Ailing Luo

Research on fault identification for complex system based on generalized linear canonical correlation analysis

Complex system exists extensively in modern process manufacturing industry. One major problem of its fault diagnosis is how to extract the inner partial relationship, with which we can model the fault performance and then identify the faults. In this paper, based on Generalized Linear Model (GLM), an improved CCA algorithm (GLCCA) is proposed to extract both the linear and nonlinear relationship in the complex system. A pneumatic experiment table as a complex system with some fault simulation is obtained from the state key laboratory. The data is composed of the pressure signature of six reducing valves and the signature of another four unit state. Simulated and experimental results show that this method is adequate enough to extract the inner relationship in the complex system.

A motion rehabilitation self-training and evaluation system using Kinect

Stroke patients usually have difficulties to conduct rehabilitation training themselves, due to no rehabilitation evaluation in time and dependence on doctors. In order to solve this problem, this paper proposes a motion rehabilitation and evaluation system based on the Kinect gesture measuring technology combining VR technology as well as traditional method of stroke rehabilitation. Real-time rehabilitation motion feedback is achieved by using Kinect motion capturing, customized skeleton modeling, and virtual characters constructed in Unity3D. The jitter problem of virtual characters following motion using Kinect is solved. Fidelity and interactivity of virtual rehabilitation training is improved. Our experiment validated the feasibility of this system preliminarily.

EEG signal co-channel interference suppression based on image dimensionality reduction and permutation entropy

It is well known that electroencephalogram (EEG) signals collected from scalps are highly contaminated by various types of artifacts and background noise. The perturbations induced by artifacts and random noise are particularly difficult to correct because of their high amplitude, wide spectral distribution, and variable topographical distribution. Therefore, de-noising of EEG is a very challenging pre-processing step prior to qualitative or quantitative EEG signal analysis. To address this issue, some de-noising approaches have been proposed for noise suppression. However, most of these methods are only available for multi-electrode EEG signal processing, besides, the co-channel interference are always left unprocessed. Aiming at the obstacles encountered by the conventional approaches in single electrode EEG signal co-channel interference suppression, a method based on time-frequency image dimensionality reduction is proposed in this paper. The innovative idea of the proposed method is that it is applicable for single electrode EEG signal enhancement and the background noise can be suppressed in entire time-frequency space. The proposed method is experimentally validated by a group of real EEG data. The experimental results indicate that the proposed method is effective in EEG single electrode co-channel interference suppression.

The Role of Visual Noise in Influencing Mental Load and Fatigue in a Steady-State Motion Visual Evoked Potential-Based Brain-Computer Interface

As a spatial selective attention-based brain-computer interface (BCI) paradigm, steady-state visual evoked potential (SSVEP) BCI has the advantages of high information transfer rate, high tolerance to artifacts, and robust performance across users. However, its benefits come at the cost of mental load and fatigue occurring in the concentration on the visual stimuli. Noise, as a ubiquitous random perturbation with the power of randomness, may be exploited by the human visual system to enhance higher-level brain functions. In this study, a novel steady-state motion visual evoked potential (SSMVEP, i.e., one kind of SSVEP)-based BCI paradigm with spatiotemporal visual noise was used to investigate the influence of noise on the compensation of mental load and fatigue deterioration during prolonged attention tasks. Changes in α, θ, θ + α powers, θ/α ratio, and electroencephalography (EEG) properties of amplitude, signal-to-noise ratio (SNR), and online accuracy, were used to evaluate mental load and fatigue. We showed that presenting a moderate visual noise to participants could reliably alleviate the mental load and fatigue during online operation of visual BCI that places demands on the attentional processes. This demonstrated that noise could provide a superior solution to the implementation of visual attention controlling-based BCI applications.

Steady-State Motion Visual Evoked Potential (SSMVEP) Based on Equal Luminance Colored Enhancement

Steady-state visual evoked potential (SSVEP) is one of the typical stimulation paradigms of brain-computer interface (BCI). It has become a research approach to improve the performance of human-computer interaction, because of its advantages including multiple objectives, less recording electrodes for electroencephalogram (EEG) signals, and strong anti-interference capacity. Traditional SSVEP using light flicker stimulation may cause visual fatigue with a consequent reduction of recognition accuracy. To avoid the negative impacts on the brain response caused by prolonged strong visual stimulation for SSVEP, steady-state motion visual evoked potential (SSMVEP) stimulation method was used in this study by an equal-luminance colored ring-shaped checkerboard paradigm. The movement patterns of the checkerboard included contraction and expansion, which produced less discomfort to subjects. Feature recognition algorithms based on power spectrum density (PSD) peak was used to identify the peak frequency on PSD in response to visual stimuli. Results demonstrated that the equal-luminance red-green stimulating paradigm within the low frequency spectrum (lower than 15 Hz) produced higher power of SSMVEP and recognition accuracy than black-white stimulating paradigm. PSD-based SSMVEP recognition accuracy was 88.15±6.56%. There was no statistical difference between canonical correlation analysis (CCA) (86.57±5.37%) and PSD on recognition accuracy. This study demonstrated that equal-luminance colored ring-shaped checkerboard visual stimulation evoked SSMVEP with better SNR on low frequency spectrum of power density and improved the interactive performance of BCI.

An optimum preload method for machine tool spindle ball bearings

To provide a suitable rotation rate for different machining processes, a single machine tool spindle should work over a wide range of speeds. This study considered the effects of speed on dynamic behaviour of ball bearings and combined the fatigue life model and ball bearing internal load distribution model to determine the appropriate preload. First, the influence of speed on internal load distribution and ball bearing contact angles was analysed. The preload was calculated using a ball bearing internal load distribution model. Next, assuming constant bearing fatigue life, the theoretical preload curves were determined using the fatigue life model by changing the reliability factor. Finally, at low speeds, the maximum designed preload (design value) was set as the initial preload. With the increase in speed, the optimum preload was hierarchically obtained within the internal region between the theoretical preload curves. An experimental test rig for the optimum preload of ball bearings, which can automatically adjust the preload, was developed. The proposed method for determining the optimum preload was verified using the measured performance indicators, including the temperature, motor currents, and vibration of ball bearings. The results showed that the optimum preload suggestion made the test ball bearings exhibit excellent behaviour.

Study on Information Integration of Condition Monitoring and Fault Diagnosis System in Manufacturing

Aimed at the problems of equipment condition information share and integration in the enterprise monitoring and diagnosis systems, the XML-based model of equipment condition information was presented by analysing the structure of equipment condition information. Afterwards, with the help of the model, the idea of intranet-based software “bus” for machine monitoring and fault diagnosis system was proposed, which is a criterion that specifies the data presentation, management and communication protocol. According to the difference of work mode, the information provider is divided into pusher and puller, the information applicant passively or actively receives the information correspondingly. Then, the technique has been applied to the information integration between the portable condition monitoring system and the online condition monitoring system, Results show that it have enough flexibility and expansibility, and can be applied to large-scale information integration of equipment condition information.

Recognition of SSMVEP signals based on multi-channel integrated GT 2 circ statistic method

Brain-computer interface (BCI) is a modern useful tool of bypassing usual channels of muscle and peripheral nervous system to establish a direct connection between brain and external devices and to restore fundamental communication and control skills. Steady-state visual evoked potential (SSVEP), as one of the most popular EEG modality, has been widely used in BCI applications. For SSVEP BCI, the most challenging task is to effectively improve the accuracy, especially in minimum number of recording electrodes and short stimulation duration. In this study, a novel multi-channel integrated GT2circ statistic method was proposed for the frequency recognition in a four-class steady-state motion visual evoked potential (SSMVEP)-based BCI. The proposed method was compared with the widely used canonical correlation analysis (CCA) and verified with three-channel EEG data from three healthy subjects. Results indicated that a higher recognition performance with shorter recording time and few electrodes can be achieved by using of this novel method rather than CCA method, making multi-channel integrated GT2circ statistic a robust approach for the implementation of SSVEP BCIs.

Human action recognition based on kinematic similarity in real time

Human action recognition using 3D pose data has gained a growing interest in the field of computer robotic interfaces and pattern recognition since the availability of hardware to capture human pose. In this paper, we propose a fast, simple, and powerful method of human action recognition based on human kinematic similarity. The key to this method is that the action descriptor consists of joints position, angular velocity and angular acceleration, which can meet the different individual sizes and eliminate the complex normalization. The angular parameters of joints within a short sliding time window (approximately 5 frames) around the current frame are used to express each pose frame of human action sequence. Moreover, three modified KNN (k-nearest-neighbors algorithm) classifiers are employed in our method: one for achieving the confidence of every frame in the training step, one for estimating the frame label of each descriptor, and one for classifying actions. Additional estimating of the frame’s time label makes it possible to address single input frames. This approach can be used on difficult, unsegmented sequences. The proposed method is efficient and can be run in real time. The research shows that many public datasets are irregularly segmented, and a simple method is provided to regularize the datasets. The approach is tested on some challenging datasets such as MSR-Action3D, MSRDailyActivity3D, and UTD-MHAD. The results indicate our method achieves a higher accuracy.

The Evaluation of the CNC Machine's Dynamic Performance Based on Rough-Set Theory

In order to make a fast and valid evaluation of CNC machines dynamic performance, this paper puts forward a new method to evaluate the dynamic property based on Rough-Set theory. Take the example of translation axis, acquire the dynamic data of round test through secondary development of numerical control system, describe the circular error map by dimensionless parameter and using Rough-Set theory to discrete processing and attribute reduction the data and features. The experiment result shows that the method proposed in this paper can evaluate CNC machines dynamic performance accurately and efficiently. This method also provides new solutions for dynamic performance evaluation of machine tools.