Research Themes|Smart Home

Sensing in the Smart Home

Recently, implementation and utilization of smart homes, which enable the occupants to live more affluent and comfortable lives by introducing the technology of network composed of various sensors to recognize activities of occupants, has been spreading step by step against a backdrop of rising awareness of the environmental issues and energy-saving. Since energy-saving is especially a pressing issue, various efforts like the demand response function that is realized by collaboration of smart grid and HEMS (Home Energy Management System) are made. On the other hand, not only energy-saving, but health and communication among family members are also important factors to live an affluent and fulfilling life. Therefore smart homes should achieve both functions of energy management and daily-life support.

Our research team have designed and developed SLSA (Smart Life Support Agent) system, which helps to improve the quality of human life in smart homes. By using various sensors, this system detects daily-life activities of occupants such as eating meal or sleeping in the bedroom, generates advices automatically which enable occupants to improve the quality of daily life from various aspects like electricity costs or health, and suggests the advices to the occupants. For example, if an occupant uses home appliances that uses electricity (e.g. cleaning rooms with a vacuum cleaner) when the electricity rate is high, it advises him that he can save some money by do it later, or if an occupant had insufficient exercise, appropriate amount of exercise and calorie consumption are advised. To recognize activities of occupants in smart homes, it is appropriate to use infrared human detection sensors which have less concern of privacy violation and low implementation cost, but it has some difficulties of calibration: adjusting each residence, and mapping locations of each sensor; complex deployment is required. These difficulties of installation will be stagnation to wide implementation of the system especially for aged households, therefore the system have to be as cheap and easy to deploy as possible.

In this research, I propose the in-home activity recognition system, which uses portable motion sensors and power monitoring sensors that are easy to deploy, because activity recognition is required for SLSA to generate advices. The system automatically estimates whether the motion sensor mainly detects stay or passage of occupants from the data over few days in the target household only by deploying motion sensors and power monitoring sensors, which do not correspond to certain locations or appliances. It also estimates the type of the location (living room, kitchen, bedroom, etc.) where the occupants stay by estimating which appliance corresponds to the power consumption data and associating appliance usage to location of occupants stay. Herewith, I aim for realizing the system which can comprehend the in-home activity of occupants correctly without registering the locations and the appliances correspondence with sensors.

Smart Life Support Adviser

We propose a system that helps to improve the quality of human life insmart homes. We take a model-based approach where smart home residents,appliances, energy sources and their correlation are comprehensivelymodeled. Using the proposed model and activity recognition based onMarkov logic networks, our system generates smart life tips, which arekind advice to increase smart life metrics (such as health index,quality of time spent at home, energy cost saving and air comfort index)in a non-intrusive way.

Copyright(c) 2015 Higashino Lab. All Rights Reserved. Design by http://f-tpl.com