John Q. Gan

Classifying mental tasks based on features of higher-order statistics from EEG signals in brain-computer interface

In order to characterize the non-Gaussian information contained within the EEG signals, a new feature extraction method based on bispectrum is proposed and applied to the classification of right and left motor imagery for developing EEG-based brain-computer interface systems. The experimental results on the Graz BCI data set have shown that based on the proposed features, a LDA classifier, SVM classifier and NN classifier outperform the winner of the BCI 2003 competition on the same data set in terms of either the mutual information, the competition criterion, or misclassification rate.

Constructing L2-SVM-Based Fuzzy Classifiers in High-Dimensional Space With Automatic Model Selection and Fuzzy Rule Ranking

In this paper, a new scheme for constructing parsimonious fuzzy classifiers is proposed based on the L2-support vector machine (L2-SVM) technique with model selection and feature ranking performed simultaneously in an integrated manner, in which fuzzy rules are optimally generated from data by L2-SVM learning. In order to identify the most influential fuzzy rules induced from the SVM learning, two novel indexes for fuzzy rule ranking are proposed and named as alpha-values and omega-values of fuzzy rules in this paper. The alpha-values are defined as the Lagrangian multipliers of the L2-SVM and adopted to evaluate the output contribution of fuzzy rules, while the omega-values are developed by considering both the rule base structure and the output contribution of fuzzy rules. As a prototype-based classifier, the L2-SVM-based fuzzy classifier evades the curse of dimensionality in high-dimensional space in the sense that the number of support vectors, which equals the number of induced fuzzy rules, is not related to the dimensionality. Experimental results on high-dimensional benchmark problems have shown that by using the proposed scheme the most influential fuzzy rules can be effectively induced and selected, and at the same time feature ranking results can also be obtained to construct parsimonious fuzzy classifiers with better generalization performance than the well-known algorithms in literature.

A self-paced brain-computer interface for controlling a robot simulator: an online event labelling paradigm and an extended Kalman filter based algorithm for online training

Due to the non-stationarity of EEG signals, online training and adaptation are essential to EEG based brain–computer interface (BCI) systems. Self-paced BCIs offer more natural human–machine interaction than synchronous BCIs, but it is a great challenge to train and adapt a self-paced BCI online because the user’s control intention and timing are usually unknown. This paper proposes a novel motor imagery based self-paced BCI paradigm for controlling a simulated robot in a specifically designed environment which is able to provide user’s control intention and timing during online experiments, so that online training and adaptation of the motor imagery based self-paced BCI can be effectively investigated. We demonstrate the usefulness of the proposed paradigm with an extended Kalman filter based method to adapt the BCI classifier parameters, with experimental results of online self-paced BCI training with four subjects.

EMG-based hands-free wheelchair control with EOG attention shift detection

This paper presents a novel hands-free control system for an electric-powered wheelchair, which is based on EMG (Electromyography) signals recorded from eyebrow muscle activity. By using a simple CyberLink device, one-dimensional continuous EMG signals are obtained, analysed, and then translated into multi-directional control commands (forward, left, right, etc.) for the wheelchair that supports multi-directional control. At the same time, EOG (Electrooculography) signals detected from eye movements are used to adjust wheelchair speed. The system also allows a user to choose either control state or non-control state so that any non-intended muscle activity can be ignored during the non-control state. The system is reliable, easy to set up, and easy to use.

A complete analytical solution to the inverse kinematics of the Pioneer 2 robotic arm

For robotic manipulators that are redundant or with high degrees of freedom (dof), an analytical solution to the inverse kinematics is very difficult or impossible. Pioneer 2 robotic arm (P2Arm) is a recently developed and widely used 5-dof manipulator. There is no effective solution to its inverse kinematics to date. This paper presents a first complete analytical solution to the inverse kinematics of the P2Arm, which makes it possible to control the arm to any reachable position in an unstructured environment. The strategies developed in this paper could also be useful for solving the inverse kinematics problem of other types of robotic arms.

Inverse kinematics learning for robotic arms with fewer degrees of freedom by modular neural network systems

Artificial neural networks have been traditionally employed to learn and compute the inverse kinematics of a robotic arm. However, the inverse kinematics model of a typical robotic arm with joint limits is a multi-valued and discontinuous function. Because it is difficult for a multilayer neural network to approximate this type of function, an accurate inverse kinematics model cannot be obtained by using a single neural network. In order to overcome the difficulties of inverse kinematics learning, we propose a novel modular neural network system that consists of a number of expert modules, where each expert approximates a continuous part of the inverse kinematics function. The proposed system selects one appropriate expert whose output minimizes the expected position/orientation error of the end-effector of the arm. The system can learn a precise inverse kinematics model of a robotic arm with equal or more degrees of freedom than that of its end-effector. However, there are robotic arms with fewer degrees of freedom, where the system cannot learn their precise inverse kinematics model. We have adopted a modified Gauss-Newton method for finding the least-squares solution to address this issue. Through the modifications presented in this paper, the improved modular neural network system can obtain a precise inverse kinematics model of a general robotic arm.

Extracting Takagi-Sugeno Fuzzy Rules with Interpretable Submodels via Regularization of Linguistic Modifiers

In this paper, a method for constructing Takagi-Sugeno (TS) fuzzy system from data is proposed with the objective of preserving TS submodel comprehensibility, in which linguistic modifiers are suggested to characterize the fuzzy sets. A good property held by the proposed linguistic modifiers is that they can broaden the cores of fuzzy sets while contracting the overlaps of adjoining membership functions (MFs) during identification of fuzzy systems from data. As a result, the TS submodels identified tend to dominate the system behaviors by automatically matching the global model (GM) in corresponding subareas, which leads to good TS model interpretability while producing distinguishable input space partitioning. However, the GM accuracy and model interpretability are two conflicting modeling objectives, improving interpretability of fuzzy models generally degrades the GM performance of fuzzy models, and vice versa. Hence, one challenging problem is how to construct a TS fuzzy model with not only good global performance but also good submodel interpretability. In order to achieve a good tradeoff between GM performance and submodel interpretability, a regularization learning algorithm is presented in which the GM objective function is combined with a local model objective function defined in terms of an extended index of fuzziness of identified MFs. Moreover, a parsimonious rule base is obtained by adopting a QR decomposition method to select the important fuzzy rules and reduce the redundant ones. Experimental studies have shown that the TS models identified by the suggested method possess good submodel interpretability and satisfactory GM performance with parsimonious rule bases.

Multiresolution analysis over simple graphs for brain computer interfaces

OBJECTIVE Multiresolution analysis (MRA) offers a useful framework for signal analysis in the temporal and spectral domains, although commonly employed MRA methods may not be the best approach for brain computer interface (BCI) applications. This study aims to develop a new MRA system for extracting tempo-spatial-spectral features for BCI applications based on wavelet lifting over graphs. APPROACH This paper proposes a new graph-based transform for wavelet lifting and a tailored simple graph representation for electroencephalography (EEG) data, which results in an MRA system where temporal, spectral and spatial characteristics are used to extract motor imagery features from EEG data. The transformed data is processed within a simple experimental framework to test the classification performance of the new method. MAIN RESULTS The proposed method can significantly improve the classification results obtained by various wavelet families using the same methodology. Preliminary results using common spatial patterns as feature extraction method show that we can achieve comparable classification accuracy to more sophisticated methodologies. From the analysis of the results we can obtain insights into the pattern development in the EEG data, which provide useful information for feature basis selection and thus for improving classification performance. SIGNIFICANCE Applying wavelet lifting over graphs is a new approach for handling BCI data. The inherent flexibility of the lifting scheme could lead to new approaches based on the hereby proposed method for further classification performance improvement.

Asynchronous BCI control of a robot simulator with supervised online training

Due to the non-stationarity of EEG signals, online training and adaptation is essential to EEG based brain-computer interface (BCI) systems. Asynchronous BCI offers more natural human-machine interaction, but it is a great challenge to train and adapt an asynchronous BCI online because the users control intention and timing are usually unknown. This paper proposes a novel motor imagery based asynchronous BCI for controlling a simulated robot in a specifically designed environment which is able to provide users control intention and timing during online experiments, so that online training and adaptation of motor imagery based asynchronous BCI can be effectively investigated. This paper also proposes an online training method, attempting to automate the process of finding the optimal parameter values of the BCI system to deal with non-stationary EEG signals. Experimental results have shown that the proposed method for online training of asynchronous BCI significantly improves the performance.

Mercer Kernel, fuzzy C-means algorithm, and prototypes of clusters

In this paper, an unsupervised Mercer kernel based fuzzy c-means (MKFCM) clustering algorithm is proposed, in which the implicit assumptions about the shapes of clusters in the FCM algorithm is removed so that the new algorithm possesses strong adaptability to cluster structures within data samples. A new method for calculating the prototypes of clusters in input space is also proposed, which is essential for data clustering applications. Experimental results have demonstrated the promising performance of the MKFCM algorithm in different scenarios.