In our previous post, we introduced ITTIA DB Lite for data computing on STMicroelectronics devices, such as STM32H7, STM32U5, and STM32H5. We showed examples on how embedded software developers use ITTIA DB Lite to efficiently update and query data stored on local flash media. By moving data computing tasks to the embedded systems where sensor data and other data sources originate, device manufacturers can offer responsiveness and autonomy that isn’t possible with remote Cloud solutions.

Data management for microcontrollers such as STM32 devices play a crucial role in modern automation by enabling real-time data collection, processing, and control, which is essential for optimizing processes, improving efficiency, and ensuring safety at the place of origin. For the edge computing, embedded systems must also facilitate data communication, monitoring, and control, ultimately leading to more intelligent and interconnected data operations.

Real-time stream processing is another key requirement for data management at the edge because it enables device applications to clean data before it is stored, transmitted, or combined with an AI model for machine learning inference. With ITTIA DB Lite stream processing queries, you can write code to dynamically filter, aggregate, and combine live sensor data. With complete control over task priority and memory resources allocated to each stream processing query, you can ensure that high priority tasks respond to critical events with deterministic performance. If you are interested in our comprehensive examples demonstrating how ITTIA DB Lite and STM32 devices will add value to your new design, as well as data cleaning, contact ITTIA and ask for STM32 examples.

Let’s look at a real-time example. STM32U5 is a series of advanced, power-saving microcontrollers based on Arm® Cortex®-M33 for smart applications like wearables, HMI, personal medical devices, home automation, and industrial systems. With up to 4 Mbytes of flash memory and 3 Mbytes of SRAM, the STM32U5 series can operate at a maximum frequency of 160 MHz. To minimize power consumption, STM32U5 device applications must strictly limit how much data is stored or transmitted.

ITTIA DB Lite stream processing examples for STM32U5—and many other STM32 devices—showcase how you can easily write code to downsample live sensor data in a background RTOS thread or task. Downsampling is a data management technique that decreases the number of data samples in a dataset, and it aims to correct imbalanced data which has benefits such as improving model performance and reducing the size of the dataset.

Our example for STM32 stream processing demonstrates real-time data aggregation with ITTIA DB Lite for monitoring, collecting, and processing data streams on-device. This enables timely decision-making and real-time responsiveness directly on the STM32 device itself.

The STM32H5 series of microcontrollers is based on the 32-bit Arm® Cortex®-M33 core, running as high as 250 MHz. Offering a balance of performance and security, it includes up to 2 Mbytes of flash memory and up to 640 Kbytes of SRAM. The increased MCU performance makes STM32H5 ideal for more advanced data cleaning algorithms.

The ITTIA DB Lite sensor fusion example for STM32H5 and other STM32 devices demonstrates how to streamline and automate data cleaning by identifying patterns across multiple sensors to correct errors and ultimately improve data quality and efficiency.

ITTIA DB Lite address data management with microcontroller challenges due to limited resources, including processing power, memory, and power consumption. ITTIA DB Lite also empowers devices to store data persistently, even recovering from a power failure. A major benefit is the ability to organize and retrieve data efficiently within the limited memory space of microcontrollers.

ITTIA DB Lite is available for evaluation for various STM32 devices.