Real-time affective states identification for human-robot interaction /
Under the context of using robot to perform rehabilitation on stroke patient, the robot must be able to understand the patient so to be able to interact with the patient more effectively. Current practise to rehabilitate stroke patient depends on the resource available that mostly involve trained hu...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
Kuala Lumpur:
Kulliyyah of Engineering, International Islamic University Malaysia,
2013
|
| Subjects: | |
| Online Access: | Click here to view 1st 24 pages of the thesis. Members can view fulltext at the specified PCs in the library. |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Under the context of using robot to perform rehabilitation on stroke patient, the robot must be able to understand the patient so to be able to interact with the patient more effectively. Current practise to rehabilitate stroke patient depends on the resource available that mostly involve trained human therapists. Progresses basically are monitored continuously in qualitative manner and the therapy session needs to be done in regular basis repetitively. However, the practice is costly and does not provide a quantitative way to measure the progress of the affected person. A robot on the other hand can work precisely and continuously and able to record the progresses of a patient quantitatively. The therapy using a robotic system can be made more efficient when the physiological state of the affected muscle of the patient complemented with the affective state (psychological state) of the patient is known. For this research, the focus of the affective state is the engagement level of the patient when subjected to rehabilitation procedure of his upper limb (i.e. moving his arm to follow specific trajectory). For evaluating engagement level, the lectrooculogram (EOG) signal is captured when the patient is doing the therapy. The signals are fed into fuzzy classifier to deduce the engagement level of the patient. In developing the fuzzy classifier, the related data is required to deduce the engagement level. A series of experiments are designed where the patients are asked to track a set of prescribed paths on the computer screen which has different level of difficulties within the allocated times and have to obey different speed constraints. The position error from the trajectory tracking is measured together with the electrooculogram (EOG) signal which is recorded by using a G-tec data acquisition system simultaneously. The information on the endogenous type of eye blinking is extracted from the electrooculogram (EOG) and it plays an important role to study the engagement level. Following the experiment, a series of questionnaires that has been carefully designed are given to the subjects to verify the engagement level deduced from the experiment done earlier by the subjects. A robotic platform is then used to verify the engagement level in real-time. The engagement model in the form of fuzzy classifier is used to adapt the speed of the robotic platform which is useful for the human-robot interaction. In particular, if the level of engagement is high, the subject is subjected to more challenging trajectory to be tracked. This is useful especially for the robot assisted type of stroke rehabilitation. The analysis on the questionnaire and the deduction of the level of engagement from the experimental results shows an accuracy of 95%. The robotic system is also able to adapt its speed whenever the level of engagement level changes. The research is only limited to one physiological signal namely the electrooculogram (EOG). Besides that, this research only consider onto one affective state which is the engagement. |
|---|---|
| Physical Description: | xv, 172 leaves : ill. ; 30cm. |
| Bibliography: | Includes bibliographical references (leaves 106-109). |