Research on Intention/Situation Understanding and Communication Control based on Stochastic Model
| Project/Area Number |
17500128
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Perception information processing/Intelligent robotics
|
|---|
| Research Institution | Osaka Institute of Technology |
|---|
Principal Investigator |
SANO Mutsuo Osaka Institute of Technology, Faculty of Information Science and Technology, Professor (30351464)
|
|---|
| Project Period (FY) |
2005 – 2007
|
| Project Status |
Completed (Fiscal Year 2007)
|
| Budget Amount *help |
¥3,050,000 (Direct Cost: ¥2,900,000、Indirect Cost: ¥150,000)
Fiscal Year 2007: ¥650,000 (Direct Cost: ¥500,000、Indirect Cost: ¥150,000)
Fiscal Year 2006: ¥1,000,000 (Direct Cost: ¥1,000,000)
Fiscal Year 2005: ¥1,400,000 (Direct Cost: ¥1,400,000)
|
|---|
| Keywords | communication control / stochastic model / intention understanding / situation recognition / communication robot / basian netwoork / interaction control / mutual entrainment / 確率推論 |
|---|
| Research Abstract |
The aim of this research is to clarify how communication robots should understand human's intention and situation based on various kinds of uncertain information, and find a technical solution to communicate naturally without stress. So we launched the studies of (i) intention and situation understanding based on stochastic model, (ii) interaction control based on hierarchical stochastic state transition model, (iii) communication control with mutual entrainments, and (iv) sensing method for intention and situation understanding. Consequently, the following four results were obtained.
1. We proposed the method to understand human's intention and situation based on the characteristics of time series curve of estimated probability acquired by dynamic basian network, and the effectiveness was verified by communication experiments.
2. It was clarified that the interaction model based on hierarchical stochastic state transition made interaction module created efficiently through a communication robot produced by ourselves.
3. The algorism of mutual entrainment which made communication smooth and natural using the timing control of nodding and throwing in words was applied to robot communication, and so the effectiveness was demonstrated
4. We developed the software of human location estimation, eye gaze recognition, speaker direction detecting, and gesture recognition which were robust against environment change and noises for improving the accuracy of intention and situation.
The remained study is to clarify mutual entrainment communication model among more than two persons and a communication model without boring.
|
Report
(4 results)
Research Products
(23 results)
