strategies into such architectures is challenging due to computational complexity and high power usage. A prospective solution is to increasingly incorporate neuromorphic technologies for prosthesis sensing and control. In this work, we demonstrate the feasibility of using spiking neural networks for performing two key functions in controlling upper limb prosthesis: touch and slippage detection. We found that a simple two-layered architecture with a few tens of neurons is sufficient to identify slip and touch with nearly 100% accuracy. The first layer consists of a heterogeneous population of neurons with their parameters adjusted to model slowly and fast-adapting type 1 mechanoreceptors. An output layer classifies touch or slip events. Furthermore, simulations showed that implementing the network in ultra-low-power neuromorphic hardware, such as Loihi, has the potential to reduce power consumption by up to 162 times, with respect to a von Neumann processor. [ABSTRACT FROM AUTHOR]

[Display omitted] • UV/VIS Machine vision system to determine CQAs of fully white tablets. • Rapid evaluation of compression force impact and API content using UV/VIS imaging. • Accurate analysis of critical quality attributes with multivariate data techniques. • Rapid analysis suitable for both off-line and in-line use. The paper provides a demonstration of how UV/VIS imaging can be employed to evaluate the crushing strength, friability, disintegration time and dissolution profile of tablets comprised of solely white components. The samples were produced using different levels of compression force and API content of anhydrous caffeine. Images were acquired from both sides of the samples using UV illumination for the API content prediction, while the other parameters were assessed using VIS illumination. Based on the color histograms of the UV images, API content was predicted with 5.6 % relative error. Textural analysis of the VIS images yielded crushing strength predictions under 10 % relative error. Regarding friability, three groups were established according to the weight loss of the samples. Likewise, the evaluation of disintegration time led to the identification of three groups: <10 s, 11–35 s, and over 36 s. Successful classification of the samples was achieved with machine learning algorithms. Finally, immediate release dissolution profiles were accurately predicted under 5 % of RMSE with an artificial neural network. The 50 ms exposition time during image acquisition and the resulting outcomes underscore the practicality of machine vision for real-time quality control in solid dosage forms, regardless of the color of the API. [ABSTRACT FROM AUTHOR]

Food quality is a fuzzy category, which could be evaluated using fuzzy logic. Our approach to food quality evaluation is based on mapping food quality attributes into a fuzzy domain as a multi-dimensional fuzzy sets. First, the data representing quality attributes are mapped into orthogonal coordinates using PCA to reduce dimensionality. Second, subtractive clustering (SC) is applied to determine a representative number of clusters. Each point in the dataset is associated with each cluster by credibilistic fuzzy C-means clustering (CFCM). After data organized in fuzzy clusters, an artificial neural network (ANN) is trained to associate each point in the dataset with its membership degree in each cluster. Trained ANN serves as a predictive model to convert real-time data stream into the multi-dimensional fuzzy domain. The application of this methodology is illustrated for real-time quality evaluation in shrimp batch drying. In this study 27 quality attributes have been merged into 9 orthonormal vectors, which have been clustered into 10 fuzzy sets. This structuring of the experimental fuzzy domain allowed the development of a multi-dimensional kinetic model, which improved the quality of shrimp drying. The computational time for quality identification in the fuzzy domain is below 1 sec, which is satisfactory for most real-time applications. This data-driven algorithm is completely automated and has unlimited potential for real-time fuzzy control and optimization. Multi-dimensional fuzzy sets are unique identifiers of food quality Extracting principal information in orthonormal coordinates Using the artificial neural network for predicting membership functions A multi-dimensional fuzzy kinetics model was developed Structuring of fuzzy domain decreased computational time for fuzzy control [ABSTRACT FROM AUTHOR]

Sustainable management of water supply systems is a major challenge within the framework of the water-energy nexus. The main strategies to improve the operation of these systems are related to increasing the hydraulic and energy efficiency of pumping systems. In this context, this work presents a new artificial neural network (ANN) controller to improve the operation of water distribution systems (WDSs) that includes in its algorithm the specific energy consumption (SEC) as a decision parameter. Therefore, pressure control at the measuring points is also based on the energy efficiency of the pumps. The technique was applied to control the pressures in an experimental setup that emulates a WDS with two consumption zones with different topographies. For this purpose, the controller acted on a conventional pump, a booster pump and a control valve. To analyze the performance under the controller action, tests were performed emulating water-demand scenarios, introducing perturbations and changing the pressure setpoints. The real-time control performance was proven based on the dynamic performance, steady-state performance and SEC. The experimental results showed that the proposed controller kept the pressures close to the setpoints and provided a reduction in the SEC between 15.1% and 17.8%, compared with the uncontrolled system, and an economy that varied from 2.5% to 8.1% compared with the performance of the ANN based only on pressure control. [ABSTRACT FROM AUTHOR]

It is necessary to grasp the moisture content of wheat in the process of hot air drying in real time to save energy and improve quality. According to the result of the hot air drying experiment, a backpropagation (BP) neural network prediction model with a topology of 2‐4‐1 is established. The mean absolute error (MAE) and degree of fit (R2) of the predicted results are 0.216% and 0.9804. The BP neural network model is optimized with a genetic algorithm (GA‐BP) to improve accuracy. The MAE and R2 are down to 0.069% and 0.9945. The generalization ability of the GA‐BP prediction model is verified, when predicting the samples out of the training set, the MAE is 0.37% and R2 is 0.9993. It shows that the GA‐BP prediction model is of good generalization ability. This study provides a new conception for the real‐time control of moisture content in hot air drying wheat. Novelty impact statement: The prediction model of wheat moisture content in the hot air drying process was established based on BP neural network algorithm and optimized by the genetic algorithm. The wheat moisture content can be accurately predicted at different hot air temperatures and drying times, which provides a new method for the real‐time monitoring of moisture content in the drying process. According to the predicted results, the changes in the moisture content of wheat are grasped in advance, which solve the problem that the moisture content of wheat is difficult to obtain in real time during the hot air drying process and hard to adjust the hot air temperature in time. [ABSTRACT FROM AUTHOR]

The transmission line is an integral part of the electrical power system; however, a fault has a negative impact on the system, like blackout, power loss, financial losses, and socio-economic impact. This fault occurs due to ageing conductors, lightning stroke, switching surge and human interference. We reviewed the protection scheme implemented in the Nigerian transmission network, which has challenges relating to the environment's terrain and a long-distance transmission line of about 20,000 km. The different approach of fault classification, detection and location was analyzed and critically summarized. This review paper proposes a hybrid artificial neural network (ANN) and distance protection scheme that can automatically identify, locate, isolate, predict, correct faults, and real-time monitor and control the entire network. It can also detect the shortest possible trip time of 0.02 s and 0.03 s of line current and fault losses, respectively, during fault to avert damage on the line. However, this method has its challenges, such as the volume of data generated from load flow analysis, training time, and the total distance covered by the network. However, these can be averted by segmenting the entire network for easy evaluation and monitoring to achieve set goals. [ABSTRACT FROM AUTHOR]

Estimation of combustion phasing and power production is essential to ensuring proper combustion and load control. However, archetypal control-oriented physics-based combustion models can become computationally expensive if highly accurate predictive capabilities are achieved. Artificial neural network (ANN) models, on the other hand, may provide superior predictive and computational capabilities. However, using classical ANNs for model-based prediction and control can be challenging, since their heuristic and deterministic black-box nature may make them intractable or create instabilities. In this paper, a hybridized modeling framework that leverages the advantages of both physics-based and stochastic neural network modeling approaches is utilized to capture CA50 (the timing when 50% of the fuel energy has been released) along with indicated mean effective pressure (IMEP). The performance of the hybridized framework is compared to a classical ANN and a physics-based-only framework in a stochastic environment. To ensure high robustness and low computational burden in the hybrid framework, the CA50 input parameters along with IMEP are captured with a Bayesian regularized ANN (BRANN) and then integrated into an overall physics-based 0D Wiebe model. The outputs of the hybridized CA50 and IMEP models are then successively fine-tuned with BRANN transfer learning models (TLMs). The study shows that in the presence of a Gaussian-distributed model uncertainty, the proposed hybridized model framework can achieve an RMSE of 1.3 × 10−5 CAD and 4.37 kPa with a 45.4 and 3.6 s total model runtime for CA50 and IMEP, respectively, for over 200 steady-state engine operating conditions. As such, this model framework may be a useful tool for real-time combustion control where in-cylinder feedback is limited. [ABSTRACT FROM AUTHOR]

Although powder bed fusion joining (TYB) metal additive manufacturing is frequently preferred in the production of complex geometry parts today, real-time monitoring of part manufacturing processes is insufficient. Therefore, the machine control system remains largely open loop. While some metal additive manufacturing machines present the powder bed monitoring with images, it has not been found that they can automatically detect the defects that may occur in the powder bed layer and stimulate the control system. In the study, an exemplary machine learning-based approach is presented for on-site monitoring and defect detection of powder bed images, which can be a component of a real-time control system in any TYB metal additive manufacturing machine. Using the deep learning method, which is one of the subfields of machine learning, a classification was made to detect the defects that may occur in creating a layer of the process. Detection and classification of defects were carried out using the convolutional neural networks model. The data set for training and performance of the model was created with photographs of a three-dimensional sample structure manufactured on the EOS M290 machine. The best performance was obtained in the VGG16 model with 88.3% accuracy by performing transfer learning from VGG-16, Inception V3, and DenseNet pre-learning models. [ABSTRACT FROM AUTHOR]

Rainfall–runoff modelling is at the core of any hydrological forecasting system. The high spatio-temporal variability of precipitation patterns, complexity of the physical processes, and large quantity of parameters required to characterize a watershed make the prediction of runoff rates quite difficult. In this study, a hyper-complex artificial neural network in the form of an octonion-valued neural network (OVNN) is proposed to estimate runoff rates. Evaluation of the proposed model is performed using a rainfall time series from a raingauge near a Canadian watershed. Results of the artificial intelligence-generated runoff rates illustrate its capacity to produce more computationally efficient runoff rates compared to those obtained using a physically based model. In addition, training the data using the proposed OVNN vs. a real-valued neural network shows less space complexity (1*3*1 vs. 8*10*8, respectively) and more accurate results (0.10% vs. 0.95%, respectively), which accounts for the efficiency of the OVNN model for real-time control applications. [ABSTRACT FROM AUTHOR]

Wave energy is a type of abundant and dense renewable energy. Wave force prediction is a critical technology that influences power absorption efficiency in the real-time control of wave energy converter (WEC). Could wave elevation be used to predict wave excitation force directly by training artificial neural network? This method results in rapid and suitable prediction for real-time control. A long short-term memory recurrent neural network (LSTM RNN) algorithm is introduced to identify characteristics of wave excitation forces based on wave elevations. In this method, the wave elevations in front of the structure are measured to obtain sufficient time to actuate the control manipulation. A total of 180 regular wave and 12 irregular wave tests are conducted, and the LSTM RNN model is trained based on the experimental results. The performance of the LSTM algorithm is verified. According to the regular cases in the study, the LSTM prediction can identify high-order harmonic loads, and the anti-noise capability of the LSTM algorithm can filter random noises from the measure signals. In the irregular cases, the LSTM RNN algorithm performs effectively to predict the wave force excited on the structure using wave elevations measured by wave probes. The best combinations of the test setting parameters are determined to guide experimental tests and WEC prototypes. • Introducing LSTM algorithm to predict wave excitation force based on experimental tests. • Identifying the high-order harmonic loads and decreasing the influence of noises. • Determining the optimal number and spacing distance of input wave probes for experimental test of wave load measurement. [ABSTRACT FROM AUTHOR]

In order to adequately leverage the advantages of coaxial compound helicopters (CCH) to enhance overall performance and efficiency through variable rotor speed, and to improve the fuel economy of the composite propulsion system, a real-time optimization control method of variable rotor speed based on helicopter/turboshaft engine on-board composite system is proposed. Firstly, a steady-state model of the CCH required power and an on-board inverse model of the turboshaft engine based on an improved deep neural network propulsion system model (AdamDNN-PSM) are established. On this basis, an on-board adaptive model based on adaptive square root Kalman filter (ASRUKF) estimator is employed to correct the precision of the on-board inverse model and to track the real engine state. Finally, by adopting the feasible sequence quadratic programming (FSQP) method, an optimal speed rotor (OSR) is obtained through real-time optimization control, which reduces fuel consumption during cruising while considering both the geometric and aerodynamic constraints of the rotor. The simulation results demonstrate that the accuracy and real-time performance of the on-board inverse model have been significantly improved. The real-time optimization control method for OSR effectively reduces fuel consumption during cruising and enhances the CCH performance of the entire flight envelope. Please find our suggested Highlights listed below. • A real-time optimization control method of variable rotor speed based on helicopter/turboshaft engine on-board composite system of CCH is proposed • A high-precision and high real-time airborne comprehensive model is conducted. • The precision of the on-board composite system is corrected by an on-board adaptive model based on ASRUKF. • Consider both the geometric and aerodynamic constraints of the CCH in real-time optimization control. [ABSTRACT FROM AUTHOR]

• Neural network model predictive controller proposed for fluidized bed gasification (FBG) process. • Effective spatiotemporal temperature control achieved using the proposed controller. • Real-time computing capabilities of the controller enable efficient control actions. • Controller successfully tested on a physics-based model of the FBG process. In this study, an advanced controller for the fluidized bed biomass gasification (FBG) process is proposed. The controller is based on the model predictive control method, utilizing a long short-term memory neural network to accurately predict the controlled variable in the FBG process. Then, employing a gradient-based optimization algorithm, it optimizes the inputs to achieve the desired control objective. The controller specifically aims for temperature regulation in three spatial regions (fluidized bed, freeboard and outlet), using primary air, secondary air, and biomass flow rates as input variables. Open-loop simulations are used to fine-tune the controller's parameters (prediction and control horizon). For unbiased controller testing, a computational fluid dynamics (CFD) model of the FBG process is also developed, which is integrated with the controller in a closed-loop system, simulating real-process feedback for performance evaluation. The proposed controller effectively maintains temperatures (steady state error < 1.5 %) at desired set points (800–900 °C) with a short prediction horizon of 24 time steps, reducing response time to under 5 s, making it suitable for real-time FBG process control. [ABSTRACT FROM AUTHOR]

• A two-layer control method is proposed to deal with calculation complexity. • The clustered temperature transfer processes of bitumen tanks are analyzed. • A data-driven method is introduced to enhance computational efficiency. • A real-time predictive control method is developed to deal with uncertainties. Uncertainties and computational complexity are two growing challenges in scheduling industrial heating loads. In this paper, a data-driven real-time predictive control approach is proposed to deal with these challenges in the industrial scheduling of bitumen tanks. Specifically, predictive control technology is utilized to leverage the updated information to mitigate the negative impact of past uncertainties in equipment parameters and external environmental factors, which may lead to temperature constraint violations in the bitumen tank operation processes. Meanwhile, a data-driven method using artificial neural networks (ANN) is developed to ensure efficient computation for real-time predictive control. Moreover, a two-layer control method is devised to reduce the calculation time for day-ahead optimal scheduling of a large scale of bitumen tanks, aiming to generate sufficient high-quality data for training ANN. In the two-layer control method, the clustered temperature transfer processes of bitumen tanks are analyzed and modeled for the first time. Simulation results indicate that the two-layer control method can significantly reduce the computational time required for the day-ahead optimal scheduling of bitumen tanks, facilitating the generation of a large amount of high-quality data for training ANN. Subsequently, the application of ANN enables real-time predictive control, helping to eliminate the negative impact of uncertainties. [ABSTRACT FROM AUTHOR]

For optimizing and scheduling multi-energy system's operation in real applications, forecasts are needed. Therefore, forecasts impact the scheduling results and the degree to with an optimized schedule can be implemented. In the present investigation, real data from a German hospital with a combined heat and power unit, gas boiler and thermal energy storage (TES) is used as case study. For this case study, different heat load forecasts are generated by artificial neural networks and a persistence forecast, optimal operation is scheduled via MILP based on these forecasts, and the determined schedules are implemented stepwise in a simulative environment. Forecasts are evaluated by mean absolute percentage error (between 6 and 13%) and normalized mean error (between −2 and 12%). Two performance indicators are selected for evaluation, comparing the planned, realized, and idealized operation (the latter based on perfect foresight) with each other: balancing energy and degree of self-generation. Using the best-performing forecasts, self-generation is half a percentage point lower than in the idealized case and 2% negative balancing energy are needed. In general, underestimating heat load is beneficial to avoid balancing energy (high reliability), whereas overestimating is beneficial to achieve a high degree of self-generation (low operating costs). The gained experience helps to implement optimal scheduling and control effectively. • Multi-model FFNNs provide accurate hospital's heat load forecasts. • To evaluate forecast accuracy both MAPE and nME are important. • Comparison of ideal, planned, and realized operation. • Self-generation introduced as meaningful performance indicator. • Forecast deviations impact the need for balancing energy. [ABSTRACT FROM AUTHOR]

This study proposes a two-step method for generating switching angles in renewable energy systems that use multi-level inverters (MLIs) to reduce low-order harmonics. The Selective Harmonic Elimination Pulse Width Modulation (SHE-PWM) technique is used for MLI control, but it can be computationally intensive for real-time applications. To address this challenge, the proposed approach consists of a two-stage process. In the first stage, the SHE equations are solved offline using artificial ecosystem-based optimization (AEO). The obtained switching angles are then used to train an artificial neural network (ANN) prediction model in the second stage. The AEO-ANN-based SHE-PWM technique is applied to a reduced-switching, 3-phase, 7-level MLI. Simulations in MATLAB/SIMULINK show that the proposed method achieves accurate voltage control with less than 0.2% error, even for changing voltages, and reduces selected harmonics to less than 0.05%. The desired output voltage exhibits minimal total harmonic distortion (THD). This method offers a promising way to generate switching angles in renewable energy systems that use MLIs, improving power quality and reducing harmonic distortion. [ABSTRACT FROM AUTHOR]

Modular multilevel converter (MMC) has attracted much attention for years due to its good performance in harmonics reduction and efficiency improvement. Model predictive control (MPC) based controllers are widely adopted for MMC because the control design is straightforward and different control objectives can be simply implemented in a cost function. However, the computational burden of MPC imposes limitations in the control implementation of MMC because of many possible switching states. To solve this, we design machine learning (ML) based controllers for MMC based on the data collection from the MPC algorithm. The ML models are trained to emulate the MPC controllers which can effectively reduce the computation burden of real-time control since the trained models are built with simple math functions that are not correlated with the complexity of the MPC algorithm. The ML method applied in this study is a neural network (NN) and there are two types of establishing ML controllers: NN regression and NN pattern recognition. Both are trained using the sampled data and tested in a real-time MMC system. A comparison of experimental results shows that NN regression has a much better control performance and lower computation burden than the NN pattern recognition. [ABSTRACT FROM AUTHOR]

Robotic arms are widely used in many industrial applications at present. The control of robotic arms involves position coordination Cartesian space by a forward/inverse kinematics solution method. The inverse kinematics is difficult for real-time control applications, computational requirements are intensive and the run-time is high. The traditional solution methods used geometric, algebraic, and numerical iterative techniques are inadequate and slow in the inverse kinematics solution. Recently, alternative solution methods based on artificial intelligence techniques have been developed to solve the inverse kinematics problem. In this study, a multilayered feed-forward Artificial Neural Network model was developed to solve the inverse kinematics of the 5 degrees of freedom robotic arm. Using the Proportional-Integral control algorithm combined with this Artificial Neural Network model, the real-time position control of the robotic arm was accomplished. The obtained results were compared with the PI control supported by an analytical inverse kinematics solution in real-time. The results showed that the PI control combined with Artificial Neural Network has superior tracking ability, smaller control error, and better absolute fit to the reference. [ABSTRACT FROM AUTHOR]

The main principles of neuro-evolutionary technology for hierarchical systems control optimization under conflict and uncertainty are considered. The problem statement is formulated as a hierarchical game under uncertainty with the right first move, which uses the principle of a coordinated stable-effective compromises generalizing the Stackelberg hierarchical equilibrium principle. To construct the optimal control law, the evolutionary synthesis of coordinated stable-effective compromises method is applied. An artificial neural network based on radial-basis functions multi-criteria synthesis method that implements the optimal control law in real time is developed. The problem of a hierarchical FANET-based (Flying Ad Hoc Network) remote monitoring system control optimizing in real time is solved. [ABSTRACT FROM AUTHOR]

This paper presents the results of nonlinear model identification of an unmanned vehicle in the Gazebo simulator based on a neural network autoregressive model. The dynamic characteristics of the control object can vary significantly that complicates the problem. The training sample of movement along the spatial path was obtained at the Robotics Center of the FRC CSC RAS. The model parameters were found by the particle swarm optimization method. Using the identified model, real-time control methods were experimentally compared. For a more accurate assessment of the methods, the model was subjected to random disturbances, and the tracking path of the control object was significantly complicated. [ABSTRACT FROM AUTHOR]

Today, making quality control systems with reliable accuracy is very important in producing industrial products with zero defects. In this respect, it is an essential issue that camera control systems work with reliable control algorithms. In this study, a real-time control algorithm using a pattern matching algorithm has been developed to optimize the minimum contrast parameter with an Artificial Neural Network (ANN). In the study, the comparison of three algorithms included in pattern matching in terms of time was made using LabVIEW image control tools. Besides, one of the most critical parameters in the low-discrepancy sampling algorithm, which gives good results in time, minimum contrast parameter is discussed. The optimization of this parameter is done by using the Levenberg-Marquardt training algorithm in ANN. The obtained results show that the proposed pattern matching algorithm using ANN for optimizing the minimum contrast parameter is fast and effective for quality control applications. [ABSTRACT FROM AUTHOR]

Animals exhibit remarkable feats of behavioral flexibility and multifunctional control that remain challenging for robotic systems. The neural and morphological basis of multifunctionality in animals can provide a source of bioinspiration for robotic controllers. However, many existing approaches to modeling biological neural networks rely on computationally expensive models and tend to focus solely on the nervous system, often neglecting the biomechanics of the periphery. As a consequence, while these models are excellent tools for neuroscience, they fail to predict functional behavior in real time, which is a critical capability for robotic control. To meet the need for real-time multifunctional control, we have developed a hybrid Boolean model framework capable of modeling neural bursting activity and simple biomechanics at speeds faster than real time. Using this approach, we present a multifunctional model of Aplysia californica feeding that qualitatively reproduces three key feeding behaviors (biting, swallowing, and rejection), demonstrates behavioral switching in response to external sensory cues, and incorporates both known neural connectivity and a simple bioinspired mechanical model of the feeding apparatus. We demonstrate that the model can be used for formulating testable hypotheses and discuss the implications of this approach for robotic control and neuroscience. [ABSTRACT FROM AUTHOR]

For predicting the conversion rate of vinyl chloride monomer (VCM) in the polyvinyl chloride (PVC)production process, a neural network soft-sensor model based on data dimensionality reduction strategies was proposed. In order to solve the problem of complex neural network topology and long training time caused by the excessive input vector dimension, seven kinds of data dimensionality reduction methods, such as principal component analysis (PCA), locality preserving projection (LPP), kernel principal component analysis (KPCA), expectation max principal component analysis (EMPCA), local tangent space alignment (LTSA), T-distributed stochastic neighbor embedding (TSNE) and neighboring preserving embedding (NPE), are used to reduce the dimension of the high-dimensional input data used in the neural network soft-sensor model. Then the radial basis function (RBF) neural network based on the gradient learning, orthogonal least squares and clustering learning methods and the dynamic fuzzy neural network (D-FNN) were utilized to realize the prediction on the VCM conversion rate. Simulation results show that the proposed neural network soft-sensor models based on seven data dimensionality reduction strategies can effectively predict the key economic and technical indicators of PVC polymerization process and meet the real-time control requirements of PVC production process. [ABSTRACT FROM AUTHOR]

The aim of this work is to deploy an advanced Nonlinear Model Predictive Control (NMPC) approach for reducing the tower fatigue of a wind turbine (WT) tower while guaranteeing efficient energy extraction from the wind. To achieve this, different Artificial Neural Network (ANN) architectures are trained and tested in order to estimate the tower fatigue as a surrogate of the traditional Rainflow Counting (RFC) method. The ANNs receive data stemming from the tower top oscillation velocity and the previous fatigue state to directly estimate the fatigue progression. The results are compared to select the most convenient architecture for control implementation. Once an ANN is selected, an economic-tracking NMPC (etNMPC) solution to reduce the fatigue of the WT tower is deployed in real-time. The closed-loop results are then compared to a baseline controller from a renowned WT simulation tool and a classic etNMPC implementation with indirect fatigue minimisation to demonstrate the improvement achieved with the proposed strategy. Finally, conclusions regarding computational cost and real-time deployment capabilities are discussed, as well as future lines of research. • We study the effects of fatigue on a multi-megawatt wind turbine tower. • We use machine-learning to estimate the evolution of fatigue on the wind turbine tower. • A real-time economic-tracking NMPC to improve efficiency while reducing tower fatigue is proposed. • The designed controller is tested with data stemming from different wind speeds. • Comparison between the proposed strategy and the FAST controller. [ABSTRACT FROM AUTHOR]

The application of augmented reality (AR) technology for assembly guidance is a novel approach in the traditional manufacturing domain. In this paper, we propose an AR approach for assembly guidance using a virtual interactive tool that is intuitive and easy to use. The virtual interactive tool, termed the virtual interaction panel (VirIP), is an easy-to-use tool that can be used to interactively control AR systems. The VirIP is composed of virtual buttons, which have meaningful assembly information that can be activated by an interaction pen during the assembly process. The interaction pen can be any general pen-like object with a certain colour distribution. It is tracked using a restricted coulomb energy (RCE) network in real-time and used to trigger the relevant buttons in the VirIPs for assembly guidance. Meanwhile, a visual assembly tree structure (VATS) is used for information management and assembly instructions retrieval in this AR environment. VATS is a hierarchical tree structure that can be easily maintained via a visual interface. It can be directly integrated into the AR system or it can alternatively act as an independent central control station on a remote computer to control the data flow of the assembly information. This paper describes a typical scenario for assembly guidance using VirIP and VATS. The main characteristic of the proposed AR system is the intuitive way in which an assembly operator can easily step through a pre-defined assembly plan/sequence without the need of any sensor schemes or markers attached on the assembly components. Several experiments were conducted to validate the performance of the proposed AR-based method using a monitor and a head-mounted display. The results show that the AR-based method can provide an efficient way for assembly guidance. [ABSTRACT FROM AUTHOR]

In this paper, we present the use of deep neural networks to estimate physical parameters from complex optical emission spectra of the D β /H β transition. Specifically, we focus on estimating the radio frequency electric field vector of the lower hybrid wave and isotope ratio within the scrape-off-layer plasma of the WEST tokamak. Fitting the spectral data using a traditional non-linear least squares analysis requires many free parameters and is computationally expensive, rendering the data unusable for real-time control. By implementing relatively small neural networks, the physical parameters can be directly extracted from the spectral data with reasonable accuracy in a few milliseconds. The deep neural network prediction can serve as input for a reduced model using least-squares fitting or for real-time control. We show that deep neural networks can be an effective tool for analyzing complex multicomponent spectra, providing a speedup of more than 1 0 5 times compared to least residual analysis, with an accuracy of 0.5% for the isotope ratio, and 0.09 kV/cm and 0.38 kV/cm for the RF radial and poloidal electric field respectively. • Deep Neural Network as surrogate model for multicomponent complex spectra analysis. • Principal component analysis applied to neural network dimensionality reduction on experimental spectra. • Estimation of the spectra parameters with high throughput. [ABSTRACT FROM AUTHOR]

We propose a new method for the analysis and design of a robotic system that minimizes the energy consumption of a six-axis robot arm by controlling the velocity and acceleration of each arm of the robot to achieve the specified trajectory of the robot determined from a lean manufacturing method. A dynamic model of the PUMA 560 robot has been simulated on MATLAB, while the Robotics Toolbox and particle swarm optimization (PSO) are utilized to search for optimal paths and the optimal velocity and acceleration of the robot arms. The optimal velocity and acceleration are described as those giving minimum overall energy consumption constrained by a specified cycle time of the entire robotic system. Typically, the picking and placing of materials are carried out by humans, causing a variation in production rate, whereas our system using a robot arm ensures a stable production rate. Moreover, the optimal results obtained from PSO are adopted to train an artificial neural network (ANN) to extend the design system from discrete optimal values to a continuous and near-optimal value. In other words, the ANN is used to obtain an approximate optimal value between those obtained from PSO to make the system applicable to a real-world system. As shown by the simulation results, this method reduces the energy consumption of 12.3% from the initial energy and reduces the time for optimization by 99.8% compared with that for the PSO technique. [ABSTRACT FROM AUTHOR]