Loo Chu Kiong

Abstract: Action recognition is important in the eld of intelligent security and surveil- xD;lance. However, most surveillance cameras can only capture in one direction with limited viewing angle. This paper proposes an edge enhancement template-based method of omnidirectional action recognition that is able to detect specic actions at a 360 degree of view. A MACH lter captures intra-class variability by synthesizing a single action MACH lter for a given action class. The proposed method, based on the wavelet MACH lter, provides additional xD;exibility of an adaptive choice of wavelet scale factors and, in doing so, enables the selection of the size and orientation of the smoothing function in edge enhancement to optimize the performance of the MACH lter. Moreover, the xD;use of wavelet transform improves the performance of the MACH lter by enhancing the cross-correlation peak intensity in the recognition process. The unwarping of an omnidirectional image into a panoramic image further enables action recognition in 360 degree wide angle of view.

Abstract: The head pose and movement of a user is closely related with his/her intention and thought, recognition of such information could be useful to develop a natural and sensitive user-wheelchair interface. This paper presents an original integrated approach to a head gesture based interface (HGI) which can perform both identity verification and facial pose estimation. Identity verification is performed by two-factor face authentication which is implemented by the combination of topographic independent component analysis (TICA) and multispace random projection (MRP). Modified synergetic computer with melting (Modified SC-MELT) is introduced to classify facial poses. Motion profile generator (MPG) is thoroughly developed during the integration to convert each estimated facial pose sequence into motion control signal to actuate motor movements. The HGI is intended to be deployed as a user-wheelchair interface for disabled and elderly users in which only users with genuine face and valid token may be granted authorized access and hence pilot an electric powered wheelchair (EPW) using their faces. The integration has been verified under a number of experiments to justify the feasibility and performance of the proposed face-based control strategy. (C) 2011 Elsevier B.V. All rights reserved.

Abstract: A method for assessing balance, which was sensitive to changes in the postural control system is presented. This paper describes the implementation of a force-sensing platform, with force sensing resistors as the sensing element. The platform is capable of measuring destabilized postural perturbations in dynamic and static postural conditions. Besides providing real-time qualitative assessment, the platform quantifies the postural control of the subjects. This is done by evaluating the weighted center of applied pressure distribution over time. The objective of this research was to establish the feasibility of using the force-sensing platform to test and gauge the postural control of individuals. Tests were conducted in Eye Open and Eye Close states on Flat Ground (static condition) and the balance trainer (dynamic condition). It was observed that the designed platform was able to gauge the sway experienced by the body when subject�s states and conditions changed.

Abstract: A brain computer interface BCI enables direct communication between a brain and a computer translating brain activity into computer commands using preprocessing, feature extraction, and classification operations. Feature extraction is crucial, as it has a substantial effect on the classification accuracy and speed. While fractal dimension has been successfully used in various domains to characterize data exhibiting fractal properties, its usage in motor imagery-based BCI has been more recent. In this study, commonly used fractal dimension estimation methods to characterize time series Katz's method, Higuchi's method, rescaled range method, and Renyi's entropy were evaluated for feature extraction in motor imagery-based BCI by conducting offline analyses of a two class motor imagery dataset. Different classifiers fuzzy k-nearest neighbours FKNN, support vector machine, and linear discriminant analysis were tested in combination with these methods to determine the methodology with the best performance. This methodology was then modified by implementing the time-dependent fractal dimension TDFD, differential fractal dimension, and differential signals methods to determine if the results could be further improved. Katz's method with FKNN resulted in the highest classification accuracy of 85%, and further improvements by 3% were achieved by implementing the TDFD method.

Abstract: In this paper, a new optical approach based on omnidirectional thermal xD;visualization system is proposed. It will provides observer or image processing xD;tool a 360 degree viewing of surrounding area using a single thermal camera. By xD;applying the proposed omnidirectional thermal visualization system even in poor lighting condition, surrounding area is under proper surveillance and the xD;surrounding heating machineries/items can be monitored indeed. Infrared(IR) xD;reflected hyperbolic mirrors have been designed and custom made for the xD;purpose of reflecting omnidirectional scenes in infrared range for the surrounding area to be captured on a thermal camera, thus producing omnidirectional thermal visualization images. Five cost effective and market-common IR reflected materials used to fabricate the designed hyperbolic mirror are studied, i.e. stainless steel, mild steel, aluminum, brass, and chromium. Among these materials, chromium gives the best IR reflectivity, withεr = 0.985. Specifically, we introduce log-polar mapping for unwarping the captured omnidirectional thermal image into a panoramic view, hence providing observers or image processing tools a complete wide angle of view. Three mapping techniques are proposed in this paper namely the point sample, mean sample and interpolation mapping techniques. Point sample mapping technique provides the greatest interest due to its lower complexity and moderate output image quality.

Abstract: Log-polar or spatially variant image representation is an important component of active vision system in tracking process for many robotic applications due to its data compression ability, faster sampling rates and hence, direct to faster image-processing speed in machine vision system. In this paper, we try to implement log-polar mapping techniques on Xilinx FPGA (Field Programmable Gate Array) board for unwarping the omnidirectional images into panoramic images to present a wide angle of view by preserving fine output image quality in a higher data compression manner. Simulations are also run on MATLAB to find out the optimum mapping criterion. Some significant advantages of this new approach are: lighter processing system, lesser space utilisation, cost saving, faster processing speed and faster reset time (boot time) compared to a laptop computer that uses MATLAB for doing the unwarping process.

Abstract: Gabor wavelet is considered the best mathematical descriptor for receptive fields in the striate cortex. Besides, as a basis function, it is suitable to sparsely represent natural scenes due to its property in maximizing information. It is argued that Gabor-like receptive fields are emerged by sparseness-enforcing or infomax method. In this paper, we incorporate Gabor overcomplete representation into quantum holography for image recognition tasks, with suggestions in improvements through iterative method for reconstruction.

Abstract: Nowadays most factories rely on machines to help boost up their production and process Therefore an effective machine condition monitoring system plays an important role in these factories to ensure that their production and process are running smoothly all the time In this paper a new and effective machine condition monitoring system using log-polar mapper quaternion based thermal image correlator and max-product fuzzy neural network classifier is proposed Two classification characteristics namely peak to sidelobe ratio (PSR) and real to complex ratio of the discrete quaternion correlation output (p-value) are applied in this proposed machine condition monitoring system Large PSR and p-value showed a good match among correlation of the input thermal image with a particular reference image but reversely for small PSR and p-value match In the simulation log-polar mapping is found to have solved the rotation and scaling invariant problems in quaternion based thermal image correlation Besides log-polar mapping can possess two fold data compression capability Log-polar mapping helps smoothen up the output correlation plane hence making better measurement for PSR and p-values The simulation results have also proven that the proposed system is an efficient machine condition monitoring system with an accuracy of more than 94% (C) 2010 Elsevier B V All rights reserved

Abstract: Gabor wavelet is considered the best mathematical descriptor for receptive fields in the striate cortex. As a basis function, it is suitable to sparsely represent natural scenes due to its property in maximizing information. It is argued that Gabor-like receptive fields emerged by the sparseness-enforcing or infomax method, with sparseness-enforcing being more biologically plausible. This paper incorporates Gabor over-complete representation into Quantum Holography for image recognition tasks. Correlations are performed using sampled result from all frequencies as well as the optimum frequency. Correlation is also performed using only those points of least activity, which shows improvements in recognition. Analysis on the use of conjugation in reconstruction is provided. The authors also suggest improvements through iterative methods for reconstruction.

Abstract: A fuzzy model based on an enhanced supervised fuzzy clustering algorithm is presented in this paper. The supervised fuzzy clustering algorithm [6] allows each rule to represent more than one output with different probabilities for each output. This algorithm implements k-means to initialize the fuzzy model. However, the main drawbacks of this approach are that the number of clusters is unknown and the initial positions of clusters are randomly generated. In this work, the initialization is done by the global k-means algorithm [1], which can autonomously determine the actual number of clusters needed and give a deterministic clustering result. In addition, the fast global k-means algorithm [1] is presented to improve the computation time. The model is tested on medical diagnosis benchmark data and Westland vibration data. The results obtained show that the model that uses the global k-means clustering algorithm [1] has higher accuracy when compared to a model that uses the k-means clustering algorithm. Besides that, the fast global k-means algorithm [1] also improved the computation time without degrading much the model performance.

Abstract: Thermography, or thermal visualization is a type of infrared visualization. Thermographic cameras are used in many heavy factories like metal recycling factories, wafer production factories and etc for monitoring the temperature conditions of the machines. Besides, thermographic camera can be used to detect trespassers in environment with poor lighting condition, whereby, the xD;conventional digital cameras are less applicable in. This paper proposed an efficient omnidirectional surveillance system using thermal camera. In this surveillance system, the omnidirectional scenes in a machine room, production plant, pump house, laboratory, etc within a factory site are first captured using a thermal camera attached to a custom made hyperbolic IR (infrared radiation) reflected mirror. The captured scenes with some machines to be monitored are then fed into a laptop computer for image processing and alarm purposes. Log-polar mapping is proposed to map the captured omnidirectional thermal image into panoramic image, hence providing the observer or image processing tools a complete wide angle of view. Two simple and fast detection algorithms are embedded into the thermal imaging surveillance system. This surveillance system is not only used for monitoring the functioning condition of different xD;machines/items in a factory site, but can also use for detecting the trespassers in a poor lighting condition. The observed significances of this new proposed omnidirectional thermal imaging system include: it can cover a wide angle of view (360° omnidirectional), using minimum hardware, low cost and the output thermal images are with higher data compression. Experimental results show that the proposed surveillance system achieves high accuracy in monitoring machines conditions and detecting trespassers.

Abstract: Aiming at the implementation of brain�machine interfaces (BMI) for the aid of disabled people, this paper presents a system design for real-time communication between the BMI and programmable logic controllers (PLCs) to control an electrical actuator that could be used in devices to help the disabled. Motor imaginary signals extracted from the brain's motor cortex using an electroencephalogram (EEG) were used as a control signal. The EEG signals were pre-processed by means of adaptive recursive band-pass filtrations (ARBF) and classified using simplified fuzzy adaptive resonance theory mapping (ARTMAP) in which the classified signals are then translated into control signals used for machine control via the PLC. A real-time test system was designed using MATLAB for signal processing, KEP-Ware V4 OLE for process control (OPC), a wireless local area network router, an Omron Sysmac CPM1 PLC and a 5 V/0.3 A motor. This paper explains the signal processing techniques, the PLC's hardware configuration, OPC configuration and real-time data exchange between MATLAB and PLC using the MATLAB OPC toolbox. The test results indicate that the function of exchanging real-time data can be attained between the BMI and PLC through OPC server and proves that it is an effective and feasible method to be applied to devices such as wheelchairs or electronic equipment.

Abstract: This paper introduces a new approach called hybrid particle swarm optimization like algorithm (hybrid PSO) with fine tuning operators to solve optimisation problems. This method combines the merits of the parameter-free PSO (pf-PSO) and the extrapolated particle swarm optimization like algorithm (ePSO). In order to accelerate the PSO algorithms to obtain the global optimal solution, three fine tuning operators, namely mutation, cross-over and root mean square variants are introduced. The effectiveness of the fine tuning elements with various PSO algorithms is tested through three benchmark functions along with a few recently developed state-of-the-art methods and the results are compared with those obtained without the fine tuning elements. From several comparative analyses, it is clearly seen that the performance of all the three PSO algorithms (pf-PSO, ePSO, and hybrid PSO) is considerably improved with various fine tuning operators and sometimes more competitive than the recently developed PSO algorithms.

Abstract: This paper presents the expectation-maximization (EM) variant of probabilistic neural network (PNN) as a step toward creating an autonomous and deterministic PNN. In the real world, faulty reading sensors can happen and will create input vectors with missing features yet they should not be discarded. To overcome this, regularized EM is put in place as a preprocessing step to impute the missing values. The problem faced by users when using random initialization is that they have to define the number of clusters through trial and error, which makes it stochastic in nature. Global k-means is used to autonomously find the number of clusters using a selection criterion and deterministically provide the number of clusters needed to train the model. In addition, fast Global k-means will be tested as an alternative to Global k-means to help reduce computational time. Tests are conducted on both homoscedastic and heteroscedastic PNNs. Benchmark medical datasets and also vibration data collected from a US Navy CH-46E helicopter aft gearbox known as Westland were used. The tests' results fully support the usage of fast Global k-means and regularized EM as preprocessing steps to aid the EM-trained PNN.

Abstract: In recent years, many researchers have studied the generation of rhythmic movement emerge as a stable limit cycle from global entrainment between biological neural networks that include central pattern generators (CPGs) and physical systems (i.e. robot arm) interacting with external environment. However, the e�ect of global entrainment on biped locomotion energy has yet to be addressed. This paper delineates an approach to minimize biped�s locomotion energy by means of global entrainment. The rhythmic movements for biped locomotion such as walking and running are generated by CPG using coupled nonlinear oscillators of Van Der Pol (VDP). Tuning VDP oscillator parameters to meet certain criteria (i.e. minimum energy) is a di�cult task to accomplish. This is due to the nonlinearity and the coupling of the oscillators. To overcome these di�culties, response surface methodology (RSM) has been proposed to characterize VDP parameters to achieve e�cient energy for biped locomotion. Then, global entrainment has been implemented between �ve-link biped robot and the constructed CPG. As a result, the sensory information modulates VDP oscillators� frequencies and tuned them to the resonance frequencies of the biped link lengths. The obtained results show an evidence of the importance of global entrainment in achieving lower energy for biped locomotion while maintaining the same or even higher forward velocity.

Abstract: The computation of optimal control variables for a two-stage steel annealing process which comprises of one or more furnaces is proposed in this paper. The heating and soaking furnaces of the steel annealing line form the two-stage hybrid systems. Three algorithms including particle swarm optimisation (PSO) with globally and locally tuned parameters (GLBest PSO), a parameter free PSO algorithm (pf-PSO) and a PSO-like algorithm via extrapolated PSO (ePSO) are considered to solve this optimal control problem for the two-stage steel annealing processes (SAP). The optimal solutions including optimal line speed, optimal cost and job completion time obtained through these three methods are compared with one another and those obtained via conventional PSO (cPSO) with time varying inertia weight (TVIW) and time varying acceleration coefficient (TVAC). From the results obtained through the five algorithms considered, the efficacy and validity of each algorithm are analysed.

Abstract: This paper focuses on the statistical based Probabilistic Neural Network (PNN) for pattern classification problems with Expectation � Maximization (EM) chosen as the training algorithm. This brings about the problem of random initialization, which means, the user has to predefine the number of clusters through trial and error. Global k-means is used to solve this and to provide a deterministic number of clusters using a selection criterion. On top of that, Fast Global k-means was tested as a substitute for Global k-means, to reduce the computational time taken. Tests were done on both homescedastic and heteroscedastic PNNs using benchmark medical datasets and also vibration data obtained from a U.S. Navy CH-46E helicopter aft gearbox (Westland)

Abstract: The noteworthy point in the advancement of Brain xD;Machine Interface (BMI) research is the ability to accurately extract xD;features of the brain signals and to classify them into targeted control xD;action with the easiest procedures since the expected beneficiaries xD;are of disabled. In this paper, a new feature extraction method using xD;the combination of adaptive band pass filters and adaptive xD;autoregressive (AAR) modelling is proposed and applied to the xD;classification of right and left motor imagery signals extracted from xD;the brain. The introduction of the adaptive bandpass filter improves xD;the characterization process of the autocorrelation functions of the xD;AAR models, as it enhances and strengthens the EEG signal, which xD;is noisy and stochastic in nature. The experimental results on the xD;Graz BCI data set have shown that by implementing the proposed xD;feature extraction method, a LDA and SVM classifier outperforms xD;other AAR approaches of the BCI 2003 competition in terms of the xD;mutual information, the competition criterion, or misclassification xD;rate.

Abstract: Here we present a novel approach to detect P300 wave in single trial Visual Event Related Potential (VERP) signals using improved principal component analysis to enable a faster brain-computer interface (BCI) design. In the process, the principal components (PCs) are selected using novel methods, namely spectral power ratio (SPR) and sandwich spectral power ratio (SSPR). We set out to assess the improved performances of our proposed methods, SPR and SSPR over standard PC selection methods like Kaiser and residual power for speller BCI design. Concluding, the P300 parameters extracted through our proposed SPR and SSPR methods showed improved detection of target characters in the speller BCI.

Abstract: The critical analysis of the data glove-based signature identification and forgery detection system emphasizes the essentiality of noise-free signals for input. Lucid inputs are expected for the accuracy enhancement and performance. The raw signals that are captured using 14- and 5-electrode data gloves for this purpose have a noisy and voluminous nature. Reduction of electrodes may reduce the volume but it may also reduce the efficiency of the system. The principal component analysis (PCA) technique has been used for this purpose to condense the volume and enrich the operational data by noise reduction without affecting the efficiency. The advantage of increased discernment in between the original and forged signatures using 14-electrode glove over 5-electrode glove has been discussed here and proved by experiments with many subjects. Calculation of the sum of mean squares of Euclidean distance has been used to project the advantage of our proposed method. 3.1% and 7.5% of equal error rates for 14 and 5 channels further reiterate the effectiveness of this technique.

Abstract: Data analysis is indispensable for engineering, for data is the only link between the theory and the reality. Traditional data analysis methods such as Fourier analysis are all based on linear and stationary assumptions i.e. the signal to be processed must be linear and temporally stationary; otherwise, the resulting Fourier spectrum will make little physical sense. Practical signals such as speech, machine vibrations, biomedical measurement and communications most likely to be both nonlinear and non-stationary. Hence new methods are needed to analyze the data from nonlinear and non-stationary process. Hilbert-Huang Transform(HHT) is a new data processing technology developed by NASA Goddard Space Flight Center. The HHT is derived from the principles of empirical mode decomposition (EMD) and the Hilbert Transform. This paper presents the suitability of using HHT for nonlinear and non-stationary data analysis. The efficiency of the new method is tested on a set of vibration data collected from Westland helicopter gearbox, a non-linear and non-stationary system. Both simulation and the experimental results indicate that this new method HHT is more suitable for non-linear and non-stationary process.

Abstract: This Study investigates the processing of sonar signals with ensemble neural networks for robust recognition of simple objects such as plane, corner and trapezium surface. The ensemble neural networks can differentiate the target objects with high accuracy. The simplified fuzzy ARTMAP (SFAM) and probabilistic ensemble simplified fuzzy ARTMAP (PESFAM) are compared in terms of classification accuracy. The PESFAM implements an accurate and effective probabilistic plurality voting method to combine outputs from multiple SFAM classifiers. Five benchmark data sets have been used to evaluate the applicability of the proposed ensemble SFAM network. The PESFAM achieves good accuracy based on the twofold cross-validation results. In addition, the effectiveness of the proposed ensemble SFAM is delineated in sonar target differentiation. The experiments demonstrate the potential of PESFAM classifiers in offering an optimal solution to the data-ordering problem of SFAM implementation and also as an intelligent classification tool in mobile robot application.

Abstract: This paper presents a new approach via hybrid particle swarm optimization (HPSO) scheme to solve the unit commitment (UC) problem. HPSO proposed in this paper is a blend of binary particle swarm optimization (BPSO) and real coded particle swarm optimization (RCPSO). The UC problem is handled by BPSO, while RCPSO solves the economic load dispatch problem. Both algorithms are run simultaneously, adjusting their solutions in search of a better solution. Problem formulation of the UC takes into consideration the minimum up and down time constraints, start-up cost, and spinning reserve and is defined as the minimization of the total objective function while satisfying all the associated constraints. Problem formulation, representation, and the simulation results for a ten generator-scheduling problem are presented. Results clearly show that HPSO is very competent in solving the UC problem in comparison to other existing methods.

Abstract: Most supervised neural networks are trained by minimizing the mean square error (MSE) of the training set. In the presence of outliers, the resulting neural network model can differ significantly from the underlying model that generates the data. This paper outlines two robust learning methods for a dynamic structure neural network called incremental growing multi-experts network (IGMN). It is convincingly shown by simulation that by using a scaled robust objective function instead of the least squares function, the influence of the outliers in the training data can be completely eliminated. The network generates a much better approximation in the neighborhood of outliers. Thus, the two proposed robust learning methods namely robust least mean squares (RLMSs) and least mean log squares (LMLSs) are insensitive to the presence of outliers unlike the least mean squares (LMSs) cost function. Moreover, various types of supervised learning algorithms can easily adopt LMLS, which is a parameter-free method.

Abstract: Theoretical and simulational evidence, as well as experimental indications, are accumulating that quantum associative memory and imaging are possible. We compare these data with biological evidence, since we find them to a significant extent compatible. This paper presents a computationally implementable integrative model of appearance-based viewpoint-invariant recognition of objects. The neuro-quantum hybrid model incorporates neural processing up to V1 and quantum associative processing in V1, achieving together an object-recognition result in V2 and ITC. Results of our simulation of the central quantum-like parts of the bio-model, receiving neurally pre-processed inputs, are presented. This part contains our original simulated storage by multiple quantum interference of image-encoding Gabor wavelets done in a Hebbian way, especially using the Griniasty et al. pose-sequence learning rule.

Abstract: A holographic experimental procedure assuming use of quantum states of light is simulated. It uses merely interference-based image storage and nonunitary image retrieval realized by wave function collapse. Successful results of computational view-invariant recognition of object images are presented. As in neural net theory, recognition is selective reconstruction of an image from a database of many concrete images (simultaneously stored in an associative memory) after presentation of a different version of that image. That is, in the first step, we store many high-resolution images of objects into quantum memory (a hologram). In the second step, we present a �nonlearned� noisy image version. We thereby trigger memory-influenced reorganization of the state of the system so that it finally encodes those corrected object images that correspond to the newly presented version. The holographic procedure seems to be implementable with present-day quantum optics.

Abstract: In this paper, the application of Active Force Control (AFC) incorporated with a conventional Proportional-Derivative (PD) controller to a five-link biped robot has been studied and simulated. The efficacy and robustness of the AFC strategy in suppressing external disturbances was examined on a model using Crude Approximation (CA) method. However, the task of tuning the PD controller parameters and the inertia matrix coefficient is tedious and time-consuming. Thus, an evolutionary strategy - Differential Evolution (DE) has been proposed in order to tune automatically the parameter gains in a systematic approach. The effectiveness of the proposed method is investigated, and it is found that the system is robust and stable even under influence of disturbances

Abstract: In this paper, an accurate and effective probabilistic plurality voting method to combine outputs from multiple simplified fuzzy ARTMAP (SFAM) classifiers is presented. Five ELENA benchmark problems and five medical benchmark data sets have been used to evaluate the applicability and performance of the proposed probabilistic ensemble simplified fuzzy ARTMAP (PESFAM) network. Among the five benchmark problems in ELENA project, PESFAM outperforms the SFAM and multi-layer perceptron (MLP) classifier. In addition, the effectiveness of the proposed PESFAM is delineated in medical diagnosis applications. For the medical diagnosis and classification problems, PESFAM achieves 100 percent in accuracy, specificity, and sensitivity based on the 10-fold crossvalidation and these results are superior to those from other classification algorithms. In addition, a posteri probability of the predicted class can be used to measure the prediction reliability of PESFAM. The experiments demonstrate the potential of the proposed multiple SFAM classifiers in offering an optimal solution to the data-ordering problem of SFAM implementation and also as an intelligent medical diagnosis tool.

Abstract: It is computationally demonstrated how quantum associative networks, implemented using quantum holography, could be harnessed for object recognition. These simulated quantum nets alone execute efficient image recognition, i.e., reconstruction of an image selected from associative memory (hologram). However, optically implementable neural-net preprocessing of object-images is needed for appearance-based viewpoint-invariant recognition of objects. We present computer simulation results of two methods: Moore-Penrose orthogonalization and encoding of object-images with Gabor wavelets. A computer-supported quantum Gabor-wavelet holography is proposed.

Abstract: This paper, written for interdisciplinary audience, presents computational image reconstruction implementable by quantum optics. The input-triggered selection of a high-resolution image among many stored ones, and its reconstruction if the input is occluded or noisy, has been successfully simulated. The original algorithm, based on the Hopfield associative neural net, was transformed in order to enable its quantum-wave implementation based on holography. The main limitations of the classical Hopfield net are much reduced with the simulated new quantum-optical implementation.

Abstract: The shortest/optimal path generation is essential for the efficient operation of a mobile robot. Recent advances in robotics and machine intelligence have led to the application of modern optimization method such as the genetic algorithm (GA), to solve the path-planning problem. However, the genetic algorithm path planning approach in the previous works requires a preprocessing step that captures the connectivity of the free-space in a concise representation. In this paper, GA path-planning approach is enhanced with feasible path detection mechanism based on traversability vectors method. This novel idea eliminates the need of free-space connectivity representation. The feasible path detection is xD;performed concurrently while the GA performs the search for the shortest path. The performance of the proposed GA approach is tested on three different environments consisting of polygonal obstacles with increasing complexity. In all experiments, the GA has successfully detected the near-optimal feasible traveling path for mobile. xD;

Abstract: A quantum associative memory, much more natural than those of �quantum computers�, is presented. Neural-net-like processing with real-valued variables is transformed into processing with quantum waves. Successful computer simulations of image storage and retrieval are reported. Our Hopfield-like algorithm allows quantum implementation with holographic procedure using present-day quantum-optics techniques. This brings many advantages over classical Hopfield neural nets and quantum computers with logic gates.

Abstract: This work presents two novel approaches to determine optimum growing multi-experts network (GMN) structure. The first method called direct method deals with expertise domain and levels in connection with local experts. The growing neural gas (GNG) algorithm is used to cluster the local experts. The concept of error distribution is used to apportion error among the local experts. After reaching the specified size of the network, redundant experts removal algorithm is invoked to prune the size of the network based on the ranking of the experts. However, GMN is not ergonomic due to too many network control parameters. Therefore, a self-regulating GMN (SGMN) algorithm is proposed. SGMN adopts self-adaptive learning rates for gradient-descent learning rules. In addition, SGMN adopts a more rigorous clustering method called fully self-organized simplified adaptive resonance theory in a modified form. Experimental results show SGMN obtains comparative or even better performance than GMN in four benchmark examples, with reduced sensitivity to learning parameters setting. Moreover, both GMN and SGMN outperform the other neural networks and statistical models. The efficacy of SGMN is further justified in three industrial applications and a control problem. It provides consistent results besides holding out a profound potential and promise for building a novel type of nonlinear model consisting of several local linear models.

Abstract: This paper deals with a novel idea of identification of nonlinear dynamic systems via a constructivism inspired neural network. The proposed network is known as growing multi-experts network (GMN). In GMN, the problem space is decomposed into overlapping regions by expertise domain and local expert models are graded according to their expertise level. The network output is computed by the smooth combination of local linear models. In order to avoid over-fitting problem, GMN deploys a redundant experts removal algorithm to remove the redundant local experts from the network. In addition, growing neural gas (GNG) algorithm is used to generate an induced Delaunay triangulation that is highly desired for optimal function approximation. A variety of examples are taken from literature to establish the efficacy of GMN. Discrete time nonlinear dynamic system modeling and water bath temperature control have been found to give excellent results via this novel neural network.

Abstract: Artificial neural networks (ANNs) have been used to construct empirical nonlinear models of process data. Because networks are not based on physical theory and contain nonlinearities, their predictions are suspect when extrapolating beyond the range of original training data. Standard networks give no indication of possible errors due to extrapolation. This paper describes a sequential supervised learning scheme for the recently formalized Growing multi-experts network (GMN). It is shown that certainty factor can be generated by GMN that can be taken as extrapolation detector for GMN. On-line GMN identification algorithm is presented and its performance is evaluated. The capability of the GMN to extrapolate is also indicated. Four benchmark experiments are dealt with to demonstrate the effectiveness and utility of GMN as a universal function approximator.

Abstract: This paper investigates the efficacy of the implementation of the conventional Proportional-Derivative (PD) controller and different Active Force Control (AFC) strategies to a 5-link biped robot through a series of simulation studies. The performance of the biped system is evaluated by making the biped walk on a horizontal flat surface, in which the locomotion is constrained within the sagittal plane. Initially, a classical PD controller has been used to control the biped robot. Then, a disturbance elimination method called Active Force Control (AFC) schemes has been incorporated. The effectiveness and robustness of the AFC as �disturbance rejecter� has been examined when a conventional crude approximation (AFCCA), and an intelligent active force control scheme, which is known as Active Force Control and Iterative Learning (AFCAIL) are employed. It is found that for both of the AFC control schemes proposed, the system is robust and stable even under the influence of disturbances. An attractive feature of the AFCAIL scheme is that inertia matrix tuning becomes much easier and automatic without any degradation in the performance.

Abstract: This paper presents a Hybrid Particle Swarm Optimization (HPSO) to solve the Unit Commitment (UC) problem. Problem formulation of the unit commitment takes into consideration the minimum up and down time constraints, start up cost and spinning reserve, which is defined as the minimization of the total objective function while satisfying all the associated constraints. Problem formulation, representation and the simulation results for a 10 generator-scheduling problem are presented. Results shown are acceptable at this early stage.

Abstract: An endeavor is made in this paper to describe a self-regulating constructive multi-model neural network called Self-regulating Growing Multi-Experts Network (SGMN) that can approximate an unknown nonlinear function from observed input-output training data. The proposed network is devised to overcome the redundancy problems of Gaussian neural networks that use square mesh partition method. In the SGMN, the problem space is decomposed into overlapping regions by expertise domain and the local expert models are graded according to their expertise level. The network output is computed by a smooth combination of local polynomial models. In order to avoid an over-fitting problem, the SGMN deploys a Redundant Experts Removal Algorithm to remove the redundant local experts from the network. In addition, the Fully Self-Organized Simplified Adaptive Resonance Theory (FOSART) is modified and adopted to generate an induced Delaunay triangulation that is highly desired for optimal function approximation. Self-adaptive learning rates Gradient Descent learning rules are employed in a supervised learning phase. A parametric control at epoch terminations and performance based on local incremental experts insertions are incorporated. A variety of examples is solved from literature to establish the efficacy of SGMN. Discrete time nonlinear dynamic system modeling and water bath temperature control have been found to give excellent results via this novel neural network.