Abstract:
Most of the countries in the world are facing the problems of aging population and disabilities among the population. Among di erent problems faced by these individuals, self feeding can be identi ed as an important aspect that should get more attention from the research community. In addition, self feeding re ects the interdependency of an individual and thus relate to their mental health. Taking care of these individuals using care takers is becoming more and more di cult due to diminishing workforce for such tasks. Therefore assistive robotic technologies play a major role in providing feeding solutions to these individuals with disabilities. Meal assistance robot is a device designed to assist the individuals in need with self feeding. The research work of this thesis is focused on developing an EEG signal based Brain Machine Interface for a meal assistance robot. Meal assistance robot is capable of handling solid food items using the spoon mounted on the end e ector. Identifying user's food selection is carried out using a Steady State Visually Evoked Potential based Brain Machine Interface where 3 LED matrices icking at 6Hz, 7Hz and 8Hz are used to generate the stimulations in the brain. User has to gaze at a LED panel to activate the motion path of the robot which will feed the solid food from the container associated with the gazed LED panel. System is incorporated with a visual servoing algorithm to identify the user's mouth position and adapt the food feeding location according the mouth location. Further, Mouth open/close status detection system is developed to measure the user's willingness to intake the food. The developed meal assistance robot is experimentally validated using 15 subjects in di erent experiments. After detailing the research methods carried out, discussion of the results obtain are presented at the end of the thesis with limitations of the research and possible future improvements.