It may have seemed unlikely when industry analysts started to estimate there would be 50 billion or more IoT devices coming online just a few short years ago. But now, over 30 billion Wi-Fi enabled devices are in service, according to Wi-Fi Alliance, just 20 years after the technology was introduced. In this context, it seems perhaps more probable that Wi-Fi-enabled IoT will grow to comprise tens of billions of connected devices.
Of course, Wi-Fi isn’t the only wireless technology in IoT devices and many devices still use wired communications, but Wi-Fi has evolved quickly over those 20 years and it continues to adapt to new demands as they emerge. What isn’t so apparent, perhaps, is how the industry is using Wi-Fi for applications other than just wireless communications. For example, research shows how Wi-Fi signals can be used to transfer power, as well as data, using a method commonly referred to as ambient backscatter. But perhaps more tangible is how Wi-Fi is already being used as a motion-sensing and detection technology.
Motion Sensing Using Wi-Fi
Before getting into how let’s just explore why we should consider using Wi-Fi as a motion-sensing solution. First, as already explained, Wi-Fi is now virtually everywhere, so the infrastructure is already all around us. Second, the IoT provides the perfect ecosystem for using data like motion sensing in more useful ways. Together, motion-sensing using Wi-Fi can be applied to new and existing use-cases. This includes conventional residential monitoring, to detect an intrusion or scheduled maintenance in and around our homes. This can easily be extended to smart buildings, using motion and occupancy detection as part of a building management system for HVAC and lighting control. Perhaps more impactful, it can be used to help us monitor elderly and vulnerable people in their homes for peace-of-mind.
The most important indicators for performance in Wi-Fi-based motion sensing are the same for other methods of motion detection, specifically accuracy, precision and latency. Accuracy relates to detecting motion when someone is moving or present while minimizing the number of false positives the system may generate. This emphasizes the need for total Wi-Fi coverage in the home, buildings or other areas being monitored. Precision, in this context, relates to how the system differentiates between types of movement being detected, including the speed and direction. This is important if the system is being used to monitor an elderly or vulnerable person in the event of a fall, for example. Latency, as it is generally understood, is the delay between detecting a movement and reporting the event. In any motion detection solution, this needs to be measured in milliseconds, rather than minutes.
Building on Existing Hardware
Fundamentally, using radio frequency (RF) to detect objects or people relies on a transmitter to send signals and a receiver which characterizes the signals it receives. Any object, whether organic or inanimate, will influence the characteristics of the signal as it travels through the medium or channel. In this way, the channel itself can be characterized and in doing so, changes in the channel can be detected.
Clearly this approach means that, while there is really no precise way to identify objects in the channel, it can detect that there are objects or persons present. The absorption or reflection of RF energy as it travels through the channel will result in measurable differences in the signal between transmitter and receiver. The key is turning this measurable difference into actionable data and this is where the underlying protocols used in Wi-Fi come into play.
In a typical home or business environment, a wireless router or Access Point (AP) has devices associated to it. The Wi-Fi protocol requires the connected device to respond to the AP with feedback that allows the AP to optimize the way it communicates with each connected device. Part of this feedback contains data, referred to as the Channel State Information or CSI. The AP can use the information contained in the CSI to comprehend its environment.
Figure 1: CSI data can be used to describe the wireless landscape
Once the AP knows what its wireless environment ‘looks’ like, it is relatively simple to identify changes in that landscape. It is these changes in the way the RF energy is absorbed or reflected that correlates with something moving or being moved in that environment. Part of the potential here is that the connected devices are essentially oblivious to their role in the application, which means the scalability of Wi-Fi-based motion detection is entirely dependent on the technology inside the AP.
Figure 2: Using CSI data to identify movement
It is also possible to pass the data extracted from the CSI to a trusted 3rd party applications, potentially at the edge or in the cloud to process data and intelligently sense motion. Several ON Semiconductor partners are already doing this, using AI to analyze raw CSI data to deliver motion detection as a service. This can enable advanced features, such as detecting when someone falls, is on the floor or when more people than expected are in a building.
Higher Performance Wi-Fi Delivers Better Motion Sensing
The performance of motion sensing using Wi-Fi is largely dependent on the quality of the Wi-Fi coverage, with better coverage systems that can improve on the key parameters of accuracy, precision and latency. As we all probably know, many things can degrade Wi-Fi performance, but the industry is always striving to improve. This is where MIMO (Multiple-Input, Multiple-Output) technology is making a real difference. MIMO and Multi-user MIMO (MU-MIMO) use multiple antennas to send and receive multiple data streams to multiple clients concurrently, taking advantage of multipath reflections of Wi-Fi radio waves in the environment. MIMO also uses beamforming to extend the range and improve the performance of a Wi-Fi network. In addition, the information contained within the multipath CSI data provides greater resolution in terms of motion detection by moving from 8-bit to 12-bit or even 16-bit CSI data.
Figure 3: A graphical representation of multipath
The evolution of Wi-Fi has seen a progressive and sustained expansion in the channel bandwidth used. The benefits here are significant in terms of motion sensing, as moving from 40MHz to 80MHz channel bandwidth delivers a corresponding increase in motion-sensing accuracy of 40%. This increased accuracy is now being used to support value-added services, as mentioned earlier.
Motion sensing using Wi-Fi is the latest example of how this versatile technology continues to evolve, providing valuable new services. The development of our MU-MIMO technology will further enable manufacturers to fully leverage the potential of Wi-Fi.
Learn more about ON Semiconductor’s Wi-Fi Solutions.
