腦 工 程 研 究

 
計 畫 內容
建構以安全駕駛為觀點之車輛與人類生物反應模型系統利用虛擬實境動態模擬系統模擬駕駛者與真實交通環境狀況之互動下,駕駛者在接收到外界訊號時所承受的心理負擔 (mental work load, MWL)。在分析車輛的控制演算法上,以multiple fuzzy regression作為分析理論,而在駕駛者精神壓力方面,將著重於人類腦電波(EEG)以及心血管系統相關生理資訊(如ECG)。

空間迷向問題之探討及模擬器之控制飛行錯覺之空間迷向現象是造成飛機失事的重大因素之一,本研究以模擬器來訓練飛行駕駛得知空間迷向的感受,並記錄生理訊號以探究飛行員於空間迷向狀態下之生理現象。

仿生物電腦視覺科技之研究仿生物電腦視覺科技將包含四個主要的視覺處理功能:目標辨識(object recognition)、運動偵測(motion detection)、色彩恆常(color constancy)、及立體視覺(stereo vision)。以人類視覺系統之操作原理及其神經生物學模型作為仿生物電腦視覺科技之研發基礎。研究成果可應用於駕駛輔助系統,以解決車道/道路偵測及障礙物偵測等問題。
Topic
To develop a system that performs real-time EEG signal analysis in order to generate control commands for environmental control, communication, or even simple driving commands.

Objectives
Although there are many brain-computer interface (BCI) research groups studying how to provide a more immersed and intimate interaction between human and computer, and developing the so-call "bionic" applications in the world, current BCI has only the maximum information transfer rates of 5-25 b/min. Achievement of greater speed and accuracy depends on improvements in signal process, translation algorithms, and user training. Our intentions here are made to provide the disabled users with a more speedy and accurate BCI.

Backgrounds
BCIs give their users communication and control channel that do not depend on the brain's normal output channels of peripheral nerves and muscles. Current interest in BCI development comes mainly from the hope that this technology could be a valuable new augmentative communication option for those with severe motor disabilities that prevent them from using conventional augmentative technologies, all of which require some voluntary muscle control. Over the past five years, the volume and pace of BCI research have grown rapidly. In 1955 there were no more than six active BCI research groups, but there are more than 20 in 2000. They are focusing on brain electrical activity, recorded from the scalp as electroencephalographic activity (EEG) or from within the brain as single-unit activity, as the basis for this new communication and control technology. Table 1 lists several research groups working on communication channels between the brain and the computer recently.
 Table 1. Brain-computer interfaces’ research groups in the world.

University

Researchers

Year

EEG-Signal

Feed-back

Country

University of Michigan Biomedical Engineering Department

Huggins et al.

1999

Oscillatory Freq. Comp.

No

USA

University Rochester Department of Computer Science

Bayliss and Ballard

1999

P300

No

USA

University of Technology Graz Institute of Biomedical Engineering

Ramoser et al. Guger et al.

1999, 2000/01

Oscillatory Freq. Comp.

No     Yes

Austria

University of Tübingen Institute of Medical Psychology and Behavioral Neurobiology

Kotchoubey  Kübler Birbaumer

1997

1999

Slow wave

Yes

Germany

University degli Studi Tor Vergata

Babiloni et al.

1999

Oscillatory Frequ. Comp.

No

Italy

Wadsworth Center Wadsworth Center for Laboratories and Research

Wolpaw et al.

1998

Oscillatory Frequ. Comp.

Yes

USA

When the real-time EEG signals are picked up from the scalp by the array electrodes, the activity electrical signals related to the response of brain can be obtained. EEG signal consists of voltage changes of tens of microvolts at frequencies ranging from below 1Hz to about 50 Hz. It can be analyzed and quantified in the time domain, as voltage versus time, or in the frequency domain, as voltage or power versus frequency. In the time domain, the form or magnitude of the voltage change evoked by a stereotyped stimulus, referred to as an evoked potential or evoked response, can serve as a command. For example, the evoked potential produced by the flash of a certain letter can indicate whether the user wants to select that letter. In the frequency domain, the amplitude of the EEG in a particular frequency band, referred to as a rhythm, can function as a command. For example, that amplitude can be used to control movement of a cursor on a computer screen. Figure 1 summarizes the protocol of BCI.

        Figure 1. Flow diagram of major processing steps in the BCI strategy
Research Highlights
This research is to develop a system that performs real-time EEG signal analysis in order to generate control commands for environmental control or communication. To achieve this goal, the following tasks will be performed.
1. Study on the definition and essential features of a brain-computer interface
BCI input consists of particular features of brain activity. In time-domain, these features include slow cortical potentials, P300 potentials, and the action potentials of single cortical neurons. In frequency-domain, these features include EEG m or b rhythms occurring in specific areas of cortex. It is needed for more improving the evoked EEG signal recognition. The above typical features of EEG will provide the information for designing the algorithm.
2. Study on matching the BCI and its input to the user
Matching the user with his or her optimal BCI input features is essential if BCI is ever to be broadly applied to the communication needs of users with different disabilities.
3. Study on the EEG signal analysis
The goal of signal analysis in a BCI system is to maximize the signal-to-noise ratio of the EEG or signal-unit features that carry the user's messages and commands.
Autoregression (AR) model parameter estimation is a useful method for describing EEG activity, and can prove valuable for BCI application. When additive outlier contamination is present, a generalized robust maximum likelihood estimate can be valuable. This method is based on a modified Kalman filter. The wavelet transform is combined the time-frequency analysis in a nonlinear scale.
4. Development of a new BCI translation algorithm
A translation algorithm is a series of computations that transforms the BCI input features derived by the signal processing stage into actual device control commands. Different BCIs use different translation algorithms. An artificial neural network or an recursive self-organization fuzzy neural network can be considered as a nonlinear transfer function between the input features and output commands.
5. BCI technology application in locomotion
A big signal like the P300 can be recognized via single trial recognition in order to trigger commands on the virtual reality (VR) car. On the contrary, we shall also build the mental workload model of a driver based on the VR-car driving simulator (see figures below) by correlating the relationship between driver's EEG signals and his/her reactions in different driving sceneries.