Japanese vendor OKI has unveiled a highly sensitive human detecting sensor technology that can distinguish large movements, like a walking person, to minute actions like breathing.
OKI said that its technology can detect even the minute movements of otherwise motionless persons, making it suitable for use in various applications, including advance warnings of health problems. OKI is currently seeking to apply this technology to areas ranging from security to the monitoring of elderly or people requiring long-term care.
Conventional sensors rely on microwaves to detect movements, but instead of just capturing the targeted human body movement, they end up capturing movements of other objects.
The firm developed a statistical model that can extract human movement with high precision while disregarding the effects of objects moving in the surroundings.
In partnership with the Chuo University Statistical Data Analysis, OKI analysed test data that simulated home environments, and came up with a technology that is able to distinguish in real-time between resting and moving states of a human body.
"Future work will refine this technology to incorporate functions capable of detecting breathing and heartbeats or anomalous behaviour," said Takeshi Kamijoh, general manager of Research and Development Center at OKI. "The associated data can be aggregated and analysed over a cloud network to ensure early detection of problems, such as falling or health-related irregularities, and rapid response."
The microwave sensors on which this technology is based do not require exposed sensor units, allowing the configuration of discreet, non-intrusive systems to reduce discomfort in privacy-sensitive locations.
The sensors are capable of detecting motion through furniture or bedding and are unaffected by ambient temperature fluctuations or heat sources such as heaters, allowing use even in hot and humid locations like bathrooms. The technology is expected to be ideal for monitoring patients in homes or in hospitals.
Sign up for Computerworld eNewsletters.