Q. Liu’s research while affiliated with Wuhan University of Technology and other places

These years have been witnessing an increasing emphasis on researches of brain computer interface (BCI) which becomes a novel communication method from the brain to the output device, independent on normal peripheral nerve and muscle. And electroencephalogram (EEG) signal processing is one of the key research topics. In this paper, wavelet packet transform and common spatial patterns (CSP) are utilized for feature extraction. Finally, support vector machine (SVM) and Mahalanobis-distance are chosen to classify two kinds of motor imagery signal of left and right hands. Through experiments, we can recognize various factors affecting classification accuracy and the maximum accuracy rate could be up to 90.00%.

Weight calculation of customer requirements (CRs) is critical in quality function deployment (QFD). In order to solve the problems such as vagueness of the customer demand information, the fuzzy theory has been applied to the weight calculation. The assessment information about CRs has been represented by using trapezoidal fuzzy number, the lacking of information problem in information aggregation has been solved by using two-tuple linguistic theory, and a convenient method was given to transform different granularity linguistic information. The experimental results proved the feasibility of the proposed method.

The EEG acquisition system is put forward based on the character of electrical signal of human scalp. The electric circuit that consists of biological signal preamplifier circuit, filter circuit, 50 Hz frequency trap circuit can effectively filter the signal noise and amplify the EEG signals. Simultaneously, the system in which the technology of high-speed sampling and serial communication based on MSP430 microcontroller and PC software developed on LabVIEW are adopted could complete the work of digital EEG signal acquisition and real-time waveform display at low cost. The result of system circuit simulation and experiment of data acquisition test indicates that the system put forward could process human bioelectricity signal effectively in real time, and it could be a support for BCI research.

Delays are the major factor in the performance analysis of the control network. Topologies most in use and data characteristics of the network control system are considered, and corresponding delay models under different scheduling policies are derived based on stochastic network calculus. Then the distribution of delays and stochastic upper bound on delays of real-time data in the control network are calculated. The results are compared with the guaranteed upper bound on delays derived by deterministic network calculus, and the pessimism of the latter bound is evaluated.

The machining accuracy of computer numerical control (CNC) machine tools largely depends on the interpolation algorithms implemented. In this paper, a real-time high-precision interpolation algorithm for a general-typed parametric curve is proposed. As with other time-partition interpolation algorithms, the linear motion trajectory (employed to approximate the real trajectory) in a sampling cycle is determined from the current position of the tool to another point (target point) on the target trajectory. To determine the target point in a sampling cycle, a neighbourhood interval containing the point is first determined, the target point is then iteratively approached by means of the bi-section or secant method for equation resolutions. Simulation results show that the order of magnitude of the maximum feedrate relative errors (the relative deviation of the computed feedrate from the desired one) is −8 or less, while the order of magnitude of the maximum chord errors in the unit of meter is −6 or less. The performance indices can be further improved with more computation iteration in each sampling cycle. The algorithm is general, and applicable to any smooth curve that can be formulated in parametric equations.

A service-oriented configurable and re-configurable network numerical control system is proposed in the paper. In the system, each machine tool is only necessarily equipped with a few necessitated embedded-based numerical control (NC) field execution units, which only provide part of the overall NC functionality. The field execution units from several machine tools are then connected to an NC server which simultaneously controls the operations of those machine tools. The system is component-based, in which the NC server software is composed of components of pre-defined interfaces. The composition is based upon configuration settings and can be re-configured.

This paper proposes a new watermarking scheme for three-dimensional (3D) triangular mesh models. In this new scheme, feature points from the rapid changing regions are used to uniquely partition a 3D mesh model into Voronoi patches. Each Voronoi patch is then projected to a reference square plane and the corresponded range image is generated. By employing the existing two-dimensional (2D) image watermarking algorithms, watermark information is embedded into the transform domain of each range image. The watermarked 3D mesh model is obtained by perturbing the vertexes of each patch according to its watermarked range image. In order to make the watermarking scheme robust against various forms of attacks while preserving the perceptibility of the model, the length of the watermark information, the number of the feature points and the embedding strength of the watermark are made adaptive with respect to the geometry complexity of the 3D mesh models. Experimental results show that the proposed watermarking scheme has good imperceptibility and performs well under common 3D watermarking attacks such as mesh simplification, adding noise and cropping.

A DC-DC chopper is employed in the exhaust gas recycling system to stable the output voltage of the system. As the process of DC-DC chopping is non-linear with considerable disturbance, it is difficult to achieve the best performance by using traditional control approaches such as PID etc. Based on the analysis of the Buck-Boost chopper in continuous conduction mode, sliding mode control strategy is utilized. Simulation demonstrates that the sliding mode control method can achieve excellent results with shorter rising time on start stage and small system overshoot on sliding mode, which ensures a better dynamical performance of the system.

The manager of manufacturing tasks and the executive of these tasks (including needed manufacturing resources) usually belong to different enterprises and don't have common authority in this kind of distributed manufacturing environment. The central methods of task allocation cannot be adapted to this kind of manufacturing environment. Hence, a distributed task allocation based on combinatorial auction and combined with multi-agent technology is proposed in this paper. And a greedy algorithm is applied to solve the problem of bidder selection in the process of combinatorial auction.

In the current rapidly changing manufacturing environment, most manufacturing companies have realised that the ability to quickly develop a customised product at competitive prices is very crucial for them to survive in the keen competitive global market. Product modelling has been recognised as one of the key factors in determining the success of various product development strategies and for industrial competitiveness now and in the future. Standard for Exchange of Product model data (STEP) is an international standard designed to provide a complete, unambiguous, computer-readable definition of the physical and functional characteristic of a product through its lifecycle. This paper proposes a STEP-compliant product data model to fully support the complete product information representation and management for injection moulding product development. The data model is developed based on the structure of an injection moulding company. The entire data model consists of six data models which are defined to express information from various departments of the company. The data model is described in EXPRESS language, which can be integrated with other STEP-based product data models and databases. Consequently, the proposed system can be further extended for helping other injection moulding manufacturers to improve their product development process. An information sharing prototype system is implemented based on the proposed product data model to demonstrate the feasibility of the approach.