Ge Yunjian

The structure optimization of radial basis probabilistic neural networks based on genetic algorithms

In this paper, a genetic algorithm (GA) is introduced into optimizing the radial basis probabilistic neural networks (RBPNN). The encoding method proposed in this paper involves not only the number and the locations of selected hidden centers but also the shape parameter of the Gaussian kernel function. We use the telling-two-spirals-apart problem as an example to validate the genetic algorithm for optimizing the RBPNN. Consequently, we obtain an optimal interval of the shape parameter of the kernel function for this problem except the reduced RBPNN structure (including the optimal number of the hidden centers and their optimal locations). The experimental results show that with the shape parameters in the optimal interval and with the optimized hidden centers the designed network is not only parsimonious but also of better generalization performance.

Research on Nano-SiO2/Carbon Black Composite for Flexible Tactile Sensing

In this paper, aiming at pressure sensitive conductive rubber materials based on flexible tactile sensing, studying the influence of adding nano-SiO2 materials on the electrical and mechanical performances. Analyses are made on the phenomenon produced by filling nano-SiO2 with different contents through experiments, researching the micro and macro mechanism of nano-SiO2 modification theoretically. The results showed that the combination of suitable nano-SiO2 and carbon black can effectively increase the conductivity of composites, and improve piezoresistive properties and stability of composites. It can provide pressure sensitive conductive rubber with high conductivity and fine mechanical properties for flexible tactile sensors.

On-line signature verification based on template matching approach and support vector data description

Automatic on-line signature verification is an intriguing intellectual challenge with wide attention and practical applications. To solve the difficulty of how to choose the verification threshold in signature verification, we present a novel method based on template matching approach and support vector data description (SVDD) in this paper. Investigates the details of the method and examines the validation of features extracted from 3-axis force of pen-tip to writing tablet. The effectiveness of the proposed solution is demonstrated by the experiment..

Application of genetic algorithms to the structure optimization of radial basis probabilistic neural networks

The genetic algorithm (GA) is applied in this paper to select hidden centers of radial basis probabilistic neural networks (RBPNN). The encoding method of individuals for GA, proposed in this paper, embodies not only the number but also the positions of selected centers. In addition, precision control is integrated into definition of the fitness function. Finally, we use the two-dimensional Gaussian distribution classification problem to illustrate the performance of the GA.

A Study of Resistance Relaxation Phenomenon Based on Carbon Black/Silicone Rubber System Flexible Tactile Sensor

The paper discusses resistance relaxation phenomenon of carbon black/silicone rubber system under constant temperature and the mechanism of relaxation phenomenon and microstructure analysis. Through the analysis of piezoresistive properties, resistance relaxation properties under static constant pressure are determined in room temperature and resistance relaxation properties under constant pressure and resistance relaxation properties under different constant pressures are obtained. The influences of different test conditions on resistance relaxation time and relaxation amplitude are analyzed as well as the changes of conductive network structure during resistance relaxation course.

Analysis of Methane Outburst Features and Real-time Alarm

This research proposed a real-time alarm method for methane outbursts. The method is established on the realtime analysis of the increasing velocity and the acceleration of the methane concentration curve. A real-time alarm algorithm based on fuzzy mathematics is presented. A prototype instrument and experiment results are presented and analyzed, which show good performance of this proposed method. This method can be used to give triggering signal to the explosion suppressing system at very early stage of outburst

Boolean Symbolic Expression Method for the Material Layer of Information Acquisition

This research proposed a symbolic expression method which classifies sensory mechanisms by Boolean symbolism. This classification belongs to the universal model research of the material layer of the discipline of information acquisition science and technology. Based on basic physics theories, sensory mechanisms were classified using a binary tree structure. Every type mechanism has been expressed with Boolean symbol. The symbolic expression method presented in the paper also describes the performances deference of sensors, including accuracy and dynamic performance. One actual application was analyzed to demonstrate the relation concluded from the Boolean classification presented in this paper. The results showed that this symbolic expression method could be used to forecast the performances of new sensor on a certain extend, especially in inertial sensor field.

Study of calibration and dynamic characteristic of SAFMS-T

This paper introduces the theory of six axes force measuring table (SAFMS-T). It will be proved that redundant technology can improve its characteristic, and then a new method of calibration and decoupling should be produced which helps to minimize the high requirement for the rigidity of the material. This also decreases cross-sensitivity errors. Finally the dynamic characteristic and compensation of it should be discussed.

A new adaptive controller-observer design for manipulator

An adaptive velocity observer, and an adaptive combined controller-observer design which is simple and easily implementable, are proposed for a manipulator. The mechanical parameters of the manipulator are uncertain and there are only joint angle measurements. The locally asymptotic convergence of the observation errors and tracking errors of the manipulators joints is proved. The simulation results demonstrate the effectiveness of the algorithm.