Estimation of Basic Activities of Daily Living Using ZigBee 3D Accelerometer Sensor Network

Abstract

Recently, how to support and care elderly people is serious problem. Modern approach is trying to support to keep health (body and mental) seniors living alone, but it is hard to keep a health by oneself. Our aim is monitoring a user's life for providing suitable services to keep his/her health. To providing suitable service, life log is very useful information. However, traditional monitoring system is only focusing elementary motion, walking, sleeping, and so on. We deal with recognizing advanced motions by using transportation data from an area to other areas in a house. These advanced motion logs support not only elderly but also care stuff or family who plans supports. In daily life, our motion is very variety. Therefore we focused on Basic Activities of Daily Living (ADL) which categories normally doing in daily life such as feeding, bathing, homemaking, and so on. In this research, we pick up 11 motion from ADL for recognition.

Integrating monitor functions of physiological data into the robotics service grid

“…In the previous works, we got the experiences of the data exchange and the Bluetooth communication. After Shimokawara [9] built the ZigBee sensor network environment to estimate the user's activity, we realize that we need to follow a standard mechanism for combining and exchanging sensor data in our robot service environment. RT-middleware [10] can solve this problem.…”

Context-Aware Adaptive Framework for e-Health Monitoring

“…Movement or fall detection systems are ambulatory activities detection including dynamic activities (like walking) and static postures (like standing), location tracking and accidental falls [9]. Activity of Daily Living (ADL) monitoring is covering a list of basic daily activities (such as eating and washing) and allows the elderly to live independently and to provide accurate services for caregivers [7] [10]. Existing approaches lack a true understanding about the person's activities that should be considered in monitoring.…”

“…However, such schemes are not automatically adjusted at the execution time of monitoring, therefore we need a dynamic choice of monitoring modes. Some studies have resorted to specific types of wireless communication technology such as Zigbee [10] or Bluetooth Low Energy (BLE) [12] to reduce the energy consumption. This remains restrictive, since dealing with all available sensors and devices for full visibility of a person's context is imperative.…”

A predictive approach for efficient e-Health monitoring

“…Support Vector Machines (SVMs) were used in [9]. Clustering approaches and fuzzy membership functions are used to define fuzzy rules of data collected for activity prediction [10]. A data driven system was presented in [4] by using temporal and spatial contextual activity data.…”