A Device Edge Solution for Active Assisted Living

Tuesday 12/15/20 09:34am
|
Posted By Rajan Mistry
  • Up0
  • Down0

Qualcomm products mentioned within this post are offered by
Qualcomm Technologies, Inc. and/or its subsidiaries.

As the population ages, the demand for technology which can facilitate independent living is increasing. Supporting this trend is both a challenge and a goal for healthcare systems around the world. As a result, many are turning to active assisted living (AAL) technology which encompasses concepts, products, services, new technologies, and the social environment to improve the quality of life for this aging demographic.

A long-established and key requirement in AAL is to monitor the movement of people as a way to detect incidents and emergencies (e.g., injuries from falls). This has traditionally been accomplished using passive infrared sensors, but such devices were not always capable of detecting critical situations accurately or quickly enough. In addition, many conventional AAL devices require manual operation, which isn’t always practical because the person involved may be confused, suffer from dementia, or may even be unconscious. Furthermore, many falls happen at night while the user is out of reach of the emergency button which is commonly installed by the bedside.

Recognizing these challenges, Berlin-based pikkerton aims to help with their Grannyguard intelligent, fall-detection system. Using two infrared sensors, the Grannyguard builds a heat map of the person’s local area (e.g., a room) and analyzes that data locally with sophisticated software algorithms to infer if a person has fallen. This happens offline and inside the device – the protection of private data matters. The device can then send out an alert to caregivers.

Figure 1 - Overview of the Grannyguard device.
Figure 1 - Overview of the Grannyguard device.

In this blog we talk to pikkerton’s CEO, Lothar Feige, to find out more about this unique solution, how the idea came about, and the technology that makes it all possible.

Actively Taking the Passive to the Next Level

According to the Grannyguard website, falls account for more than 75% of accidental deaths at home for people aged 65 and over. Lothar says that timing is critical in such emergencies, so devices should send alerts about critical situations immediately after recognizing a lack of activity by a person. Unfortunately, most traditionally passive devices fail to meet this requirement. With such dire statistics and consequences, pikkerton felt that technology could better detect falls and provide faster alerts.

Building on the idea of using 100% passive sensors, pikkerton sought to develop a more intelligent solution that takes advantage of today’s IoT chipsets and communications technology while still remaining energy efficient. Using ultra-sensitive sensors to analyze infrared radiation from the human body, the Grannyguard detects and analyzes thermal energy signatures, avoiding the need for cameras and other sophisticated sensors. Figure 2 illustrates how this could look from the device’s perspective:

Figure 2 - The Grannyguard analyzes heat maps to detect a fallen user.
Figure 2 - The Grannyguard analyzes heat maps to detect a fallen user.

When a fall is detected, Grannyguard’s built-in 2G/LTE RF cellular module then attempts to contact one or more pre-defined contacts (e.g., caregivers, family members, etc.) via voice or SMS. If voice contact is established (i.e., the contact answers their phone), then the Grannyguard’s microphones and speakers provide hands-free, two-way conversation.

Figure 3 - The user pre-configures the Grannyguard with one or more persons to contact in an emergency.
Figure 3 - The user pre-configures the Grannyguard with one or more persons to contact in an emergency.

The Grannyguard detects falls within the user’s living space via a detection range of approximately 5m spanning a 170° field of view, which provides intelligent coverage of the whole room. Grannyguard can also detect suboptimal heating and ventilation, fires, humidity levels, and general inactivity (e.g., if no activity is detected in a room for certain time ranges).

Security

As with all IoT devices, especially those which incorporate intelligence at the edge, maintaining security, privacy, and confidentiality become key requirements.

The Grannyguard has a number of attributes that make it inherently secure. For starters, it has no camera technology, it simply analyzes heat maps looking for patterns that would indicate a fall. As well, with heat maps providing very generalized clusters of colors, there is no specific information to identify gestures or body parts which retains confidentiality. Grannyguard works offline 24x7 and also has no Internet connection, nor any need for one, as there is no data to transmit from the device, nor is there any cloud backend. Analysis is done directly on the edge device and if a fall is detected, communication via SMS and voice calls take place over cellular, directly with contacts who have been pre-configured on the device.

Installation

Installing an IoT device might conjure up images of complex electronic wiring, but pikkerton wanted installation of the Grannyguard to be as simple as any other hard-wired electrical appliance. As a result, the unit is installed by hardwiring it into an existing flush-mounted AC wall socket. This is commonly done by replacing a light switch with Grannyguard so the device has a good overall view of the room. And to retain the toggle switch functionality (e.g., to control a ceiling light), the Grannyguard includes its own on-board push button. Of course, the light can also get switched on automatically, which makes sense (e.g., in the bathroom). Combined with a small white LED for floor illumination, Grannyguard might prevent falls in the dark.

Once installed, contacts are configured on the Grannyguard using a mobile device over Wi-Fi.

Development Process

Lothar says that while the hardware development process was fairly straightforward, hardware designers had to package the radio frequency (RF) and antenna hardware in a very small form factor (i.e., to fit a standard AC wall socket) which was a challenge in itself.

The primary challenges were related to the situation analysis algorithms, since viewing the world through infrared is very different than how humans perceive the world. Many variables and unexpected effects had to be identified and accounted for including pets, small obstacles and obstructions, warm and cold spells, mirror effects from glass doors, etc.

pikkerton planned for the development process to last around two and a half years for the device, and all in all it took around four years for it to mature to the point where it was ready for commercialization. In the end, certification was awarded by VDE in Germany – an organization responsible for testing and certifying electrical devices, components, and systems. pikkerton has won several international innovation awards with Grannyguard (Deep Tech Award, Elektra Awards, Building better Healthcare, Telekom Digital-X, German Stevie Awards, Demographic Excellence Award).

The Technology Behind the Tech

During early development, pikkerton used the Quectel M66 RF module to give the device 2G connectivity, and has since evolved the design to use the Quectel BG96 and BG95-M3 for NB-IoT and LTE-M multi-mode Low Power Wide Area (LPWA). Lothar says this hardware provides just the right-sized form factor, good RF performance, and is easy to integrate, while the manufacturer provides fast support.

Developers familiar with the Quectel BG95 may recall that it’s built around the Qualcomm® 9205 LTE IoT modem. With our 9205 modem, developers have the option of using the platform either as a standalone modem, or to allow its built-in cloud support using the Qualcomm® LTE for IoT SDK. Developers who are interested in learning more should check out the Quectel LTE IoT development kit.

Conclusion

The Grannyguard intelligent fall-detection system is an innovative solution that is helping to overcome common challenges with traditional AAL devices. From a technology perspective, we’re excited to see how the Qualcomm 9205 LTE IoT modem is facilitating new use cases for IoT and how Quectel is utilizing this technology to help contribute to the health, safety, and well-being of others as the demand for AAL increases.

For additional information be sure to check out the following resources:

About pikkerton GmbH

pikkerton GmbH is a Berlin-based electronics company that was founded in 2004. They design, develop, and build electronic systems for many markets including smart energy, smart grid, industrial monitoring, and AAL. One of pikkerton’s primary goals for its AAL systems is to help people live safer and longer in surroundings which are familiar to them.

Qualcomm 9205 LTE Modem and Qualcomm LTE for IoT SDK are products of Qualcomm Technologies, Inc. and/or its subsidiaries.

Sections: