Embedding graceful deception system in visual navigation for elderly health care

Abstract

In the past sensors and surveillance systems were installed in the home environment to capture user activities. However, blind spots exist in areas out of reach of sensors. Obstacles and furniture are some barriers that cause fall down of elderly in the home environment. A motion sensor is mounted on Arduino platform to design a visual navigation system in service robot. Colors and depth of images are used to predict the walking route of elderly and to navigate the service robot. The service robot is functioned to use strategy inferred from fuzzy output to make the elderly detour from any possible collision with barriers. A graceful deception model and fuzzy inference system are proposed in this paper. In case an incident is predicted to happen in the current walking route, the proposed system will calculate the dangerous degree to avoid colliding with obstacles. Experiment results are given to verify the effectiveness of the presented work.