A position-based approach of extracting elderly daily behavior habits for service robot

For the purpose of learning human daily behavior habit from sensor data to detect the unwonted behavior, we propose an approach by using the position information in home environment. Firstly, we build a layered framework which is suitable for service robot and can be applied in home environment to monitor the behavior states of elderly human. Then, we apply the improved basic sequential algorithmic scheme and k-means clustering to extract the disciplinarian of human daily living. The emulational experiment results demonstrate that our approach is fit for the problem we address.