The Linux Foundation Projects
Skip to main content

People Fall Detection via Privacy Preserving AI

By January 14, 2021No Comments

Guest Authors: Brenda Brian, Ivelin Ivanov from

For many adults, one of the most difficult decisions to make is how to care for an elderly parent or relative that needs assistance. The AARP has found that almost 90% of folks over the age of 65 prefer to remain independent by living out their golden years at home. 

Whether living alone or with family members, elderly parents need constant monitoring. Why? This is because as they age, their risk to potentially life-threatening accidents increases. 

In fact, a slew of researches reveal that seniors are more prone to fall than other age classes.  According to the US National Council on Aging:

  • One in four Americans aged 65+ falls each year.
  • Every 11 seconds, an older adult is treated in the emergency room for a fall; every 19 minutes, an older adult dies from a fall.
  • Falls are the leading cause of fatal injury and the most common cause of nonfatal trauma-related hospital admissions among older adults.
  • Falls result in more than 2.8 million injuries treated in emergency departments annually, including over 800,000 hospitalizations and more than 27,000 deaths.
  • In 2015, the total cost of fall injuries was $50 billion. Medicare and Medicaid shouldered 75% of these costs.

Medical alert systems were introduced into the home healthcare space to help prevent medical complications for independent seniors. 

At present, the global medical alert system, which includes a variety of landline and mobile personal emergency response systems and standalone or wearable devices is witnessing rapid growth. 

Examples of medical alert systems.

Some medical alert systems include a fall detection feature. They generally comprise two major components: cell phones and wearable devices. Once the wearable device detects fall, it sends an alert to the cell phone, and then the cell phone notifies the emergency contacts the user selected.

While these systems are recommended by AARP and caregivers, a recent study found that only 8% of seniors believe they need a medical device. Most are unsure or unwilling to wear one:

This is where the Ambianic Fall Detection remote monitoring system can help. Ambianic is an open-source software project, which allows the developer community to take part in solving the global challenge of healthy ageing by offering round the clock surveillance of elderly relatives or patients and instant alert to family caregivers in the event of an accident or fall. 

A recent publication in Nature discusses the potential of ambient contactless sensors in healthcare. A systematic review found that ambient sensors can detect falls with 97% accuracy vs 96% for wearables. 

What is Ambianic Fall Detection? 

Ambianic Fall Detection is based on computer vision technology. It receives video feed from a wall mounted camera, then uses PoseNet to detect a person’s body vectors and determines whether there is a rapid downward motion between frames. When a fall is detected, it alerts family caregivers and presents them with contextual information to make an informed and timely decision. Helping a fallen senior within the first golden hour drastically reduces fatalities

How Ambianic Preserves Privacy

Nowadays, safety and privacy are difficult to achieve, particularly when it involves remote home surveillance. 

Due to the advancement in technology, most remote monitoring solutions available, offer an easy plug-and-play experience which makes them popular. 

By design, these systems upload, store and process camera footage on the vendor’s cloud.  However, nearly all vendors -big and small fail to protect user data.

Notable data privacy violations by remote monitoring companies include leveraging user data for targeted advertisement, sending data to foreign governments, leaking user data due to product bugs, and sharing data with law enforcement without user consent. Data stored on the public cloud are also vulnerable to hacks. 

These data privacy concerns are essential to Ambianic’s solution, which has received multiple awards. Ambianic eliminates the stigma associated with surveillance systems by implementing privacy-preserving algorithms in three critical layers: 

  • Peer-to-Peer encrypted remote access 
  • Local device AI inference and training 
  • Local data storage

This solution monitors your environment and unlike other surveillance systems, only alerts users when it observes meaningful, actionable and interesting events. 

In other words, Ambianic only tells you about things you actually care about or want to know about. Also, camera footage and user data generated are not sent to any third party cloud server and are only available to homeowners and their family members. 

Thus, with Ambianic Fall Detection in your home, you can effectively monitor your elderly relatives and never have to worry about your sensitive data being shared, leaked or sold without your explicit knowledge and permission. 

Google’s Tensorflow Community Spotlight, and World Innovations Summit for Health are two prestigious awards under Ambianic’s belt. 

How It All Works 

Basically, Ambianic systems consist of three major components: 

They work in synergy to ensure efficient privacy preserving monitoring. 

Ambianic Edge is a Python device built to run on an IoT Edge device such as a Raspberry Pi or a NUC. It embeds video cameras and other sensors to gather data and then runs inference pipelines using AI models that detect events of interest such as objects, people and other triggers. Ambianic UI, on the other hand, is a Progressive Web application written in Javascript using Vue.js and other front end frameworks to provide an intuitive timeline of events to the client user. 

For the Ambianic UI and Ambianic Edge to discover each other seamlessly and communicate, WebRTC was introduced. WebRTC is a secure peer to peer (p2p) communication protocol and HTML5 browser API that enables browser apps to communicate directly with other browsers and IoT devices. 

Once the two peers in the communication channel – Ambianic UI and Ambianic Edge – locate each other’s UUID, they can open a signalling channel over HTTPS and Secure Web Sockets (WSS) via Ambianic PnP and negotiate the most direct possible routes to each other for their secure data exchange. User data is not exchanged outside a negotiated p2p channel at any time. 

This approach reduces data privacy risks, among other benefits. One of such benefits is lower latency. Crucial alerts travel in real-time directly from the edge device to the UI using the shortest possible Internet route with no detours or short stops at third party cloud services. Also, it eliminates the need to build and manage a complex cloud service which is in the critical communication path. More technical deep dive is available in this blog post.

How To Install Ambianic

Recommended Installation: The simplest approach to get started consists of the following steps:

  • First, assemble an Ambianic Box – an open source physical enclosure of the Ambianic Edge device using off-the shelf hardware components.
  • Next flash the Ambianic OS Image to an SD card.
  • Once you’ve flashed your SD card, you can insert it into your Raspberry Pi that powers the Ambianic Box.

Next, connect your Ambianic Box device to WiFi with these steps:

  • Use your mobile phone to scan for new WiFi networks and connect to the hotspot named Ambianic Edge WiFi Setup.
  • Wait a few moments until the captive portal opens, this portal will enable you to connect the Raspberry Pi to your local WiFi network.

After the WiFi connection is established, the Raspberry Pi will pull the latest Ambianic Edge binaries. This may take 5-10 minutes depending on your internet connection speed.

Once the update process is completed, you can proceed to pair the edge device with your Ambianic UI app as described below.

Alternate Installation: For developers and more advanced users, an alternative Docker based installation is also available.

Your Ambianic instance is up and running. 


Both adult children and caregivers constantly worry about the health of seniors under their care. With the high rate of falls among older adults, it has become increasingly crucial to figure out ways to keep eyes on them. Forward-thinking developers examined these challenges and introduced Ambianic Fall Detection. An Open Source AI-powered remote home surveillance system that can detect falls and instantly alert family caregivers, and ensure that private data doesn’t get leaked or worse, sold. 

Click here to get a pre-deployed Ambianic app.

Going forward, the development team is planning on additional features such as detecting fever or detecting anomalies in daily activities. Your thoughts and contributions are most welcome! Join the community discussion.

LF AI & Data Resources


  • Andrew Bringaze

    Andrew Bringaze is the senior developer for The Linux Foundation. With over 10 years of experience his focus is on open source code, WordPress, React, and site security.