As known as IoT (Internet of Things), commodity network and sensor are going to be installed in whole cities and indoors. Reality of pervasive computing environment is depends on the quality of sensor networking and big data technologies. We are analyzing sensor big data to know what's happen in a real world and designing scalable data processing systems.
Environmental sensing
Setting up sensor data to cities and indoors then analyze sensor data to know the circumstances. Though, raw sensor data are directly valuable for the first purpose, secondary use of the sensor data will enhance their values. Furthermore, sophisticated sensor fusion technologies are expected.
We have researched human tracking system and climate sensing system. Then, we are going to sense whole the cities by using bus transportation infrastructure which consists of bus stops and buses themselves as mobile sensors.
- Wi-Fi Probe request collection based tourist tracking system
- Live E! (Distributed climate sensors and the infrastructure)
Participatory sensing
Distributing sensors whole the cities is not sometime cost effective. The other solution is participatory sensing. There are tons of smart phones over the world. Once, we found out the method to utilize gathered sensor data from the smart phones, the developing cost of sensor network is dramatically decrease. However, due to the wide variety of specification of sensors installed in smart phones, gathered data are noisy. It's important to innovate correction technology to make the sensor data from smart phones serviceable.
We have developed a cooperative brightness map of street at night, by collecting, correcting, and sublimating illuminometer data from smart phones.
Life logging
Sensor big data is useful for not only for public but also private. For example, your smart phone can always log acceleration data then you can know the walking step and burning calories.
As increasing the variety of MEMS (Micro Electro Mechanical Systems) sensors on the private wearable devices, the value of wearing such devices is depend on the quality of sensor data analyzing technologies.
We have a plan to utilize heart beat information to detect an unknown disease or understand a mental state.
Publications
- Toyokazu Akiyama, Yukiko Kawai, Tsuneo Jozen, Tatsuro Miura, Junji Nishida, Ryuichi Yoshida, Kaori Ota, Ismail Arai, “Research and development on layered cloud computing for collection and analysis of vehicle sensor data,” IEICE Tech. Rep., Vol.117, No.299, IA2017-40, p.47, Thailand, November, 2017.