Building IoT Sensors Armed With AI, Wi-Fi 6, and Matter Capabilities

The Internet of Things (IoT) encapsulates a universe of devices and applications that intercommunicate, exchanging data and instructions. At the heart of this web of connectivity are IoT sensors. 

These devices are the nerve endings of IoT, responsible for collecting real-world data—be it temperature, light, pressure, or any other parameters—and translating it into a digital format that can be analyzed and acted upon. The information covered here is based on Silicon Labs’ 2023 Tech Talk: “Designing Connected Wi-Fi 6 Sensors Using SiWx917 and Matter.”

 

IoT Sensors in a Smart Home

 

As these sensors become more integral to our daily lives, their design and function must evolve to meet increasingly demanding criteria. Let's consider some of these critical requirements:

• Connectivity and Range: IoT sensors must reliably communicate their data to a network. Whether it's a smart thermometer sending temperature data to a home thermostat or a fleet of shipping containers reporting their locations to a central hub, the range and quality of connectivity are paramount. For example, the introduction of Wi-Fi 6 represents a significant step forward in this domain, offering faster data rates and improved performance in dense deployment scenarios.
• Size and Battery Life: The versatility of IoT applications demands that sensors be as small and as energy-efficient as possible. The smaller the sensor, the more applications it can be used in—from wearables to smart home devices. Similarly, energy efficiency impacts battery life, the device's environmental footprint, and maintenance needs.
• Ease of Use and Deployment: The explosion of IoT devices has also necessitated focusing on the ease of use and deployment of sensors. This includes simple device pairing procedures, user-friendly interfaces, easy device management, and the ability to integrate with existing IoT ecosystems seamlessly.
• Security: As gatekeepers of sensitive real-world data, IoT sensors must be inherently secure. This entails robust encryption, secure device identification and authentication procedures, and ongoing security updates to address emerging threats.
• Edge Computing: With the surge in data produced by IoT devices, the ability to process data at the edge —within the sensor itself or close to it—is becoming increasingly crucial. This reduces the need for constant communication with the cloud, thereby saving bandwidth and improving response times.

Understanding these demands is the key to unlocking the potential of IoT sensors.

 

Wi-Fi 6: The Catalyst for IoT Sensor Revolution

Wi-Fi 6, also known as 802.11ax, represents the latest generation of Wi-Fi technology. While earlier Wi-Fi standards focused primarily on achieving faster data rates, Wi-Fi 6 is a leap forward in network design, intended to better handle the growing number of devices that need a network connection, with a particular emphasis on improving performance in congested areas.

 

Smart Farming with Wi-Fi 6

 

As we step into an era increasingly driven by IoT devices, Wi-Fi 6 emerges as a key enabler of IoT sensors, offering enhancements that streamline their integration, boost their efficiency, and broaden their applications. 

Now, let's look at some critical Wi-Fi 6 features and how they benefit IoT sensors.

 

Key Feature	Description	How IoT Sensors Benefit
Full Duplex Multi-User MIMO	Full Duplex Multi-User MIMO (MU-MIMO) allows multiple devices to simultaneously communicate with the router, both sending and receiving data at the same time. This is a significant upgrade from previous iterations where data transfer was either upstream or downstream, but not both simultaneously.	It enables multiple sensors to communicate with the router simultaneously, reducing latency and improving overall system efficiency.
Beamforming	Beamforming focuses the Wi-Fi signal towards specific devices rather than broadcasting it equally in all directions.	It improves signal strength and reliability for IoT sensors, enhancing the sensor's connectivity, even in environments where signal strength may be otherwise weak.
OFDMA and Spatial Reuse (BSS Coloring)	Orthogonal Frequency Division Multiple Access (OFDMA) allows for the efficient utilization of channels by splitting them into smaller sub-channels. Spatial Reuse or BSS Coloring helps devices differentiate between overlapping networks, reducing interference.	It allows for better management of network congestion. BSS Coloring improves the recognition of signal interference, increasing network efficiency and data reliability for IoT sensors.
Target Wake Time	Target Wake Time creates scheduled communication between the Wi-Fi router (AP) and the client devices. Instead of constantly staying alert for signals, the client devices "sleep" and only wake up at scheduled times to receive or send data.	Significantly extends the battery life of Wi-Fi 6 client devices.
Extended range packet structure	Provides more guard interval options to improve immunity to signal delays and reflection	Long-guard intervals, cyclic prefix, reduced interference, and extended-range packet structures help improve indoor and outdoor coverage areas.

This table shows some critical Wi-Fi 6 features and how they each benefit IoT sensors.

 

Meet SiWx917: Silicon Labs' Answer to IoT's Rising Demands

Capitalizing on the benefits of Wi-Fi 6, Silicon Labs has developed the SiWx917, their pioneering SoC in the field. This highly efficient, high-performing SoC blends Wi-Fi 6 and Bluetooth LE 5.1 technologies for advanced IoT solutions.

 

SiWx917 features—from Tech Talk presenation by Abhilash Yarragolla Malavya, Senior Engineer at Silicon Labs

 

Designed with a forward-looking mindset, SiWx917 aims to dramatically lower power requirements for Wi-Fi IoT networks while boosting computational abilities, accelerating AI/ML, and providing stringent security. The SoC is a one-stop, Matter-ready solution featuring high memory capacity and unsurpassed battery life. 

Security-wise, SiWx917 is impressive. It's on track for PSA Level 2 certification and comes equipped with a plethora of protective features, such as Secure Boot, Secure Trust Zone, and Secure XIP. This comprehensive security umbrella helps guard devices and end-users against the growing menace of cyber threats.

To learn more about the specifics of SiWx917, you can visit its dedicated solution page.

 

AI and Machine Learning for Sensor Technology

Traditionally, IoT sensors generate a large amount of data that needs to be sent to the cloud or a centralized location for processing. This paradigm, while functional, comes with several challenges:

• Bandwidth constraints: Transmitting huge amounts of raw sensor data to the cloud can clog up network bandwidth, potentially slowing down other network activities.
• Latency: The round-trip data travel from the device to the cloud and back can introduce a significant delay, which is especially problematic for applications that require real-time responses.
• Energy usage: Constantly sending data to the cloud can quickly drain the battery of IoT devices, which is especially critical for those running on batteries.
• Privacy and security: Transmitting data to and from the cloud introduces risks of data interception and breaches.

Here's where AI/ML at the tiny edge comes into play. By processing data locally on the device itself, these challenges can be significantly mitigated:

• Efficient use of bandwidth: By analyzing data at the source, only relevant insights or reduced datasets need to be sent to the cloud, thereby easing network congestion.
• Reduced latency: Edge processing eliminates the need for round-trip data travel, enabling real-time decision-making and responses.
• Offline mode operation: In situations where constant connectivity may not be feasible due to cost or availability—like deep-sea sensors or remote agricultural IoT installations—edge AI/ML allows the system to operate effectively independent of network connection
• Lower energy usage: Sending fewer data to the cloud means lower energy consumption, thereby increasing battery life.
• Improved privacy and security: Local data processing minimizes the amount of sensitive information transmitted, reducing the risk of data breaches.

In essence, integrating AI/ML at the tiny edge can make sensor technology smarter, more responsive, and more efficient, unlocking a host of new possibilities for IoT applications.

 

SiWx917 Supports an AI/ML Accelerator

The SiWx917 comes equipped with a dedicated AI/ML accelerator. This subsystem, specialized in Matrix-Vector Processing (MVP), bolsters ML computations, making the SiWx917 a standout choice for data-heavy, real-time IoT applications.

To illustrate its application, let's consider a use case involving Google Pixel 6 as a Matter controller and the SiWx917 as a Matter node. 

 

SiWx917 use case with sensors using Matter

 

The process begins with the Google Pixel sending network credentials like the SSID and password to the SiWx917 via Bluetooth LE—a step known as BLE commissioning.

Upon receiving the credentials, the SiWx917 connects to the access point, and the Bluetooth LE is disconnected. All subsequent communication between the Matter controller (Google Pixel 6) and the Matter node (SiWx917) is conducted over Wi-Fi, using the IP-based Matter protocol.

The SiWx917 is interfaced with several sensors. It utilizes its AI/ML hardware accelerator to process data from these sensors. This data can then be uploaded to the cloud using MQTT protocols or Matter over Wi-Fi. Moreover, these sensors can be remotely controlled via the Nest Hub using Matter over Wi-Fi.

 

All images used courtesy of Silicon Labs