The number of IoT devices connected to the Internet is rapidly increasing.
A large amount of data transmitted by numerous IoT nodes,
The Liti algorithm-driven system is advantageous for processing


The adoption of IoT technology is rapidly increasing. The number of IoT devices worldwide is expected to reach hundreds of billions in the future. Representative use cases of IoT technology include monitoring environmental parameters such as temperature, humidity, and air quality.

Organizations can make decisions and take appropriate action based on data collected through IoT technology. For example, analyzing the composition of gas emitted from a plant's chemical treatment equipment can help determine whether a malfunction requires maintenance or replacement due to end of life, allowing for appropriate action.

However, effectively analyzing the massive amount of data continuously captured by sensors and determining appropriate actions is a daunting task. While processing data from a few nodes isn't difficult, smart home/farm/building/factory systems typically contain a massive number of IoT nodes.

Automated decision-making processes are essential for maximizing efficiency and achieving rapid response times. This requires advanced algorithms. With so many IoT nodes to process, simply storing a set of rules in a database and applying them to incoming data is ineffective. This quickly leads to processing limitations. Therefore, a more sophisticated yet streamlined approach is needed.

The Rete algorithm is a prime example. Already used in numerous applications, this algorithm was first developed in the late 1970s. The basic framework of the Liti algorithm is a pattern matching mechanism that allows for rapid comparison of large amounts of pattern data with a database containing various objects. This avoids data repetition and allows for the storage of specific pattern states during processing. This eliminates the need to revert to previously applied rules, resulting in faster processing.

UrsaLeo's cloud-based analytics system, based on Silicon Labs' Thunderboard 2 sensor module and a Raspberry Pi 3B+ computer pre-loaded with Yocto Linux, uses the Lity algorithm.

Rather than referencing time and then referencing a lookup table, this system applies a tree-based rule structure to captured data, matching patterns and taking appropriate action accordingly. This allows for efficient decision-making without adding unwanted delays to the system.

You can view your compiled data sets in a customized dashboard, and you can even set up alerts to alert workers when specific events requiring human intervention occur.

You can also use an intuitive visual editor tool to build a set of rules tailored to your application's needs. For example, you can set it to notify you whenever a message is received, or when no messages are received within a certain time interval.

Alternatively, for long-term monitoring purposes, you can configure notifications to occur at specific intervals, such as every hour, day, or week. You can also apply geofencing rules to notify when a mobile node enters or leaves a specific area.

This method can be used in factories to ensure that forklifts do not leave a specific area, and can also be used in fleet management or livestock tracking applications.

If captured data values are different from normal, for example, if they consistently remain the same or consistently display zero, this may indicate that the sensor is not functioning properly. You can configure the system to notify you in these cases, allowing for prompt repairs.

◇ Collects diverse environmental data, capable of processing data from up to 2 million nodes.

The Thunderboard 2 module can capture environmental data from a variety of sensors measuring things like light, air quality, barometric pressure, relative humidity, and temperature, and can even add gas sensing capabilities. It also includes a 6-axis inertial sensor (for spatial orientation) and a Hall-effect sensor (for geographic orientation).

The Raspberry Pi board connects directly to the Google Cloud Platform, acting as an IoT gateway. All compiled data can be transmitted to the cloud via wireless (Wi-Fi) or wired (Ethernet) connections. Depending on the application's needs, the hardware can be powered by USB or lithium-ion coin cells.

This system employs a tree-based methodology, enabling it to handle any number of IoT nodes connected within the network, potentially processing hundreds of thousands of events and alerts per second. A medium-capacity server can handle 500,000 messages per second. A high-capacity server can handle up to 1 to 2 million messages per second.

A system featuring optimized cloud-connectable hardware and the LIT algorithm enables rapid and efficient processing of data transmitted by IoT nodes across various industries. This accelerates the adoption of IoT technology and enables the connection of more nodes.



This article is a summary of the article, "More Effective Analysis of Environmental Sensor Data Through Automated Cloud-Based Analysis," by Mark Patrick, Technical Marketing Manager at Mouser Electronics.