Bin An         Email: ban@mail.ustc.edu.cn

Yan Ning       Email: ningyan@mail.ustc.edu.cn
Qiang Chen     Email: shawn@mail.ustc.edu.cn
Zhaohui Jiang  Email: jiangzhh@ustc.edu.cn
Huanqing Feng  Email: hqfeng@ustc.edu.cn

Department of Electronic Science and Technology,
University of Science and Technology of China,
Hefei, Anhui, China

Description:

Since the most distinctive motion imagination related character exhibits on the frequency 0~20Hz with the spectral estimation, we use the program below:
1. Common Spatial Patterns(CSP). After testing, we choose the most distinctive features exhibiting in the first and second  of the fourteen CSP components.
2. AR spectral estimation with the order 10 at 10-12Hz as character.
3. SVM as the classifer (We use the libsvm2.5 toolbox),the kernel of SVM is radial basis function(RBF).