A New Way to Detect Accidental Falls

Sensors on walls could be a faster way to send out an alert for help

21 October 2013

Falling is the second leading cause of injury and death for people ages 65 and older, according to the World Health Organization. Falls often occur when a person is home alone and not within reach of a phone. In some cases, it can be hours until help arrives. 

To solve this problem, two IEEE members from the department of electrical and computer engineering at the University of Utah, in Salt Lake City, have developed a wireless network of sensors that detect when a person has fallen. Unlike present systems, it would not require a person to wear a sensor, push a button, or install a video camera to send an alert.

Instead, the system, being developed by IEEE Member Neal Patwari and IEEE Student Member Brad Mager, can be installed in or on walls to automatically alert an emergency center or loved one when help is needed. The two are in the early stages of bringing the technology to market.

Patwari, a professor of electrical engineering at the school, says such alarm systems currently fall short of user needs. Mager is a computer engineering graduate student.

“Some people will forget to put on a sensor when getting up in the middle of the night, or they feel uncomfortable being videotaped in their own home,” Patwari says. “We saw this as a problem and knew our sensor system could help.”

Their interest in developing a better system came from a simple statistic: The world’s elderly population is already over 1 billion and climbing rapidly.

read Two researchers from the University of Utah use radio frequency sensors to build a wireless system that can detect a fall. Photo: Dan Hixson/University of Utah


RF sensors that determine where someone is in a building or a room already exist, but they won’t detect if a person is lying on the floor. The two decided to see if radio frequencies could be used to create a 3-D system for detecting both the vertical and horizontal position of a person, as well as whether the person was lying on a bed or on the floor, or had fallen down. The sensors—CC2531 USB Dongles from Texas Instruments—are low-cost radio-frequency transceivers. They operate in the 2.4-gigahertz industrial, scientific, and medical band, and they are being used to sense a person’s situation from their motions. 

RF isn’t blocked by walls or furniture, and the sensors can be hidden behind walls or inside other objects, such as a lamp or dresser. Such a system would be less invasive than, say, one employing a video camera or a sensor that must be worn at all times.

The two tried different locations for the RF sensors to determine how best to sense someone standing, sitting, or lying down. They found that placing 24 sensors around the walls of a room was sufficient: One set of 12 is placed lower down, at about ankle height, with the other set of 12 higher up, at the torso. (Note that sensors on one wall can also monitor motion in an adjacent room on the other side of the wall.) The two are working now to get the number of sensors down to as few as possible, with the goal of between 20 and 24 for a 1000-square-foot apartment.

Data from the transceivers goes to a computer for processing. A 3-D image of the person in the room, known as a radio tomographic image, helps determine whether the person fell. For example, if only the lower array of sensors is receiving a signal and not the array at the torso level, the system will register that a person is lying on the floor. By also measuring the amount of time it took the person to get to the floor, the system can determine if someone lay down voluntarily or fell.

To test the technology, Mager set up the sensors in his living room at home. He and another student walked into the room and moved around—standing, sitting, lying down—in different spots. They then fell down. Mager collected and processed data from the sensors to see whether the system could distinguish the different movements, particularly between lying down on the floor and the more rapid act of falling.

With that simple information, the system could in fact conclude if a fall had occurred. Moreover, it could detect various movements that might take place on the floor—for example, doing push-ups as opposed to lying on one’s back. The system could also calculate how long someone stayed down. Once it’s determined that someone has fallen, the system could notify a caretaker with a phone call or text message that help is needed.

“Even if it takes minutes to determine whether a person fell, it’s going to be a dramatic improvement over cases where a person lies on the floor for hours,” Patwari says.


The researchers from Utah have received a small business innovative research grant from the U.S. National Science Foundation. They have six months to demonstrate the commercial potential of the system; if successful, they can then apply for a second grant to continue development. Patwari hopes they can bring the technology to market in two to three years.

In the meantime, the accuracy of the system is being tested in different types of rooms and buildings. Patwari envisions that one day it could be installed in newly built homes or sold as a kit. With installation the system could come with a monthly charge and be operated like a home security system for about US $100 per month.

“The bigger goal here is to allow elderly people to live independently for as long as possible, and this is one component that would help them do so,” Patwari says.

IEEE membership offers a wide range of benefits and opportunities for those who share a common interest in technology. If you are not already a member, consider joining IEEE and becoming part of a worldwide network of more than 400,000 students and professionals.

Learn More