Our contextual technology is based on one key principle: simplicity. We rely on existing wireless signals to recognize people (and animals too), gather data about what they are doing and where they are located. Based on this data, we can automate workflows and provide contextual information that can be leveraged by a variety of applications.
When WiFi signals propagate through a given space, they interact with walls, doors, objects, and human bodies. A person moving through a wireless mesh (also known as sensing area) will disrupt the signals very differently based on what the person is doing. This is what we refer to as the unique signature. Every person has his or her own signature. Thanks to our machine learning engine and proprietary algorithms, this is how we can identify who is doing what, where, and when. That is the power of Aerialytix.
Our SaaS can easily be integrated into smart objects and other WiFi-enabled devices, such as routers, making Aerialytix simple to get to your end users without having to install more equipment.
Contrary to other solutions available in the market, Aerialytix’s innovative technology captures data without end users having to wear sensors or devices. It is truly passive.
Aerialytix easily integrates into existing infrastructure, therefore reducing the hardware, deployment and maintenance costs.
Our solutions are suited for multiple applications involving humans moving through spaces where wireless signals are available.
The raw data captured by our solution, data that is transformed into information through our machine learning engine, is then anaonymized, compiled and made available for various purposes (targeted advertising, healthcare players, etc.).
Meet Jeanine. Her father, George, just had hip replacement surgery.
George, a widower, is still autonomous and lives in a small apartment far away from Jeanine’s home.
Jeanine has gotten help; everyday, a caregiver visits her dad at 10 am. Jeanine still would like to monitor how he is doing the rest of the day without having to purchase a complicated, expensive or intrusive solution with cameras. George values his privacy and doesn’t want to be recorded.
Jeanine remembers reading about an innovative and affordable WiFi-based solution being promoted by her internet service provider (ISP). She contacts her father’s ISP and quickly has access to a mobile application enabling her to monitor the activity in her dad’s apartment in a non-intrusive manner. She also customizes the solution to send her alert emails if her dad.
- Stays in bed for more than 3 consecutive hours, during daytime
- Stays in the bathroom for more than 15 minutes
- Takes less than 100 steps per day
- Does not receive a visitor (the caregiver) before 10h30
Jeanine recovers peace of mind concerning her father’s condition. She is paying a minimal monthly fee on top of the internet package her dad was already paying. She is so satisfied with the application that she is considering keeping it in the long run, after her dad fully recovers. What a great way to monitor the condition of a loved one without intruding on their privacy.
Refer to the Applications section for more use case scenarios
Motion Data Generation
At Aerial, we focus on activity recognition using off-the-shelf wireless devices (access points, laptops, smart TVs, printers, desktop computers, repeaters). One of our key advantages is that our solutions do not require extensive, and often expensive, new hardware. We can leverage what is already there.
When people move through a wireless mesh, they disturb the energy waves and generate specific signatures, based what they are doing. These signatures are then captured by the physical layer of the devices generating our mesh, and transmitted to the cloud using digital signal processing models. This is where our data comes from.
Motion Data Analysis
Aerialytix analyzes data on two levels. The binary analysis is handled locally (presence detection) by the existing wireless devices that have generated a mesh. Then the data is extracted to Aerial’s cloud where more extensive data processing occurs to trigger potential follow-on actions where required (notification management, for example). This is when complex data are built into models to better understand the meaning of motion.
The motion data is segmented into two categories:
- Data interpretation
- Active – to send notifications to end users via the mobile application for the security and home care applications.
- Passive – to activate workflows based on what has been predetermined (for example, dimming the light or tuning on the stove)
- Data transmitted to customers about human behavior. It is very important to mention that this data is anonymized at the source using tokenization and that our customers cannot associate our data with specific individuals or groups of individuals
True Value of Motion
The true value of motion to our customers is twofold:
- Making decisions based on data (For example, in healthcare: a person is not moving enough, so send a caregiver).
- Triggering workflows automatically, passively and adaptively. Data machine learning will improve over time, increasing predictions’ accuracy.
Via our API, our customers will get better real-time tracking of their own customers, assets, premises, etc. Understanding the historical view of the data is what provides intelligence. Past motions and actions will be the basis for predicting the future, and with our machine learning, we will be able to offer better data models that customers can leverage to ultimately make better decisions and better planning