{{else if Image}}
{{else if Image}}
{{/if}}
{{/if}}
embedded award 2025: Tools nominees
Automated tools increase efficiency and reliability at all levels of the embedded value chain. There are tried and tested tools that have been used for years. However, in order to further optimise development processes, the need for new tools is always there. A look at the nominees in this category is definitely worthwhile ...
A customized AI accelerator, a truly generic test system for microcontrollers and a configurable workflow builder
Catapult AI NN
Exhibitor: Siemens EDA
Hall/Booth: 5-119
The High-Level Synthesis engineering team at Siemens EDA has over 20 years of experience delivering tools that compile abstract algorithms into deployable hardware designs. Hardware designers at some of the world’s leading technology companies use these tools to dramatically improve their productivity designing and verifying some of the most complex hardware designs. Now, the High-Level Synthesis team has turned their focus to AI algorithms.
AI algorithms are enormously compute-intensive. Many require some kind of acceleration to be practical, especially at the edge. Customized accelerators can outperform off-the-shelf parts or standard IP by more than a factor of 10, delivering greater performance while improving efficiency. By combining state-of-the-art synthesis technology with advances in AI frameworks and neural networks Siemens EDA‘s Catapult AI NN delivers an automated path from neural network to a customized AI accelerator.
It will automatically quantize the network to achieve an optimal power, performance, and area (PPA), reducing size cost, and energy consumption. What used to be a development effort that spanned disparate technology domains is now an automated flow, allowing system designers to more aggressively pursue the benefits of AI for their embedded products.
Would you like to delve deeper into the topic?
At embedded world Exhibition&Conference 2025 from March 11 to 13, 2025,
you will have the opportunity to exchange ideas with industry experts.
ExecutionPlatform - Digital twins for microcontroller testing
Exhibitor: embeff GmbH
Hall/Booth: 2-639d
Microcontroller code interacts with the external environment through its pins and the custom PCB. We simplify software testing dramatically by eliminating the need for the custom PCB during testing.
In our approach, the only project-specific component is the microcontroller itself, with all its pins connected directly to the test system. At the pin level, the ExecutionPlatform supports a wide range of common digital interfaces—such as GPIO, UART, CAN, I2C, and Ethernet — as well as analog interfaces.
This comprehensive interfacing allows you to reuse a standardized set of logical test Endpoints across multiple projects. Using these logical test Endpoints, you can create digital twin simulations of the environment. This innovation makes the ExecutionPlatform the first truly generic test system for microcontrollers.
The ExecutionPlatform stands out with its fully network-based design, enabling 100% remote use. There is no need for physical wiring for specific tests, making it the most automated solution for hardware-related testing in the industry.
AI Labeling
Exhibitor: EdgeImpulse Inc.
Hall/Booth: 4-505
AI Labeling innovates by seamlessly integrating advanced foundational models with users' custom-trained models within a user-friendly interface. Our configurable workflow builder allows users to chain state-of-the-art LLMs from Hugging Face and OpenAI, enabling rapid automated labeling of large datasets in minutes. This tool not only facilitates new data labeling but also excels in validating and refining existing labels, quickly identifying mislabeled data in pre-labeled datasets.
A key innovation is the system's ability to automatically process new data added to the user's data lake, applying these advanced labeling and validation techniques, and notifying users upon completion. This combination of cutting-edge AI, intuitive design, and automated workflows significantly streamlines the data preparation process, reducing time and effort while enhancing label quality and consistency across various data types.
