Falls could be predicted in older people three weeks before they happen using sensors in the home which monitor gait.
The ground-breaking system devised by US scientists is designed to alert medics to changes so they can intervene before a potentially lethal stumble.
Early tests have shown that it can help people stay in their own homes for years longer.
It is made up of several wall-based movement sensors which measure walking speed and length of stride as older people are moving around their homes.
Experts from the University of Missouri, found even small changes can predict if an elderly person is about to suffer a dangerous fall. Risk goes up more than four times if their walking speed slows.
For example when walking speed decreases by 5.1cm per second, pensioners have an 86 per cent chance of toppling within three weeks, compared to just a 20 per cent chance with no change.
They also found a drop in stride length of 7.6cm predicted a 51 per cent chance of OAPs tripping within three weeks.
The scientists developed the technology to help elderly residents live independently for longer.
Falls are one of the main causes of broken hips in the UK and can reveal undiagnosed health problems. One in three pensioners has had at least one in the past year at a cost of £2.3 billion a year to the NHS.
Elderly patients are three times as likely to die following a ground-level fall compared to their under-70 counterparts.
Lead researcher Professor Marjorie Skubic, from the University of Missouri, invented the system after her mother-in-law suffered a bad fall and damaged her shoulder.
She recently fitted the system into her parents' home in South Dakota for her mother's 93rd birthday in the hope it will allow them to stay their until their death.
"You can make a big difference to how someone is going to age," she said. "There was this assumed curve that there had to be a decline, but what we are showing is there doesn't have to be a decline, that you can in fact keep people up at a high level until they die.
"I pray that my parents die in their sleep, in their own beds, in their home. If they can die in their own home where they've been completely independent all this time that is a complete win. That is really squaring the life curve."
Falls in elderly people can often be caused by underlying infections, weakness or clashes with medication, so the system picks up problems early enough so that changes can be made, or rehabilitation given before a potentially fatal tumble happens.
The system has already been trialled on 23 elderly people with an average age of 85. It helped them remain independent for twice as long as those living in other settings.
Those monitored by the technology stayed in their own home for an average of 4.3 years, compared to 1.8 years for those who did not have it.
Elsewhere scientists have begun a project to fit thousands of digital sensors to elderly people in the hope it could help detect the very first signs of Alzheimer’s disease.
The high-tech wristbands which have been given 2,200 older people in Boston, in the US, measure everything from sleep, to balance and fall risk, to heart rate.
Scientists hope the three-year project will reveal subtle physical changes that develop during the first stages of the disease and provide an alternative test for picking up the illness.
Currently it is difficult to diagnose Alzheimer’s and requires a number of tests.
Speaking at the American Association for the Advancement of Science (AAAS) annual meeting in Boston, lead author Rhoda Au said the devices could make the process more simple.
"It's really labour-intensive to bring people into the lab for conventional dementia tests,” she said.
Although there are currently no treatments for Alzheimer’s disease, it is hoped that within the next decade, drugs to help stave off dementia will be available. And it is likely that they will work better if the disease is caught in its earliest stages.
Fellow researcher David Knopman, a neurologist at the Mayo Clinic in Minnesota, said: "The idea of preclinical Alzheimer's disease is that, for people who are destined to develop dementia due to Alzheimer's disease, in the years before they become overtly cognitively impaired, there might be subtle things that change in their daily behaviour that, if we knew what to look for, would disclose who might be at risk."