| Project/Area Number |
12480098
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (B)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
情報システム学(含情報図書館学)
|
|---|
| Research Institution | NARA INSTITUTE OF SCIENCE AND TECHNOLOGY |
|---|
Principal Investigator |
NISHITANI Hirokazu NARA INSTITUTE OF SCIENCE AND TECHNOLOGY, Graduate School of Information Science, Professor, 情報科学研究科, 教授 (10029572)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
KUROOKA Taketoshi NARA INSTITUTE OF SCIENCE AND TECHNOLOGY, Graduate School of Information Science, Research Associate, 情報科学研究科, 助手 (90273846)
YAMASHITA Yuh NARA INSTITUTE OF SCIENCE AND TECHNOLOGY, Graduate School of Information Science, Associate Professor, 情報科学研究科, 助教授 (90210426)
|
|---|
| Project Period (FY) |
2000 – 2001
|
| Project Status |
Completed (Fiscal Year 2001)
|
| Budget Amount *help |
¥7,600,000 (Direct Cost: ¥7,600,000)
Fiscal Year 2001: ¥2,100,000 (Direct Cost: ¥2,100,000)
Fiscal Year 2000: ¥5,500,000 (Direct Cost: ¥5,500,000)
|
|---|
| Keywords | Thinking state estimatioin / EEG data / Physiological indices / Artificial neural network model / Plant operations / Abnormal situation / On-line monitoring / Adaptive aiding / 思考状態 / 心理状態モニタリング / 混乱状態の検知 / プラント運転 / 数学問題解答 / 支援システム |
|---|
| Research Abstract |
We have been studying the use of multiple channel elecroencephalogram (EEG) data to infer a human's thinking state. As a result, we have confirmed off-line thinking state estimation to be effective, in experimental studies on simulator training during malfunctions and mathematical problem solving. In this research, we reconsidered the mathematical model for thinking state estimation and concluded that an artificial neural network (ANN) model is more appropriate for real-time use than the linear regression model.
We developed a prototype real-time thinking state monitoring system with the ANN model and evaluated its classification accuracy via mathematical problem solving. The average rates of correct classification and misclassification were 68 % and 9 %, respectively. We used this model for reaHime monitoring and confirmed that the estimation result by the ANN model was found to coinckie with the subject's thinking state mode judged from the operator's behavior.
In order to develop an easier way of measurement, we examined the relationships between the thinking state and six physiological indices. As a result, we found that electrocardiogram (EGG) data, etectrooculogram (EOG) data, and respiratory curve (RSP) correlate to the thinking stata However, the best index W for each thinking state mode is different. We proposed a new estimation model with these three indices, which provided stably 60-70 % classification accuracy for day-to-day variations.
With emphasis on the practical use of thinking state monitoring, we applied a simple wireless electroencephalograph to monitor the operator's thinking state in a simulator training for emergency situations in a production plant. We expect that the monitoring result can be used to change the training menu in reference to the trainee's engagement level estimated from the thinking state information.
|
