ECOMAI is entering its final phase, which means we have exciting results about our pioneering project to share. Pioneering technologies in an embedded world: The conference and exhibition embedded world Exhibition&Conference in Nürnberg in March offers the perfect stage to demonstrate what this means. In the context of ECOMAI (Ecological Motor Control and Predictive Maintenance with AI), the defined use cases had one goal: develop an Edge AI Solution which is suitable for embedded electrical motor drive systems.
The value chain covers the entire stack of hardware, software as well as a model-based design kit for its architecture: Specialized AI hardware platform, Development kit with AI compiler, model-based design and simulation environment, AI Applications for Predictive Maintenance and Energy-efficient Control to execute on the hardware.
We were happy to present 3 project results at our booth – and having them also presented in talks and tutorials at the embedded world conference.
But first, we also laid down the groundwork for those types of innovations in the tutorial of ECOMAI’s scientific leader Prof. Dr. Daniel Müller-Gritschneder:
Introduction to tinyML – Running Deep Learning Models on Low-Power Micro-Controllers.
About the practice: Our demonstrators for more energy efficiency
Demonstrator I AI supported Motor Control Development / Simulation + Software Permanent Magnet Synchronous Motor (PMSM) AI enhanced Motor control with unbalanced load from MOTEON and Infineon
This demonstrator, developed by MOTEON GmbH and Infineon Technologies and other ECOMAI partners, harness the power of neural networks for control prediction, shifting from reactive to proactive control strategies.
The approach addresses
· dynamic changes in pressure, temperature, and component lifetime,
· including factors like permanent magnetism, coil resistance, and inductivity,
· influencing system performance and efficiency.
By combining AI with traditional motor control algorithms, these variables are carefully considered in the design process, paving the way for more intelligent and responsive motor control systems.
Read here about the development process of the demonstrator
“From Simulation to Silicon”: Research insights for AI assisted Motor Control at the Conference
AI concepts offer sophisticated approaches for embedded applications. But the todays question is also about energy efficiency. In the talk from Steven Klotz at the conference, the question “What is the optimum next current value in order to use as little energy as possible to regulate the speed?”
Demonstrator II Model-based specification of embedded/IoT systems “ECOMOD” ECOMAI Modeling Toolbox
The ECOMOD modelling methodology considers the system-under-design on different levels (System, Technology) and differs strictly between “the problem” (Requirements) and “the solution” (Architecture). It has been developed in the project context of ECOMA by SparxSystems Software GmbH – Central Europe for the specific requirements for embedded systems with AI.
The ECOMOD Framework is based on SysML:
- tailors SYSMOD for the special needs of ECOMAI-specific systems
- supports the modeling process in Enterprise Architect with model patterns and a profile (extending the SysML)
- provides pre-filled project templates for the identified ECOMAI system types (Autonomous Data Collector, Autonomous Control Loop, Actor-controlled System)
ECOMOD is available as free download on GitHub, also offering a practical framework for PdM, Predicitive Maintenance.
Access ECOMOD Documentation and Downlaod on GitHub
Use Case presented at the Conference: Predictive Maintenance with AI for Platform Screen Door Systems (PSD)
How does this approach look in practice? At the conference, the collaboration between Sparx and consortium partner Albayrak Makine Elektronik San. ve Tic. A.Ş. has been presented. PSDs are used in public transportation, but PdM (Predictive Maintenance) is not yet applied. This use case discusses this innovation development
Demonstrator III Motor Control for Medical Rehabilitation Robots: biomedical engineering by neuroConn
The demonstrator by neuroConn is a real-time application with biosignal acquisition, real-time computation, control of actuators within 1-3 ms. It allows a precise operation of actuators for arm and finger rehabilitation based on body state determination and modulation.
Team-Members from ECOMAI joined the booth over these 3 busy days in Nürnberg: Fabiola Bermudez-Elsinger Daniel Mueller-Gritschneder Klaus Schellhorn Steven Klotz Dr. Veit Zöppig Philip Würfel Peter Lieber Salomé Wagner
ChipsForEurope! Chips Joint Undertaking ECOMAI is a #EURIPIDES² project of Xecs_Eureka and AENEAS Association.
We sincerely thank the State Development Corporation of Thuringia – LEG and the team of Katja Wetzel for hosting us and the hospitality at the community booth.
Impressions from the embedded world conference
Thomas Besorna from Sprax Systems Europe presenting ECOMOD
Prof. Dr. Daniel Müller-Gritschneder from TU Vienna presenting TinyML