Exhibitor: Ambiq
Hall/Booth: 4-581
heartKIT harnesses Ambiq‘s ultra-low-power hardware to deliver the most energyefficient embedded heart analytics available, enabling the development of AI-powered heart-monitoring solutions that can operate for days or even weeks on a single coin cell battery.
The heartKIT ADK is designed for maximum flexibility. Instead of offering a limited set of pre-trained models, it provides a customizable platform that allows developers to modify existing solutions and incorporate their own datasets, models, and tasks.
Moreover, heartKIT covers the entire data processing workflow, including signal denoising, segmentation, and classification. This comprehensive approach enables the creation of sophisticated, multi-stage pipelines.
For example, a pipeline could include a denoising model to eliminate motion and noise artifacts, followed by both a segmentation model and an arrhythmia detection model. The arrhythmia model identifies specific conditions, while the segmentation model provides key points of interest, enhancing the context for arrhythmia detection. This integration of models results in explainable, transparent pipelines—a critical feature for heart-monitoring applications.
Website of the product and Website of the company
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.
Exhibitor: NXP Semiconductors
Hall/Booth: 4A-222
eIQ Time Series Studio is an end-to-end tool that streamlines the development of edge AI/ML solutions in various domains, including anomaly detection for predictive and preventive maintenance, classification, and regression, based on time series data. The integrated autoML (automated machine learning) capability differentiates this tool from others. With autoML, it significantly reduces the data science and machine learning expertise required for developers, accelerating development time from weeks or months to hours.
autoML automates signal processing, model selection, hyperparameter tuning and feature engineering. Once data is curated, with a single click, models are generated with optimal accuracy, performance, and memory footprint, allowing developers to select the best model for their specific requirements.
eIQ Time Series Studio is designed to lower the ML development entry barrier, making AI development accessible even for developers with low data science and machine learning experience. Integrating automated machine learning, development time can be accelerated from months to hours. Prioritizing data privacy, it stands apart from cloudbased tools by running offline on Windows and Ubuntu based systems, eliminating the need to upload proprietary data to the cloud for training. Additionally, the tool is completely free with no recurring cloud subscription fee, making it an ideal tool for developers interested in time series development.
Exhibitor: STMicroelectronics
Hall/Booth: 4A-148
Our product revolutionizes the way you interact with edge AI and IoT solutions. We offer a seamless approach to dataset creation and management, both offline and in real-time, ensuring you have the data you need when you need it. Our simplified training process, leveraging AutoML and a no-code experience, empowers anyone to build sophisticated machine learning models effortlessly based on decision trees for inertial MEMS sensors.
With our solution, you can validate model performance directly on hardware sensor nodes via intuitive web and mobile apps, guaranteeing real-world reliability. Deploying AI solutions in an IoT context has never been easier, whether it's on sensor nodes, gateways, or the cloud. We also enable the exploration and evaluation of use case proof-of-concepts using ST reference designs and components, providing a solid foundation for innovation. Finally, our robust project and model organization management ensures that every aspect of your development process is streamlined and efficient, allowing you to focus on what truly matters: creating groundbreaking AI solutions in IoT market.
ST AIoT Craft is the first online tool designed to program ST MEMS sensors embedding a machine learning core (MLC) for inference operations, making Artificial Intelligence of Things (AIoT) accessible to all. This tool significantly lowers the barrier to entry, enabling even developers with minimal experience to have their projects ready in under an hour.
