The support vector data description (SVDD) is a method proposed to solve the problem of one-class classification. It models a hypersphere around the target set, and by the introduction of kernel functions, more flexible descriptions are obtained. In SVDD, the width parameter s and the penalty parameter c have to be given beforehand by the user. To automatically optimize the values for these parameters, the error on both the target and outlier data has to be estimated. Because no outlier examples are available, we propose a max-min range method for generating artificial outliers in this paper. By generating artificial outliers around the target set, the accuracy of classifiers will improve. At the last, we use four benchmark data sets: Iris, Wine, Balance-scale, and Ionosphere data base to validate the approach in this research indeed has better classification result.

In this article, a Cartesian coordinate nano-scale positioning platform driven by stack-type piezoelectric actuator associated with flexure hinge mechanism is developed. Structural design and analysis, system identification, and feedforward controller design to deal with the hysteresis effect, as well as and the feedback PID controller designed by use of Ziegler-Nichols method and the genetic algorithm are included. The developed Cartesian platform is able to achieve the target within 10 nm in positioning at the stroke of 19 mum

In this article, hysteresis controller design and static error compensation method for a 6-dof micro-positioning platform driven by piezoelectric actuator is studied. The nonlinear hysteresis effect of the piezoelectric actuator is analyzed by means of Preisach model. Its inverse model is used as the feedforward controller. Error compensation method is designed to compensate the manufacture error and assembly error by use of the developed 3-points-3-axes measurement method. From practical experiment, the proposed method makes improvement on the accuracy of positioning.

This paper describes the implementation of a recovery mechanism in a hot standby fault-tolerant computer system. The recovery mechanism is designed based on Win32 API and in association with the error detection and diagnosis mechanism (EDDM), the heartbeat and the backup database. The recovery mechanism consists of three primary components, the Interceptor, the Log Manager and the Recovery Manager. The Interceptor is in charge of capturing the state information of the target process. The Log Manager will take over the data transmitted from the Interceptor and backup the data in the redundant database. While application program (AP) is failed, the Recovery Manager obtains related state information from the Log Manager and resumes the processes to the normal operating status. Winpro 2000 is used to establish the platform of the recovery mechanism.

A virtual reality (VR) surgical training platform for unicompartmental knee replacement is developed in his study. The platform is constructed with VR software employed on PC-based system and connected with a low-cost manual manipulator. The primary components in the VR include soft tissue and skeletal tissue of the knee as well as various surgical tools such asscapels, cutting jig, drill, and oscillating saw, and so on. The manipulator provides 6-DOF motion and 3-DOF force measurements. During manipulation, all measurement data is displayed via a monitoring interface on the screen. Besides, an evaluation system with different assessment method for each surgical process is built for training purpose.

As a robot loses some of its capabilities due to component failure, fault-tolerant controllers isolate the faulty unit and complete the task with minimum performance loss. However, as the robot undergoes a failure and accumulates large errors, controller torque requirements usually increase and may even exceed actuator capacities. This problem can be avoided in redundant serial manipulators if the redundancy of the system is exploited to redistribute the actuator torques. This is accomplished by implementing the controller in the end-effector space. The computed torque method and PID feedback control are used in the controller. When torque demand on an actuator exceeds the allowable limit, that actuator is isolated and the required load is distributed among the remaining actuators. This controller not only avoids the saturation problem, but also allows the robot to eliminate errors and track the original trajectory. Application of the method is demonstrated on a four degree-of-freedom serial manipulator.>