Chin-Shyurng Fahn

Development of a data glove with reducing sensors based on magnetic induction

In this paper, we present the development of a data glove system using magnetic induction coils as finger movement sensors. This data glove has the capability of measuring ten degrees of freedom of a hand with only five sensors that are arranged properly on the palmar surface instead of the dorsal surface. Because these sensors are installed on the finger phalange positions, there is no contact point between the sensors and the finger joints. Hence, the shape of the sensors does not change as the fingers are bending, and the quality of measurement and the lifetime of the sensors will not decrease with time. Additionally, the motion constraints of the finger joints are investigated to simplify the development of the data glove. For the convenience of using this glove, a simple and efficient calibration process consisting of two steps is also provided, so that all required parameters can be determined automatically. The theoretical formulation of the bending angles of proximal interphalangeal, metacarpophalangeal, and thumb interphalangeal joints as well as the calibration equations are derived directly from the positions and orientations of the generator and sensor coils on the data glove. Also described and realized are the circuit block diagram and the construction of the data glove system. To prevent the interference among the generator coils, we adopt the time division method to scan the generator and sensor coils. So far, the experimental results of the sensors performing linear movement and bending angle measurements directly using an oscilloscope in less noisy environments are very satisfactory. It reveals that our data glove is available for a man-machine interface.

The multisynapse neural network and its application to fuzzy clustering

In this paper, a new neural architecture, the multisynapse neural network, is developed for constrained optimization problems, whose objective functions may include high-order, logarithmic, and sinusoidal forms, etc., unlike the traditional Hopfield networks which can only handle quadratic form optimization. Meanwhile, based on the application of this new architecture, a fuzzy bidirectional associative clustering network (FBACN), which is composed of two layers of recurrent networks, is proposed for fuzzy-partition clustering according to the objective-functional method. It is well known that fuzzy c-means is a milestone algorithm in the area of fuzzy c-partition clustering. All of the following objective-functional-based fuzzy c-partition algorithms incorporate the formulas of fuzzy c-means as the prime mover in their algorithms. However, when an application of fuzzy c-partition has sophisticated constraints, the necessity of analytical solutions in a single iteration step becomes a fatal issue of the existing algorithms. The largest advantage of FBACN is that it does not need analytical solutions. For the problems on which some prior information is known, we bring a combination of part crisp and part fuzzy clustering in the third optimization problem.

Development of a Fingertip Glove Equipped with Magnetic Tracking Sensors

In this paper, we present the development of a data glove system based on fingertip tracking techniques. To track the fingertip position and orientation, a sensor module and two generator coils are attached on the fingertip and metacarpal of the corresponding finger. By tracking the fingertip, object manipulation tasks in a virtual environment or teleoperation system can be carried out more precisely, because fingertips are the foremost areas that reach the surface of an object in most of grasping processes. To calculate the bending angles of a finger, we also propose a method of constructing the shape of the finger. Since the coils are installed on the fingertips and metacarpals, there is no contact point between the sensors and finger joints. Hence, the shape of the sensors does not change as the fingers are bending, and both the quality of measurement and the lifetime of the sensors will not decrease in time. For the convenience of using this glove, a simple and efficient calibration process consisting of only one calibration gesture is also provided, so that all required parameters can be determined automatically. So far, the experimental results of the sensors performing linear movement and bending angle measurements are very satisfactory. It reveals that our data glove is available for a man-machine interface.

A high-definition human face tracking system using the fusion of omni-directional and PTZ cameras mounted on a mobile robot

The omni-directional cameras providing 360 degrees field of view (FOV) are widely used in video surveillance and robot vision applications. However, the omni-directional cameras have an obvious drawback; that is, only low-resolution images captured. Therefore, the objects are not able to be correctly identified if they are far from the omni-directional cameras. To overcome this problem, we propose a high-definition human face tracking system using the fusion of omni-directional and pan-tilt-zoom (PTZ) cameras. Our system first detects and tracks human faces in the panoramic images received from an omni-directional camera, and then controls the PTZ camera to fixate at a target face for capturing a high-resolution image. At the beginning, the human face detection procedure obtains moving human faces by means of temporal differencing together with skin and hair color filtering. The detected human faces are subsequently fed to the face tracking procedure which employs a particle filter for iteratively tracking human faces. Once a target human face is selected, the PTZ camera is directed to stare at the target and zoom in it speedily. Then the face tracking procedure turns to use the images received from the PTZ camera for continuously tracking the target face until it is outside the FOV of the camera.

Multiple disturbances classifier for electric signals using adaptive structuring neural networks

This work proposes a novel classifier to recognize multiple disturbances for electric signals of power systems. The proposed classifier consists of a series of pipeline-based processing components, including amplitude estimator, transient disturbance detector, transient impulsive detector, wavelet transform and a brand-new neural network for recognizing multiple disturbances in a power quality (PQ) event. Most of the previously proposed methods usually treated a PQ event as a single disturbance at a time. In practice, however, a PQ event often consists of various types of disturbances at the same time. Therefore, the performances of those methods might be limited in real power systems. This work considers the PQ event as a combination of several disturbances, including steady-state and transient disturbances, which is more analogous to the real status of a power system. Six types of commonly encountered power quality disturbances are considered for training and testing the proposed classifier. The proposed classifier has been tested on electric signals that contain single disturbance or several disturbances at a time. Experimental results indicate that the proposed PQ disturbance classification algorithm can achieve a high accuracy of more than 97% in various complex testing cases.

Hierarchical Artificial Neural Networks for Recognizing High Similar Large Data Sets

This paper proposes a hierarchical artificial neural network for recognizing high similar large data sets. It is usually required to classify large data sets with high similar characteristics in many applications. Analyzing and identifying those data is a laborious task when the methods adopted are primarily based on visual inspection. In many field applications, data sets are measured and recorded continuously using automatic monitoring equipments. Therefore, a large amount of data can be collected, and manual inspection has become an unsuitable approach to recognizing those data. This proposed hierarchical neural network integrates self-organizing feature map (SOM) networks and learning vector quantization (LVQ) networks. The SOM networks provide an approximate method for computing the input vectors in an unsupervised manner. Then the computation of the SOM may be viewed as the first stage of the proposed hierarchical network. The second stage is provided by the LVQ networks based on a supervised learning technique that uses class information to improve the quality of the classifier from the first stage. The multistage hierarchical network attempts to factorize the overall input vector into a number of small groups, each of which requires very little computation. Consequently, by use of the proposed network, the loss in accuracy can be small.

A face robot for autonomous simplified musical notation reading and singing

This research is aimed to devise an anthropomorphic robotic head with a human-like face and a sheet of artificial skin that can read a randomly composed simplified musical notation and sing the corresponding content of the song once. The face robot is composed of an artificial facial skin that can express a number of facial expressions via motions driven by internal servo motors. Two cameras, each of them installed inside each eyeball of the face, provide vision capability for reading simplified musical notations. Computer vision techniques are subsequently used to interpret simplified musical notations and lyrics of their corresponding songs. Voice synthesis techniques are implemented to enable the face robot to sing songs by enunciating synthesized sounds. Mouth patterns of the face robot will be automatically changed to match the emotions corresponding to the lyrics of the songs. The experiments show that the face robot can successfully read and then accurately sing a song which is assigned discriminately.

Polygonal mesh simplification with face color and boundary edge preservation using quadric error metric

In applications such as scientific and medical visualization, highly detailed polygonal meshes are needed. Rendering these polygonal meshes usually exceeds the capabilities of graphics hardware. To improve rendering efficiency and maintain proper interactivity, the polygonal mesh simplification technique is commonly used to reduce the number of polygons of the mesh and construct a multiresolution representation. We propose a new and simple constraint scheme based on the quadric error metric proposed by Garland and Heckbert (1997, 1998) to preserve face colors and boundary edges during the simplification process. In addition, our method generates progressive meshes that store the polygonal mesh in a continuous multi-resolution representation. According to our experimental results, this new method is successful in preserving face colors and boundary edges. Moreover, we compare the latency of resolution changes for progressive meshes of various models.

Versatile Humanoid Robots for Theatrical Performances

The purpose of this research is to develop multi-talented humanoid robots, based on technologies featuring high-computing and control abilities, to perform onstage. It has been a worldwide trend in the last decade to apply robot technologies in theatrical performance. The more robot performers resemble human beings, the easier it becomes for the emotions of audiences to be bonded with robotic performances. Although all kinds of robots can be theatrical performers based on programs, humanoid robots are more advantageous for playing a wider range of characters because of their resemblance to human beings. Thus, developing theatrical humanoid robots is becoming very important in the field of the robot theatre. However, theatrical humanoid robots need to possess the same versatile abilities as their human counterparts, instead of merely posing or performing motion demonstrations onstage, otherwise audiences will easily become bored. The four theatrical robots developed for this research have successfully performed in a public performance and participated in five programs. All of them were approved by most audiences.

A distributed approach to fuzzy clustering by genetic algorithms

Fuzzy clustering (c-means) is a widely known unsupervised clustering algorithm, but it can not guarantee to find the global minimum, because it approximates the minimum of an objective function by the iterative method in solving the differentiation problem, starting from a given point. For overcoming this drawback, we incorporate the genetic search strategies in the fuzzy clustering algorithm to explore the data space from a multiple-point concept. The direct application of the genetic algorithms to the fuzzy clustering is not suitable, because sometimes the data set is enormous. Under this situation, the chromosome would be too long, so a distributed approach to fuzzy clustering by genetic algorithms is proposed to divide the huge search space into many small ones. The simulation results show our algorithm works fine.