Exploration of Advanced Applications of Triboelectric Nanogenerator-Based Self-Powered Sensors in the Era of Artificial Intelligence

The integration of Deep Learning with sensor technologies has significantly advanced the field of intelligent sensing and decision making by enhancing perceptual capabilities and delivering sophisticated data analysis and processing functionalities. This review provides a comprehensive overview of the synergy between Deep Learning and sensors, with a particular focus on the applications of triboelectric nanogenerator (TENG)-based self-powered sensors combined with artificial intelligence (AI) algorithms. First, the evolution of Deep Learning is reviewed, highlighting the advantages, limitations, and application domains of several classical models. Next, the innovative applications of intelligent sensors in autonomous driving, wearable devices, and the Industrial Internet of Things (IIoT) are discussed, emphasizing the critical role of neural networks in enhancing sensor precision and intelligent processing capabilities. The review then delves into TENG-based self-powered sensors, introducing their self-powered mechanisms based on contact electrification and electrostatic induction, material selection strategies, novel structural designs, and efficient energy conversion methods. The integration of TENG-based self-powered sensors with Deep Learning algorithms is showcased through their groundbreaking applications in motion recognition, smart healthcare, smart homes, and human-machine interaction. Finally, future research directions are outlined, including multimodal data fusion, edge computing integration, and brain-inspired neuromorphic computing, to expand the application of self-powered sensors in robotics, space exploration, and other high-tech fields. This review offers theoretical and technical insights into the collaborative innovation of Deep Learning and self-powered sensor technologies, paving the way for the development of next-generation intelligent systems.

Selective Memory Recursive Least Squares: Recast Forgetting Into Memory in RBF Neural Network-Based Real-Time Learning

In radial basis function neural network (RBFNN)-based real-time learning tasks, forgetting mechanisms are widely used such that the neural network can keep its sensitivity to new data. However, with forgetting mechanisms, some useful knowledge will get lost simply because they are learned a long time ago, which we refer to as the passive knowledge forgetting phenomenon. To address this problem, this article proposes a real-time training method named selective memory recursive least squares (SMRLS) in which the classical forgetting mechanisms are recast into a memory mechanism. Different from the forgetting mechanism, which mainly evaluates the importance of samples according to the time when samples are collected, the memory mechanism evaluates the importance of samples through both temporal and spatial distribution of samples. With SMRLS, the input space of the RBFNN is evenly divided into a finite number of partitions, and a synthesized objective function is developed using synthesized samples from each partition. In addition to the current approximation error, the neural network also updates its weights according to the recorded data from the partition being visited. Compared with classical training methods including the forgetting factor recursive least squares (FFRLS) and stochastic gradient descent (SGD) methods, SMRLS achieves improved learning speed and generalization capability, which are demonstrated by corresponding simulation results.

A Dual Filter Based on Radial Basis Function Neural Networks and Kalman Filters with Application to Numerical Wave Prediction Models

The aim of this study is to introduce and evaluate a dual filter that combines Radial Basis Function neural networks and Kalman filters to enhance the accuracy of numerical wave prediction models. Unlike the existing methods, which focus solely on systematic errors, the proposed framework concurrently targets both systematic and non-systematic parts of forecast errors, significantly reducing the bias and variability in significant wave height predictions. The produced filter is self-adaptive, identifying optimal Radial Basis Function network configurations through an automated process involving various network parameters tuning. The produced computational system is assessed using a time-window procedure applied across divergent time periods and regions in the Aegean Sea and the Pacific Ocean. The results reveal a consistent performance, outperforming classic Kalman filters with an average reduction of 53% in bias and 28% in RMSE, underlining the dual filter's potential as a robust post-processing tool for environmental simulations.

Real-time haptic characterisation of Hunt-Crossley model based on radial basis function neural network for contact environment

Dynamic soft tissue characterisation is an important element in robotic minimally invasive surgery. This paper presents a novel method by combining neural network with recursive least square (RLS) estimation for dynamic soft tissue characterisation based on the nonlinear Hunt-Crossley (HC) model. It develops a radial basis function neural network (RBFNN) to compensate for the error caused by natural logarithmic factorisation (NLF) of the HC model for dynamic RLS estimation of soft tissue properties. The RBFNN weights are estimated according to the maximum likelihood principle to evaluate the probability distribution of the neural network modelling residual. Further, by using the linearisation error modelled by RBFNN to compensate for the linearised HC model, an RBFNN-based RLS algorithm is developed for dynamic soft tissue characterisation. Simulation and experimental results demonstrate that the proposed method can effectively model the natural logarithmic linearisation error, leading to improved accuracy for RLS estimation of the HC model parameters.

Verification study on how macrofungal fruitbody formation can be predicted by artificial neural network

The occurrence and regularity of macrofungal fruitbody formation are influenced by meteorological conditions; however, there is a scarcity of data about the use of machine-learning techniques to estimate their occurrence based on meteorological indicators. Therefore, we employed an artificial neural network (ANN) to forecast fruitbody occurrence in mycorrhizal species of Russula and Amanita, utilizing meteorological factors and validating the accuracy of the forecast of fruitbody formation. Fungal data were collected from two locations in Western Hungary between 2015 and 2020. The ANN was the commonly used algorithm for classification problems: feed-forward multilayer perceptrons with a backpropagation algorithm to estimate the binary (Yes/No) classification of fruitbody appearance in natural and undisturbed forests. The verification indices resulted in two outcomes: however, development is most often studied by genus level, we established a more successful, new model per species. Furthermore, the algorithm is able to successfully estimate fruitbody formations with medium to high accuracy (60-80%). Therefore, this work was the first to reliably utilise the ANN approach of estimating fruitbody occurrence based on meteorological parameters of mycorrhizal specified with an extended vegetation period. These findings can assist in field mycological investigations that utilize sporocarp occurrences to ascertain species abundance.

Three-Dimensional Scanning Applied for Flexible and In Situ Calibration of Galvanometric Scanner Systems

Galvanometric laser scanner (GLS) systems are widely used for materials processing due to their high precision, processing velocity, and repeatability. However, GLS systems generally suffer from scan field distortions due to joint and task space relationship errors. The problem is further pronounced in robotic applications, where the GLS systems are manipulated in space, as unknown errors in the relative pose of the GLS can be introduced. This paper presents an in situ, data-driven methodology for calibrating GLS systems using 3D scanning, emphasising the flexibility, generalisation, and automated industrial integration. Three-dimensional scanning serves two primary purposes: (1) determining the relative pose between the GLS system and the calibration plate to minimise calibration errors and (2) supplying an image processing algorithm with dense and accurate data to measure the scan field distortion based on the positional deviations of marked fiducials. The measured deviations are used to train a low-complexity Radial Basis Function (RBF) network to predict and correct the distorted scan field. The proposed method shows promising results and significantly reduces the scan field distortion without the use of specialised calibration tools and with limited knowledge of the optical design of the GLS system.

Attitude tracking control design of fixed-wing UAVs having uncertain dynamics and corrupted gyro sensor outputs

Flight performance of unmanned aerial vehicles (UAVs) strongly depends on implemented attitude tracking control. For designing better controllers, nonlinear control design techniques are often opted instead of control design based on linearized models. Uncertainty in nonlinear dynamics estimation may arise due to inaccuracies in aerodynamic derivatives and simplifications/assumptions made during the derivation of nonlinear models. This paper considers attitude tracking control of fixed-wing UAVs having uncertain dynamics and corrupted gyro sensor outputs. An integral chain differentiator (ICD) is used to provide the analytical redundancy to the gyros used to measure the angular rates. Two control design schemes are proposed, a neuro-fuzzy adaptive sliding mode control (NFASMC) and an ICD approximation-based fuzzy adaptive sliding mode control (ICD-FASMC). In NFASMC, the uncertain part of the dynamics is estimated using an adaptive radial basis function neural network. Gyro sensor output errors are estimated in real-time, using ICD based error estimation scheme and used in the control law along with the sensor’s corrupted outputs. In ICD-FASMC, the uncertain dynamics and angular rates of UAV are estimated using the ICD such that the requirement of the gyro sensor outputs for control design is bypassed. The switching gain of the designed controllers is made adaptive using fuzzy logic to mitigate the chattering effect. The stability of the proposed controllers is proved using the Lyapunov approach. The proposed schemes are implemented using a nonlinear simulation of a fixed-wing UAV. Simulation results are presented to show the effectiveness of the proposed techniques.

Rapid and Non-Destructive Analysis of Corky Off-Flavors in Natural Cork Stoppers by a Wireless and Portable Electronic Nose

This article discusses the use of a handheld electronic nose to obtain information on the presence of some aromatic defects in natural cork stoppers, such as haloanisoles, alkylmethoxypyrazines, and ketones. Typical concentrations of these compounds (from 5 to 120 ng in the cork samples) have been measured. Two electronic nose prototypes have been developed as an instrumentation system comprise of eight commercial gas sensors to perform two sets of experiments. In the first experiment, a quantitative approach was used whist in the second experiment a qualitative one was used. Machine learning algorithms such as k-nearest neighbors and artificial neural networks have been used in order to test the performance of the system to detect cork defects. The use of this system tries to improve the current aromatic defect detection process in the cork stopper industry, which is done by gas chromatography or human test panels. We found this electronic nose to have near 100 % accuracy in the detection of these defects.

Reduced-Dimension, Biophysical Neuron Models Constructed From Observed Data

Using methods from nonlinear dynamics and interpolation techniques from applied mathematics, we show how to use data alone to construct discrete time dynamical rules that forecast observed neuron properties. These data may come from simulations of a Hodgkin-Huxley (HH) neuron model or from laboratory current clamp experiments. In each case, the reduced-dimension, data-driven forecasting (DDF) models are shown to predict accurately for times after the training period. When the available observations for neuron preparations are, for example, membrane voltage V(t) only, we use the technique of time delay embedding from nonlinear dynamics to generate an appropriate space in which the full dynamics can be realized. The DDF constructions are reduced-dimension models relative to HH models as they are built on and forecast only observables such as V(t). They do not require detailed specification of ion channels, their gating variables, and the many parameters that accompany an HH model for laboratory measurements, yet all of this important information is encoded in the DDF model. As the DDF models use and forecast only voltage data, they can be used in building networks with biophysical connections. Both gap junction connections and ligand gated synaptic connections among neurons involve presynaptic voltages and induce postsynaptic voltage response. Biophysically based DDF neuron models can replace other reduced-dimension neuron models, say, of the integrate-and-fire type, in developing and analyzing large networks of neurons. When one does have detailed HH model neurons for network components, a reduced-dimension DDF realization of the HH voltage dynamics may be used in network computations to achieve computational efficiency and the exploration of larger biological networks.

Incorporating Drone and AI to Empower Smart Journalism via Optimizing a Propagation Model

In the recent digital age, information and communication technologies are rapidly contributing to remodel the media and journalism. Numerous technologies can be utilized by the media industry to capture news or events, taking footage and pictures of a breaking news. Technology and the media are interwoven, and neither can be detached from contemporary society in most nations. Unsurprisingly, technology has affected how and where information is shared. Nowadays, it is impractical to discuss media and the methods in which societies communicate without addressing the rapidity of technology change. Thus, the aerial journalism term has emerged, which refers to the ability of creating and conveying media content in a timely and efficient fashion. This work aims to integrate a drone with AI to empower aerial journalism via training a neural network to obtain an accurate channel using the NN-RBFN approach. The proposed work can enhance aerial media missions including investigative reporting (e.g., humanitarian crises), footage of news events (e.g., man-made and/or natural disasters), and livestreams for short-term, large-scale events (e.g., Olympic Games). In our digital media era, such a smart journalism approach would help to become far more sustainable and an eco-efficient process. Both MATLAB and 3D Remcom Wireless Insite tools have been used to carry out the simulation work. Simulated results indicate that the proposed NN-RBFN managed to obtain an accurate channel propagation model in a 3D scenario with a high accuracy rate reaching 99%. The proposed framework also could offer various media and journalism services (e.g., high data rate, wider coverage footprint) in timely and cost-effective manners in both normal scenarios or even in hard-to-reach zones and/or short-term, large-scale events.

Global fast terminal sliding mode based radial basis function neural network for accurate fault estimation in nonlinear systems

The problem of quick and accurate fault estimation in nonlinear systems is addressed in this article by combining the technique of radial basis function neural network (RBFNN) and global fast terminal sliding mode control (GFTSMC) concept. A new strategy to update the neural network weights, by using the global fast terminal sliding surface instead of conventional error back propagation method, is introduced to achieve real time, quick and accurate fault estimation which is critical for fault tolerant control system design. The combination of online learning ability of RBFNN, to approximate any nonlinear function, and finite time convergence property of GFTSMC ensures quick detection and accurate estimation of faults in real time. The effectiveness of the proposed strategy is demonstrated through simulations using a nonlinear model of a commercial aircraft and considering a wide range of sensors and actuators faults. The simulation results show that the proposed method is capable of quick and accurate online fault estimation in nonlinear systems and shows improved performance as compared to conventional RBFNN and other techniques existing in literature.

Applying Artificial Intelligence to Improve On-Site Non-Destructive Concrete Compressive Strength Tests

In the construction industry, non–destructive testing (NDT) methods are often used in the field to inspect the compressive strength of concrete. NDT methods do not cause damage to the existing structure and are relatively economical. Two popular NDT methods are the rebound hammer (RH) test and the ultrasonic pulse velocity (UPV) test. One major drawback of the RH test and UPV test is that the concrete compressive strength estimations are not very accurate when comparing them to the results obtained from the destructive tests. To improve concrete strength estimation, the researchers applied artificial intelligence prediction models to explore the relationships between the input values (results from the two NDT tests) and the output values (concrete strength). In-situ NDT data from a total of 98 samples were collected in collaboration with a material testing laboratory and the Professional Civil Engineer Association. In-situ NDT data were used to develop and validate the prediction models (both traditional statistical models and AI models). The analysis results showed that AI prediction models provide more accurate estimations when compared to statistical regression models. The research results show significant improvement when AI techniques (ANNs, SVM and ANFIS) are applied to estimate concrete compressive strength in RH and UPV tests.

Design and Optimal Control of a Multistable, Cooperative Microactuator

In order to satisfy the demand for the high functionality of future microdevices, research on new concepts for multistable microactuators with enlarged working ranges becomes increasingly important. A challenge for the design of such actuators lies in overcoming the mechanical connections of the moved object, which limit its deflection angle or traveling distance. Although numerous approaches have already been proposed to solve this issue, only a few have considered multiple asymptotically stable resting positions. In order to fill this gap, we present a microactuator that allows large vertical displacements of a freely moving permanent magnet on a millimeter-scale. Multiple stable equilibria are generated at predefined positions by superimposing permanent magnetic fields, thus removing the need for constant energy input. In order to achieve fast object movements with low solenoid currents, we apply a combination of piezoelectric and electromagnetic actuation, which work as cooperative manipulators. Optimal trajectory planning and flatness-based control ensure time- and energy-efficient motion while being able to compensate for disturbances. We demonstrate the advantage of the proposed actuator in terms of its expandability and show the effectiveness of the controller with regard to the initial state uncertainty.

Exploration of a Framework for the Identification of Chronic Kidney Disease Based on 2D Ultrasound Images: A Survey

BACKGROUND

Chronic kidney disease (CKD) is a fatal disease that ultimately results in kidney failure. The primary threat is the aetiology of CKD. Over the years, researchers have proposed various techniques and methods to detect and diagnose the disease. The conventional method of detecting CKD is the determination of the estimated glomerular filtration rate by measuring creatinine levels in blood or urine. Conventional methods for the detection and classification of CKD are tedious; therefore, several researchers have suggested various alternative methods. Recently, the research community has shown keen interest in developing methods for the early detection of this disease using imaging modalities such as ultrasound, magnetic resonance imaging, and computed tomography.

DISCUSSION

The study aimed to conduct a systematic review of various existing techniques for the detection and classification of different stages of CKD using 2D ultrasound imaging of the kidney. The review was confined to 2D ultrasound images alone, considering the feasibility of implementation even in underdeveloped countries because 2D ultrasound scans are more cost effective than other modalities. The techniques and experimentation in each work were thoroughly studied and discussed in this review.

CONCLUSION

This review displayed the cutting-age research, challenges, and possibilities of further research and development in the detection and classification of CKD.

Ethnic comparison in takotsubo syndrome: novel insights from the International Takotsubo Registry

Background

Ethnic disparities have been reported in cardiovascular disease. However, ethnic disparities in takotsubo syndrome (TTS) remain elusive. This study assessed differences in clinical characteristics between Japanese and European TTS patients and determined the impact of ethnicity on in-hospital outcomes.

Methods

TTS patients in Japan were enrolled from 10 hospitals and TTS patients in Europe were enrolled from 32 hospitals participating in the International Takotsubo Registry. Clinical characteristics and in-hospital outcomes were compared between Japanese and European patients.

Results

A total of 503 Japanese and 1670 European patients were included. Japanese patients were older (72.6 ± 11.4 years vs. 68.0 ± 12.0 years; p < 0.001) and more likely to be male (18.5 vs. 8.4%; p < 0.001) than European TTS patients. Physical triggering factors were more common (45.5 vs. 32.0%; p < 0.001), and emotional triggers less common (17.5 vs. 31.5%; p < 0.001), in Japanese patients than in European patients. Japanese patients were more likely to experience cardiogenic shock during the acute phase (15.5 vs. 9.0%; p < 0.001) and had a higher in-hospital mortality (8.2 vs. 3.2%; p < 0.001). However, ethnicity itself did not appear to have an impact on in-hospital mortality. Machine learning approach revealed that the presence of physical stressors was the most important prognostic factor in both Japanese and European TTS patients.

Conclusion

Differences in clinical characteristics and in-hospital outcomes between Japanese and European TTS patients exist. Ethnicity does not impact the outcome in TTS patients. The worse in-hospital outcome in Japanese patients, is mainly driven by the higher prevalence of physical triggers.

Trial Registration

URL: https://www.clinicaltrials.gov; Unique Identifier: NCT01947621.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00392-021-01857-4.

Multi-Sensor Fusion for Underwater Vehicle Localization by Augmentation of RBF Neural Network and Error-State Kalman Filter

The Kalman filter variants extended Kalman filter (EKF) and error-state Kalman filter (ESKF) are widely used in underwater multi-sensor fusion applications for localization and navigation. Since these filters are designed by employing first-order Taylor series approximation in the error covariance matrix, they result in a decrease in estimation accuracy under high nonlinearity. In order to address this problem, we proposed a novel multi-sensor fusion algorithm for underwater vehicle localization that improves state estimation by augmentation of the radial basis function (RBF) neural network with ESKF. In the proposed algorithm, the RBF neural network is utilized to compensate the lack of ESKF performance by improving the innovation error term. The weights and centers of the RBF neural network are designed by minimizing the estimation mean square error (MSE) using the steepest descent optimization approach. To test the performance, the proposed RBF-augmented ESKF multi-sensor fusion was compared with the conventional ESKF under three different realistic scenarios using Monte Carlo simulations. We found that our proposed method provides better navigation and localization results despite high nonlinearity, modeling uncertainty, and external disturbances.

Gradient-based training and pruning of radial basis function networks with an application in materials physics

Many applications, especially in physics and other sciences, call for easily interpretable and robust machine learning techniques. We propose a fully gradient-based technique for training radial basis function networks with an efficient and scalable open-source implementation. We derive novel closed-form optimization criteria for pruning the models for continuous as well as binary data which arise in a challenging real-world material physics problem. The pruned models are optimized to provide compact and interpretable versions of larger models based on informed assumptions about the data distribution. Visualizations of the pruned models provide insight into the atomic configurations that determine atom-level migration processes in solid matter; these results may inform future research on designing more suitable descriptors for use with machine learning algorithms.

Effective Electricity Theft Detection in Power Distribution Grids Using an Adaptive Neuro Fuzzy Inference System

Electric power grids are a crucial infrastructure for the proper operation of any country and must be preserved from various threats. Detection of illegal electricity power consumption is a crucial issue for distribution system operators (DSOs). Minimizing non-technical losses is a challenging task for the smooth operation of electrical power system in order to increase electricity provider’s and nation’s revenue and to enhance the reliability of electrical power grid. The widespread popularity of smart meters enables a large volume of electricity consumption data to be collected and new artificial intelligence technologies could be applied to take advantage of these data to solve the problem of power theft more efficiently. In this study, a robust artificial intelligence algorithm adaptive neuro fuzzy inference system (ANFIS)—with many applications in many various areas—is presented in brief and applied to achieve more effective detection of electric power theft. To the best of our knowledge, there are no studies yet that involve the application of ANFIS for the detection of power theft. The proposed technique is shown that if applied properly it could achieve very high success rates in various cases of fraudulent activities originating from unauthorized energy usage.

Attractor Ranked Radial Basis Function Network: A Nonparametric Forecasting Approach for Chaotic Dynamic Systems

The curse of dimensionality has long been a hurdle in the analysis of complex data in areas such as computational biology, ecology and econometrics. In this work, we present a forecasting algorithm that exploits the dimensionality of data in a nonparametric autoregressive framework. The main idea is that the dynamics of a chaotic dynamical system consisting of multiple time-series can be reconstructed using a combination of different variables. This nonlinear autoregressive algorithm uses multivariate attractors reconstructed as the inputs of a neural network to predict the future. We show that our approach, attractor ranked radial basis function network (AR-RBFN) provides a better forecast than that obtained using other model-free approaches as well as univariate and multivariate autoregressive models using radial basis function networks. We demonstrate this for simulated ecosystem models and a mesocosm experiment. By taking advantage of dimensionality, we show that AR-RBFN overcomes the shortcomings of noisy and short time-series data.

Improving clinical refractive results of cataract surgery by machine learning

AIM

To evaluate the potential of the Support Vector Machine Regression model (SVM-RM) and Multilayer Neural Network Ensemble model (MLNN-EM) to improve the intraocular lens (IOL) power calculation for clinical workflow.

BACKGROUND

Current IOL power calculation methods are limited in their accuracy with the possibility of decreased accuracy especially in eyes with an unusual ocular dimension. In case of an improperly calculated power of the IOL in cataract or refractive lens replacement surgery there is a risk of re-operation or further refractive correction. This may create potential complications and discomfort for the patient.

METHODS

A dataset containing information about 2,194 eyes was obtained using data mining process from the Electronic Health Record (EHR) system database of the Gemini Eye Clinic. The dataset was optimized and split into the selection set (used in the design for models and training), and the verification set (used in the evaluation). The set of mean prediction errors (PEs) and the distribution of predicted refractive errors were evaluated for both models and clinical results (CR).

RESULTS

Both models performed significantly better for the majority of the evaluated parameters compared with the CR. There was no significant difference between both evaluated models. In the ±0.50 D PE category both SVM-RM and MLNN-EM were slightly better than the Barrett Universal II formula, which is often presented as the most accurate calculation formula.

CONCLUSION

In comparison to the current clinical method, both SVM-RM and MLNN-EM have achieved significantly better results in IOL calculations and therefore have a strong potential to improve clinical cataract refractive outcomes.

Taking advantage of hybrid bioinspired intelligent algorithm with decoupled extended Kalman filter for optimizing growing and pruning radial basis function network

The growing and pruning radial basis function (GAP-RBF) network is a promising sequential learning algorithm for prediction analysis, but the parameter selection of such a network is usually a non-convex problem and makes it difficult to handle. In this paper, a hybrid bioinspired intelligent algorithm is proposed to optimize GAP-RBF. Specifically, the excellent local convergence of particle swarm optimization (PSO) and the extensive search ability of genetic algorithm (GA) are both considered to optimize the weights and bias term of GAP-RBF. Meanwhile, a competitive mechanism is proposed to make the hybrid algorithm choose the appropriate individuals for effective search and further improve its optimization ability. Moreover, a decoupled extended Kalman filter (DEKF) method is introduced in this study to reduce the size of error covariance matrix and decrease the computational complexity for performing real-time predictions. In the experiments, three classic forecasting issues including abalone age, Boston house price and auto MPG are adopted for extensive test, and the experimental results show that our method performs better than PSO and GA these two single bioinspired optimization algorithms. What is more, our method via DEKF achieves the better results in comparison with the state-of-art sequential learning algorithms, such as GAP-RBF, minimal resource allocation network, resource allocation network using an extended Kalman filter and resource allocation network.

Haze Removal Using Radial Basis Function Networks for Visibility Restoration Applications

Restoration of visibility in hazy images is the first relevant step of information analysis in many outdoor computer vision applications. To this aim, the restored image must feature clear visibility with sufficient brightness and visible edges, while avoiding the production of noticeable artifacts. In this paper, we propose a haze removal approach based on the radial basis function (RBF) through artificial neural networks dedicated to effectively removing haze formation while retaining not only the visible edges but also the brightness of restored images. Unlike traditional haze-removal methods that consist of single atmospheric veils, the multiatmospheric veil is generated and then dynamically learned by the neurons of the proposed RBF networks according to the scene complexity. Through this process, more visible edges are retained in the restored images. Subsequently, the activation function during the testing process is employed to represent the brightness of the restored image. We compare the proposed method with the other state-of-the-art haze-removal methods and report experimental results in terms of qualitative and quantitative evaluations for benchmark color images captured in typical hazy weather conditions. The experimental results demonstrate that the proposed method is able to produce brighter and more vivid haze-free images with more visible edges than can the other state-of-the-art methods.

Motion tomography: Mapping flow fields using autonomous underwater vehicles

Since the motion of autonomous underwater vehicles is affected by ambient flow, knowledge of an environmental flow field can be used to improve the navigation of autonomous underwater vehicles. Due to imperfect knowledge of flow, the actual trajectory of an autonomous underwater vehicle deviates from the predicted trajectory. The difference between the actual and predicted trajectories is referred to as the motion-integration error, providing information of flow along the vehicle trajectory. Inspired by computerized tomography, this paper proposes motion tomography, a tomographic method for creating a fine-grid spatial map of flow based on the motion-integration error. While typical computerized tomography is a linear problem, motion tomography is a nonlinear problem because of unknown nonlinear trajectories of autonomous underwater vehicles and the dependency of the trajectories on the flow field. Therefore, motion tomography employs an iterative process consisting of two alternating steps: Trajectory tracing and flow field estimation. Starting from an initial guess of the flow field, in the trajectory tracing step, unknown nonlinear vehicle trajectories are estimated. Then, using the estimated vehicle trajectories, a spatial map of flow is constructed through either the non-parametric or parametric flow field estimation. The error bound for trajectory tracing is computed and the convergence of both the non-parametric and parametric flow field estimation algorithms is proved. Simulation and experimental data are analyzed to evaluate the performance of motion tomography when subject to changing vehicle speed and flow variability.

A growing and pruning sequential learning algorithm of hyper basis function neural network for function approximation

Radial basis function (RBF) neural network is constructed of certain number of RBF neurons, and these networks are among the most used neural networks for modeling of various nonlinear problems in engineering. Conventional RBF neuron is usually based on Gaussian type of activation function with single width for each activation function. This feature restricts neuron performance for modeling the complex nonlinear problems. To accommodate limitation of a single scale, this paper presents neural network with similar but yet different activation function-hyper basis function (HBF). The HBF allows different scaling of input dimensions to provide better generalization property when dealing with complex nonlinear problems in engineering practice. The HBF is based on generalization of Gaussian type of neuron that applies Mahalanobis-like distance as a distance metrics between input training sample and prototype vector. Compared to the RBF, the HBF neuron has more parameters to optimize, but HBF neural network needs less number of HBF neurons to memorize relationship between input and output sets in order to achieve good generalization property. However, recent research results of HBF neural network performance have shown that optimal way of constructing this type of neural network is needed; this paper addresses this issue and modifies sequential learning algorithm for HBF neural network that exploits the concept of neuron's significance and allows growing and pruning of HBF neuron during learning process. Extensive experimental study shows that HBF neural network, trained with developed learning algorithm, achieves lower prediction error and more compact neural network.