Search

Peritoneal dialysis (PD) removes waste products from blood when the kidney is malfunctioned. Since there is no clear criterion for PD recommendation for patients with kidney disease, existing machine learning models (ML), which rely on credible decision criterion, are ineffective in making PD eligibility decisions, especially when the correlated traits or indicators (patterns) inherent in the PD data are diverse and subtle. Furthermore, the lack of interpretable transparency in traditional ML also weakens the credibility of the decision they produce. Hence, an in-depth knowledge of the patients’ characteristics is needed to render a clearer picture of the decision-making process and model to detect the rare PD eligibility cases. In this paper, we extend our previous work (Attribute-Value-Association Discovery and Disentanglement (ADD)), to an extended ADD for PD data analysis (PD-ADD) to overcome these problems. We show that PD-ADD is able to discover association patterns of patient profiles and symptoms to reveal PD characteristics and detect eligible rare cases. Experimental results show that PDADD is much superior to existing unsupervised clustering (with accuracy of 89.87% vs 73.37% of K-Means). It also enables straightforward interpretation of the underlying relations of patient characteristics in an unsupervised setting.

Pattern mining has been widely used to uncover interesting patterns from data. However, one of its main problems is that it produces too many patterns and many of them are redundant. To reduce the number of redundant patterns and retain overlapping ones, delta-closed pattern pruning was introduced, yet it can only prune subpatterns if they are covered by superpatterns. Such unduly superpatterns need to be pruned. Furthermore, in order to improve the management and interpretation of patterns, pattern summarization is proposed. It renders a small number of patterns that retain the most crucial information. RuleCover algorithm was one of such algorithms. However, it tends to produce over trivial patterns, whereas more interesting and revealing ones may be pruned. To overcome these problems, this paper presents a new algorithm which integrates delta-closed, and RuleCover methods with our other two new algorithms: 1) statistically induced pattern pruning for pruning statistically induced superpatterns by strong subpatterns and 2) AreaCover algorithm for pruning overlapping patterns but retain higher order and high quality patterns with large coverage of the data “area.” Experimental results show that the proposed algorithms produce very compact yet comprehensive knowledge from patterns discovered from relational data sets.

Proteins from the same family have similar functions. Hence, it is important to discover from a protein family conserved sequence patterns with variations to unveil the functionality of a functional domain. Aligned Pattern Clusters (APCs) are knowledge-rich representations comparing with probabilistic models. If significant aligned residue associations (ARAs) were discovered in APCs, they could reveal subtle functional or subgroup characteristics. However, when ARAs corresponding to different subgroups/classes were entangled due to certain subtle factors, to disentangle them to reveal succinct ARA groups is a big challenge. This paper presents a novel method known as Aligned Residual Association Discovery and Disentanglement (ARADD), to meet such challenge. ARADD first constructs an ARA Frequency Matrix (ARAFM) and converts it into a Statistical Residual (SR) Vector Space (SRV) to suppress noise. SR measures the deviation of the observed frequency of an event against that when the occurrence is random. By applying Principal Component Decomposition (PCD) on the SRV, we obtain PCs ranked by their variance. The ARAs of an AR with others can be represented by an AR-vector whose coordinates account for its associations with others. When the projection of an AR vector on a PC Space is reprojected to the SRV (abbreviated by RSRVs), its coordinates reflect the SRs of that AR associating with other ARs. Experiments showed that the ARADD can a) disentangle entangled ARAs in APCs, b) reveal subtle AR clusters relating to classes or ARAs within or between subgroups — significant to proteomic research, drug discovery and personalized medicine

Functional region identification is of fundamental importance for protein sequences analysis for a protein family. Such knowledge not only provides a better scientific understanding but also assists drug discovery. Domain annotation is one approach but it needs to leverage existing databases. For de novo discovery, motif discovery locates and aligns locally similar sub-sequences and represents them as a position-weight matrix (PWM). However, PWM is a fixed-length model whereas protein functional region size varies. Furthermore, to obtain a PWM, a width range parameter needs to be identified through exhaustive search. Hence, it is computational intensive for large dataset. This paper presents a new method known as Pattern-Directed Aligned Pattern Clustering (PD-APCn) to discover and align residues in conserved protein functional regions. It adopts Aligned Pattern Cluster (APC) as the representation model which allows variable pattern length. It uses patterns with strong support to direct the incremental expansion of the APCs, allowing substitution and frame-shift mutations, until a robust termination condition is reached. The concept of breakpoint gap is introduced to identify uncovered conserved patterns with substitution and frame-shift mutations, where these are often rare mutants. To evaluate the performance of PD-APCn, we conducted experiments on synthetic datasets with different size and noise level. Comparing with the popular motif discovery algorithm MEME, PD-APCn has demonstrated competitive performance throughout the experiments, obtaining a higher recall and Fmeasure with up to 400× significant computational speed up comparing to MEME.

Understanding Protein-protein interaction (PPI) is of fundamental importance in deciphering cellular processes. Predicting PPIs is thus critical in making new discoveries in the biological domains. Traditionally, new PPIs are identified through biochemical experiments but such methods are labor-intensive, expensive, time-consuming and technically ineffective due to high false positive rates. Computational docking is an alternative but requires the three-dimensional structures of the target proteins which are not always accessible. Sequence-based prediction is the most readily applicable and cost-effective method. It exploits known PPI Databases to construct classifiers for predicting unknown PPIs based only on sequence data. However, existing methods, adopting features that fix the pattern length and use exact patterns, are biologically unrealistic. Also, those based on SVM and String Kernel are hardly biologically interpretable since they do not compute the features. Recently, we have developed a new method for predicting PPI known as WeMine-P2P based on our WeMine Aligned Pattern Clustering algorithm which discovers and identifies the localized and co-occurring conserved patterns and regions allowing variable length and pattern variations. As our first attempt, under 40 independent experiments, we showed that (1) WeMine-P2P outperforms the well-known algorithm, PIPE2 which also utilizes co-occurring amino acid sequence segments but does not allow variable lengths and pattern variations; (2) Unlike SVM-based methods, WeMine-P2P renders interpretable biological features; (3) WeMine-P2P achieves satisfactory PPI prediction performance, comparable to the SVM-based methods particularly in unseen protein sequences, with a potential reduction of feature dimension of 1280x. WeMine-P2P is extendable to other biosequence interactions such as predicting Protein-DNA interactions.

The evolution of m-learning has changed the ways of learning. With the increasing popularity of using mobile devices, m-learning has become important in teaching and learning in education. Most of m-learning services are provided through the Learning Management Systems (LMS). This paper aims at sharing an experience in students' usage of the mobile app in the LMS. In this paper, the study of using the LMS with mobile app is performed. The study includes the usage of LMS before and after the implementation of mobile app, the usage pattern of the web interface and mobile app, and the distribution of functions used by students for the mobile app. The value of this paper is to provide a concrete example for developers in considering the design and development of m-learning for the higher education institutes.

Understanding binding cores is of fundamental importance in deciphering Protein-DNA (TF-TFBS) binding and gene regulation. Limited by expensive experiments, it is promising to discover them with variations directly from sequence data. Although existing computational methods have produced satisfactory results, they are one-to-one mappings with no site-specific information on residue/nucleotide variations, where these variations in binding cores may impact binding specificity. This study presents a new representation for modeling binding cores by incorporating variations and an algorithm to discover them from only sequence data. Our algorithm takes protein and DNA sequences from TRANSFAC (a Protein-DNA Binding Database) as input; discovers from both sets of sequences conserved regions in Aligned Pattern Clusters (APCs); associates them as Protein-DNA Co-Occurring APCs; ranks the Protein-DNA Co-Occurring APCs according to their co-occurrence, and among the top ones, finds three-dimensional structures to support each binding core candidate. If successful, candidates are verified as binding cores. Otherwise, homology modeling is applied to their close matches in PDB to attain new chemically feasible binding cores. Our algorithm obtains binding cores with higher precision and much faster runtime ( $\geq$ 1,600x) than that of its contemporaries, discovering candidates that do not co-occur as one-to-one associated patterns in the raw data. Availability: http://www.pami.uwaterloo.ca/~ealee/files/tcbbPnDna2015/Release.zip .

Discovering protein patterns for amino acids and their biochemical properties is important for revealing the underlying biophysical models. From this, pattern clustering was introduced in order to relate the discovered protein patterns to taxonomic classes in a localized region of a protein. This paper proposes an algorithm to synthesize and re-group pattern clusters, maximizing their separability in order to reveal class characteristics of the localized region of the protein based on our previous work. To evaluate the pattern clustering and regrouping pattern clusters results, we introduce three evaluation measures: F-measure, class entropy measure, and attribute entropy measure. To validate our proposed algorithm, experiments are run on synthetic data, protein family for amino acid attributes, and chemical property attributes. The experimental results show that: a) the result for regrouping pattern clusters is more accurate in class separation than only using pattern clustering; b) The clusters after regrouping are more distinctly separable with each other than only using pattern clustering; c) two types of pattern clusters are found, with one pertaining to distinct classes and the other associating with two or more related classes; and d) class characteristics are clearly revealed in the data subspace containing the patterns in the pattern clusters. The datasets with chemical properties show that unsupervised techniques can reveal common chemical attributes in the inherent classes as more of the common properties shared by different amino acids are taken into account.

A basic task in protein analysis is to discover a set of sequence patterns that reflect the function of a protein family. This set of sequence patterns contains non-exact significant residue associations. Currently, the existing combinatorial methods are computationally expensive and probabilistic methods require richer representation of the amino acid associations. To undertake this task, we create a synthesized pattern representation called an Aligned Pattern (AP) Cluster that identifies the residue associations in the binding segment and the site variations in the aligned residues. In this paper, our algorithm identifies the binding segments for two protein families: the Cytochrome Complex and the Ubiquitin protein families. For each of the experiments, the AP Clusters obtained correspond to protein binding segments including a few beyond those identified by the other protein databases, PROSITE and pFam. Furthermore, the columns of aligned sites that exist only as a single value in the AP Clusters also corresponds to the binding residues. Additional information retained by the AP Clusters can reveal the amino acid residues of interest, thus averting time-consuming simulations and experimentation.

It's a new challenge to power system stability and control when wind farm cluster is connected to Power Grid at a substation. This paper presents work that has been conducted to research this impact on power system transient stability. The paper obtained results conducted in a realistic Power System with installed wind generation equal to 950MW. The results are generally positive that wind farm cluster should not create serious dynamic risks if the control strategy of DFIG is to support system during faults by generating reactive power. The system studies were performed using the simulation tools PSS/E. Modeling and simulation experiences for this tools achieved during the study are presented in the paper.

An essential step of protein function analysis is to discover patterns that represent functional regions in a set of protein family sequences. However, the same functional region of a protein family that occurs in different sequences may contain variations that resulted from biological substitutions, deletions, and insertions. Thus, a sequence pattern representing this functional region seldom repeats precisely at the exact position with the same amino acid residues. To capture these variable associations, we developed a pattern synthesis process. First, we used an effective sequence pattern discovery algorithm to discover high order patterns as input. Next, we group and align these similar discovered patterns into Aligned Random Pattern Clusters (ARPCs). During the clustering process, each ARPC is transformed into a probabilistic structural pattern called the Aligned Random Pattern Digraph (ARPD). The advantages of our synthesis process are 1) the synthesized patterns are not confined to a fixed protein region since the ARPCs captures the similar patterns by their variable sites, 2) the ARPDs retain both horizontal pattern associations and vertical site variations, and 3) the search space for synthesizing input patterns is smaller than that for aligning input sequences. Our method successfully discovers two functional protein regions of the Cytochrome Complex protein family: the proximal and distal binding segment that binds the iron molecule of the heme ligand from each side of the plane without relying on prior knowledge.

With the popularity of touchscreen devices, physical button-based keyboards are being replaced by finger-operated virtual keyboards. Entering text on these touchscreen devices is no longer limited by finger tapping activities on the keys. Text input can be performed by finger sliding over the virtual keyboard, which is well realized by the Swype technology on a typical QWERTY keyboard. Such shorthand gesturing for text input, however, may be inefficient when directly applied to Chinese input, especially in Chinese stroke-based input method. In this paper, a novel dynamic candidate keypad with use of unidirectional finger gesture on the stroke key for character selection is proposed for enhancing stroke-based Chinese input method. The new design can enhance frequently used Chinese characters searching and input using stroke-based input. The new method is implemented on the Android 2.2 platform for performance evaluation using Traditional Chinese characters set. Experimental results show that the proposed method enables users to input popular Chinese characters easier than conventional stroke-based input methods.

Discovering patterns from gene expression levels is regarded as a classification problem when tissue classes of the samples are given and solved as a discrete-data problem by discretizing the expression levels of each gene into intervals maximizing the interdependence between that gene and the class labels. However, when class information is unavailable, discovering gene expression patterns becomes difficult. This paper attempts to tackle this important problem. For a gene pool with large number of genes, we first cluster the genes into smaller groups. In each group, we use the representative gene, one with highest interdependence with others in the group, to drive the discretization of the gene expression levels of other genes. Treating intervals as discrete events, association patterns can be discovered. If the gene groups obtained are crisp clusters, significant patterns overlapping different clusters cannot be found. This paper presents a new method of “fuzzifying” the crisp attribute clusters for that purpose. To evaluate the effectiveness of our approach, we first apply the above described procedure on a synthetic dataset and then a gene expression dataset with known class labels. The class labels are not being used in both analyses but used later as the ground truth in a classificatory problem for assessing the algorithm's effectiveness in fuzzy gene clustering and discretization. The results show the efficacy of the proposed method.

Coastal surface currents measured remotely by high frequency (HF) radar are compared with those from in situ acoustic Doppler current profilers (ADCP) deployed in both Eulerian and Lagrangian configurations. Two 25 MHz CODAR SeaSondesreg, located at Cape May, NJ and Cape Henlopen, DE, continuously measure surface currents at the Delaware Bay mouth. The radars measure surface current hourly with a spatial resolution of approximately 1.5 km using ideal antenna patterns. Surface currents in this area are dominated by strong semidiurnal tidal exchange between the Delaware Bay and Atlantic Ocean, which can reach velocities of 190 cm/s. An upward looking 600 KHz ADCP with a vertical resolution of 0.5 m was deployed on a seafloor frame 1.5 km east of Cape Henlopen within the radar footprint for 38 days. A downward looking 600 KHz ADCP with a vertical resolution of 1 m was mounted on a research vessel and operated continuously along a transect within the radar footprint for 41 hours. Velocities from the uppermost complete bin of the seafloor mounted ADCP are compared with the nearest radar measurement. Velocities from the shallowest bin of the ship mounted ADCP are compared with the nearest radar measurement in both time and space. No interpolation was done. Regression analysis indicates correlation coefficients (R) between the radar and ADCPs of 0.8 to 0.9 for current magnitude and 0.7 to 0.8 for current direction. Complex correlation analysis yields correlation coefficients greater than 0.9 with the ADCP observations veering clockwise from the radar observations by 0.3deg to 0.6deg. The root mean square (RMS) difference of magnitudes indicates that radar measurements of surface current magnitude are 19.4% greater than those made by the moored ADCP and 8.2% less than those made by the ship ADCP.

Genome comparison is very useful for deriving evolutionary and functional relationships between genomes. Previous works on genome comparison focus mainly on comparing the entire genome at the nucleotide level. As interesting patterns exist also at the gene and segment level, we propose an algorithm called Multi- Level Genome Comparison Algorithm (MGC) that can allow genome comparison to be performed at multi-level while sequential and regional consistency of gene segments can be determined. Different genomes may have common sub-sequences that differ with each other due to processes such as mutations, lateral transfers, gene rearrangements that cannot be easily identified. The result is that not all the genes can form a certain one-to-one matching gene pair. One-to-many or many-to-many ambiguity relationships may exist . MGC takes this ambiguity into consideration and represents genomes with a new graph representation known as Multi-Level Attributed Graph Mining Algorithm (MAGMA). We tested MGC with the intra- and inter-species of Chlamydia genomes. The results show that the proposed algorithm is able to discover the similarities and dissimilarities among different genomes, while in addition, to confirm the specific role of the gene in the genomes and provide variations among species and similarity within species. KEY WORDS Genome comparison, multi-level, consistency, segment, graph

In this article the inverse kinematics problem for robot manipulators is considered under an inexact context; and a fast procedure for its solution and pseudoinverse robustness is presented. The approach is based on solving the linear system based on the symmetric matrix JJT+pl (where, J and I are the Jacobian and the identity matrix respectively, and p is a damping factor) by a Gaussian elimination process that takes into account the matrix symmetry, and automatically evaluating some simple parameters. These parameters are used in either one of two original schemes, which are theoretically justifiable, that are also proposed for the appropiate evaluation of the damping factor. In the first scheme an upper bound for the condition number of the matrix JJT+pl, in terms of some of the calculated parameters is used. Alternatively, in the second scheme the sufficiency condition for the rank preservation of the Jacobian developed in [5] is considered. First, an upper bound for the pseudoinverse matrix p, in terms of some of the evaluated parameters is found. Then, it is easily shown that the sufficiency condition for rank preservation can be established in terms of this bound and on the 00 norm of the Jacobian rate of change matrix. Furthermore, and as important, here it is also shown how to properly implement these schemes, in particular the one based on an upper bound for the pseudoinverse, in conjuction with a recently developed approach [4] for the singularities prevention of redundant manipulators. The developed algorithms are tested on the simulation of a planar redundant manipulator. From the results obtained it is observed that the proposed approach (in terms of efficiency, robustness and ability to prevent large joint speeds) compares favorably with the approaches using a Gaussian elimination procedure and with pseudoinverse robustness based on a manipulability measure. 2. The inverse Linematics problem.

This paper presents a new robust scene-recognition algorithm using points and lines in three uncalibrated color images under natural or artificial lighting without knowing the poses of the camera. Local optimization is employed by the algorithm in finding point matches to guarantee accuracy in the case when object surface is extremely uneven of having abrupt changes, thus causing a wide range of point movements between consecutive images. The algorithm is iterated to obtain maximum features until no improvement can be achieved. Experiments using a large number of images show excellent performance by this new algorithm.

Structured light systems (SLS) have been used in 3D model building for a long time. However, the comprehensive performance of such systems, in terms of data acquisition speed and depth density, is still far from being perfect. In this paper, we introduce a new pattern design for SLS in order to get a dense depth map while in principle maintaining the acquisition speed, high reliability and accuracy. Our system uses a hierarchical integration of features and textures in the projection pattern. This hierarchy is then used to guide the matching process. The effectiveness of the method has been demonstrated by extensive experiments.

A singularities avoidance approach suitable for the optimal path planning of redundant robot manipulators is presented. The approach is based on establishing proper bounds for the rate of change of the Jacobian matrix of the transformation between the joint speeds and the end effector Cartesian speed. These bounds become an additional constraint for an optimization problem that is formulated to obtain the optimal path of the robot manipulator. Here, the optimization problem is formulated globally as a state-constrained continuous optimal control problem which can consider joint (speeds) constraints and/or manipulator dynamics, and be solved by an efficient iterative numerical technique. This approach is particularly exemplified for the optimal path generation of a simulated planar redundant manipulator, and its results are compared with the results yielded by a local approach. The results obtained (although not adequate for present real-time implementation) confirm the superiority of the global approach. >

This paper presents an automatic learning technique. Its strength is demonstrated by a bench mark problem for evaluating learning systems-learning the past-tenses of English verbs. To deal with training patterns of variable length and to exploit probabilistic properties inherent in the English language, our method uses random graphs. Due to the large amount of redundancy between the base form and the past tense form of most verbs, only the endings of these forms are of interest in the classification process. These are extracted automatically. Taking advantage of the sequential nature of words, the complete graphs can be relaxed to sequences with great reduction in the computational requirements. A significant merit of this method over the connectionist models is that the knowledge acquired is expressed automatically in a rule-based form which simplifies the classification process of unknown verbs. >

This article presents a vision based online system for the robust trajectory planning of robot manipulators. It uses a 3D vision system to determine the relative position of the objects to be engaged and the obstacle to avoid, and a novel obstacle avoidance procedure for manipulator motion planning. From intensity images acquired by a CCD camera mounted on the robot arm, the salient features are first accurately and robustly detected and then grouped. Through the correspondences between the feature groupings and the model features, the 3D poses of the objects and the obstacles are determined and confirmed by back-projection. Once these poses are determined, an online procedure, based on redundancy resolution, is used to achieve obstacle avoidance. The approach utilizes a null space vector to set properly the robot configuration, and a potential field method to guide the end-effector. By pseudoinverse perturbation it also prevents singular configurations and local minima. The feasibility and effectiveness of the system is demonstrated by an experiment with online engagement and transportation of objects posed inside an aluminium frame.

Investigates the use of an attributed hypergraph (AH) as a generic and hierarchical representation for dynamic vision, and examines its mathematical perspective in the framework of category theory. As shown in Kass et al. (1988), attributed hypergraph representation (AHR) is able to render a direct, general and flexible representation for describing objects/models with geometrical and physical features. With this representation, analysis and synthesis of structural/visual patterns can be carried out at different resolution levels. This paper presents the application of AHR in motion analysis and automatic morphing.

The recovery of 2D information from intensity images, and that of 3D information from range images are the major issues in 3D objects recognition from sensory data. The analysis and interpretation of remote sensing aerial images have important applications. This paper presents an efficient method for the analysis and modeling of such scenes based on range sensory data. Unlike methods using 2D intensity images, we exploit the rich 3D data. We extract symbolic information from the 3D triangular mesh models. These are used to recognize buildings using symbolic reasoning and generic object models.

Rapidly changing requirements in manufacturing and robotics require efficient automated planning systems. In this paper, we present a method to acquire and exploit domain-knowledge. We use two examples of knowledge-extensive contexts; outdoor terrain robot navigation and mission planning. We represent the acquired sensory 3D data by triangular terrain meshes. Application independent features are automatically extracted from these and converted into symbolic entities suitable for reasoning. Their topological relations are then organized into attributed graphs. Higher-order, application dependent relations are captured by hyper-edges in attributed hypergraphs. The symbolic relations inducing hyperedges are used as the basis of symbolic reasoning operations. The resulting compact hypergraph representation of the raw data facilitates complex navigation and mission planning tasks. Domain-knowledge is thus captured in a flexible form and used to reduce the search for feasible paths.

This paper presents the automated model synthesis component of an integrated passive 3D vision system. The synthesized models can be used by the object recognition and pose determination components. The model synthesis obtains the 3D object model from images acquired by a CCD camera posed at various known positions. This paper presents developments and discusses automatic model synthesis. The tasks include robust 2D feature detection; 2D feature post-processing for eliminating noise and recovering missing features; 2D feature grouping of structurally related 2D features; stereo triangulation with a new form of epipolar line constraint; projective inversion of ellipses; synthesis for circular shape in 3D space from its projective views based on the ellipse pose hypothesis; and incremental model synthesis of model from multiple views based on the vertex triangulation. To demonstrate full automation, we use a single CCD camera mounted on the last link of a robot arm. The integrated robot system is able to move the CCD camera around the object and capture images at various vantage points and furnish the camera pose corresponding to each image acquired. The intelligent system then synthesizes the extracted features from each image to obtain a 3D model of the object. Such an approach, though more difficult than the direct use of range data through range sensors, is of great importance for space and industrial automation where cost and flexibility are of concern.

In distributed synthesis and control, one well-known potential hazard is conflict between modular designs. In a modular approach to the supervisory control of discrete-event systems, modular supervisors that are individually nonblocking may nevertheless conflict and thus produce blocking, when operating concurrently. A scheme of resolving this potential conflict between the modular supervisors would be to accord priorities to the conflicting supervisors. We formalize this scheme with reporter maps from a hierarchical approach to the supervisory control of discrete-event systems. Sufficient conditions on these reporter maps for conflict resolution are obtained. With these conditions, a constructive solution is developed.

A challenge in 3D computer vision is to automatically acquire 3D models of objects through a CCD camera and to use the acquired models to recognize objects and estimate their poses. The PAMI System works on images acquired from a single CCD camera. It first detects salient features from an image and then groups them according to their types as well as their spatial, geometrical and topological relations. The feature grouping types include: a) four corner points and triplets of lines forming corners; b) curve segments fitted into ellipses. The use of matching hypotheses generated based on feature groupings is usually more robust and effective than the combinatorial matching of point features.

To uncover qualitative and quantitative patterns in a data set is a challenging task for research in the area of machine learning and data analysis. Due to the complexity of real-world data, high-order (polythetic) patterns or event associations, in addition to first-order class-dependent relationships, have to be acquired. In this paper, we propose a novel method to discover qualitative and quantitative patterns (or event associations) inherent in a data set. It uses the adjusted residual analysis in statistics to test the significance of the occurrence of a pattern candidate against its expectation. To avoid exhaustive search of all possible combinations of primary events, techniques for eliminating impossible pattern candidates have been developed. Test results on artificial and real-world data are discussed towards the end of the paper.

In this article a simple and fast procedure for (on-line) synchronous motion planning of multi-manipulatory systems is presented. The method is based on solving for each manipulator a linear system of equations taking into account a simple null space vector for motion planning, and a novel scheme for the appropriate robustness of the pseudoinverse matrix. This method pursues simultaneously both motion: collision-free, coordination and singularities prevention in real time in a sensor based environment.

We develop a new technique for 3D object modeling by integrating the 2D vision methodology with a laser based high precision measurement technology. After calibrating the spatial relationship between the CCD camera and the range laser, the system captures an image of the inspected object, extracts 2D image features such as edges, curves and corners, and provides intelligent guidance for the range laser sensor to obtain 3D measurements on these features. The available range data are analysed and synthesized to construct a CAD model of the object.

The authors (1990) have shown that from the Jacobian rate of change matrix is it possible to derive a dexterity measure for robot manipulators. In this article this criterion is further extended for its application on the evaluation of several redundant wrist mechanisms for a commercial manipulator design. >

A simple scheme for the appropriate pseudoinverse perturbation, in the inverse kinematics computation, of robot manipulators is presented. The approach is based on properly implementing a bound for a sufficiency condition for the rank preservation of the Jacobian matrix. This bound is established in terms of the Jacobian matrix and on its rate of change matrix infinity norms. >

Two pattern recognition techniques capable of handling unordered discrete data are applied to the analysis and classification of clinical data. The first technique uses a dependence-tree approach for classifying and datecting patterns from the discrete data. The second technique is based on modulo-2 linear transformation and approximation of probability distributions. Both techniques are applied to clinical data of two categories of liver diseases: acute viral hepatitis and chronical active hepatitis. The data selected by a physician for identifying and discriminating these two liver diseases consists of 12 features, each feature having a range of two or three discrete values. Experimental results using the two techniques are presented.

A newly developed classification scheme for samples with discrete valued features is presented in this paper. In it, we first map the discrete feature space into a Euclidean space called logarithm of likelihood ratio (LLR) space. The likelihood ratios are formed from the estimated distributions based on the dependence tree structure obtained through minimizing the error probability. By discriminant analysis, we then transform the LLR space into one-dimensional space on which classification is conducted. We have applied this new scheme to several sets of biomedical data and have obtained significantly high classification rates.