Fall-detection tech gets smart

A report by the Australian Institute of Health and Welfare (AIHW) shows that more than 100,000 people aged over 65 were hospitalised due to a fall in 2016–17, with injuries to the hip and thigh (22%) and head (26%) being the most common fall-related injuries. The Australian & New Zealand Falls Prevention Society has outlined that falls represent a major health issue, with the incidence and cost to the health system set to increase as the population ages.

To help address this problem, an Australian company has created an artificial intelligence (AI) device that can detect if someone suffers a fall in the home, hospital or aged-care facility.

Similar in size to a small portable speaker, the HomeGuardian.ai device uses sensors to monitor the interaction of objects and people within its surroundings and alerts carers or family members if a fall occurs — the development team spent more than one year teaching the device to recognise objects such as furniture and people, along with the ability to identify abnormal behaviours.

HomeGuardian.ai Co-founder and CEO Kane Sajdak said the device’s ability to track objects without saving or sending images or video will revolutionise the aged-care, disability and hospital sectors.

“The device knows it is normal for an object it recognises as a person to lay on a bed, but it also knows it’s not normal for that same object to lay on the bedroom floor or bathroom tiles.

“It takes less than two seconds for the device to recognise someone may be at risk and immediately sends a notification to their emergency contact, be it a neighbour, loved one or nurse in an aged-care facility or hospital.

Sajdak explained that an alert can also be escalated to other people if the primary contact fails to respond within a designated period.

“Everything we have developed focuses on getting help to the user as soon as possible.

“Falls cost Australian governments more than $3 billion in hospital admissions each year and it’s directly related to how long people are on the floor. The sooner you get to them, the better their chance of avoiding major issues or complications.

“On top of this, many elderly people go from living independently in their own home into an aged-care facility as a direct result of a traumatic incident such as a fall,” he said.

To ensure the technology would enable better care for those using it, as well as provide staff efficiency gains for care providers, HomeGuardian.ai worked with aged-care groups and Southern Cross University.

Sajdak said that aged-care homes, hospitals and childcare centres can link the devices to an alerting dashboard, allowing staff to respond where and when they are needed.

“It’s akin to having a staff member monitoring each room 24/7.

“This device is the perfect union of cutting-edge technology and practical application. It will allow people to live independently for longer, provide care providers with the best incident detection technology in the world and, crucially, save lives.”

Q&A with Co-Founder and CTO of HomeGuardian.ai Adam Carroll

What led you to pursue AI technology to assist people at risk of falls?

The idea came from my wife, who is an aged-care nurse. She often discussed the challenges associated with monitoring large numbers of patients, emphasising how long a round of checks could take, as well as what could happen in the time away from each patient.

Once the team began chatting with friends and family, we realised everyone had an experience — whether it be a mother falling in the middle of the night and left on the ground for hours helpless, or a partner susceptible to fainting due to an ongoing medical condition. From there, we identified a need for a device that was minimally invasive, but had enhanced monitoring precision — to be in the corner of a user’s home without the need for them to interact.

What is involved in teaching AI to recognise objects and identify abnormal behaviours?

In its simplest form, AI is developed by feeding imagery and video content into the computer so it begins to recognise behaviour and activity. In this instance, we shared thousands of images including people or objects on the ground and a dog jumping off furniture — all aspects that could be involved in a fall or perceived as a fall. From there, we teach the AI to identify what is a fall and what is not. Through the education process, the technology begins to understand the difference between a pet and human, and so on. The device then has the capability to continue to recognise habits and behaviour in the household and can adapt to its surroundings.

What was the main challenge you faced in developing the technology?

To maintain security and privacy, our AI processes everything within the device — we do not transmit information to the internet for processing like other AI products. Conducting this type of processing on the device is very difficult as one of the main challenges is ensuring we are processing high levels of data as quickly as possible. The second is making sure we’re doing it securely on the device to maintain the privacy and dignity of its user.

In what other applications might the technology be used?

AI has endless possibilities and opportunities to fit into everyday life. Our team are continuously testing the device’s capabilities and have a continued focus on developing the technology to identify further dangers and risks. This could include the detection of weapons in a school classroom, monitor the existence of drowning in a pool or supervising newborns to detect indications of SIDS (Sudden Infant Death Syndrome). At its core, HomeGuardian.ai aims to develop technology that can detect risks beyond what is humanly possible and most importantly, save lives.

How do you see the technology coming into its own in the face of the COVID-19 crisis?

HomeGuardian can put families and care workers at ease in this chaotic time. With the strict visitation restrictions put in place on aged-care facilities, hospitals and homes, families can have peace of mind knowing their elderly relatives can be monitored without additional physical contact and interaction. The device can not only safeguard our elderly population but also dramatically reduce the risk of infection and provide relief to overworked and under-resourced medical staff.

Are there plans to develop any associated technologies to assist the healthcare sector?

At this time, HomeGuardian is the best artificial intelligence-based fall detection product on the market, combating a very real problem for our older generation. Naturally, as research and implementation continues, the extent of its capabilities will broaden, and we endeavour to offer further tech to assist the healthcare sector in future years as the device becomes a staple in homes and facilities.