The world’s inhabitants is rising older, and that could be a development that’s anticipated to proceed for the foreseeable future. The proportion of the inhabitants that’s over age 65 is predicted to almost double by 2050, whereas the variety of people over age 80 is predicted to triple in the identical time interval. Except we collectively, particularly within the Western world, determine to begin having bigger households, we’re going to face many new challenges within the years forward. With many extra older people, and fewer younger folks to assist handle them, at this time’s healthcare issues will probably be drastically amplified.
One of many largest points confronting older people is the danger of falls. A fall can result in damaged bones, head accidents, and even dying. The implications are so extreme that folks at excessive threat for falls usually should be positioned in around-the-clock assistive care. This not solely locations a heavy load on the healthcare system, but additionally robs the people of their independence. Clearly this isn’t the perfect resolution, and because the inhabitants ages, it could not even be possible to proceed down this path.
An outline of the system structure (đź“·: H. Solar et al.)
A pair of researchers at Binghamton College consider that expertise often is the resolution to this drawback. They’ve developed a fall detection system that may be deployed in-home to quickly alert medical professionals and relations when a fall happens. This isn’t a brand new thought. Technological advances have enabled the event of quite a few fall detection units lately. Nevertheless, present programs all have one flaw or one other that limits their practicality in real-world conditions.
Usually, these units both use an accelerometer, which has restricted accuracy in detecting falls, or cameras, which pose an usually unacceptable violation of 1’s privateness, particularly when positioned within the dwelling. To beat these issues, the staff developed a system known as Speedy Response Aged Security Monitoring (RESAM). RESAM leverages each an accelerometer and a Kinect sensor to gather details about the individual underneath commentary. The Kinect sensor produces a simplified human skeleton mannequin consisting of 20 key factors. This, along with the motion information supplied by the accelerometer, offers the system wealthy information to work with, but it additionally protects the privateness of the consumer.
A consultant skeleton mannequin captured by the Kinect sensor (đź“·: H. Solar et al.)
With a view to interpret this information and detect falls, the researchers developed a machine studying evaluation pipeline. Testing confirmed it to be fairly correct because it achieved a 99 % common accuracy charge. Responses had been fast, as nicely, which is essential when an accident has occurred — predictions could be made in simply over one second. Moreover, the algorithm can run completely domestically on a laptop computer, desktop, or perhaps a smartphone. That is an extra privateness safety as information doesn’t should be transmitted over the web for processing on distant servers.
As a subsequent step, the staff plans to discover the potential for incorporating extra sensors into RESAM. By doing so, they could be capable of monitor extra health-related metrics. Trying even additional forward, they envision the system being constructed right into a robotic canine that may comply with its proprietor round because it retains shut watch over them.