Enhance the electric motor drive systems with an embedded AI system running on a specialised AI hardware platform: this is the heart of the PENTA project ECOMAI.
Our research activities are focusing on the efficiency and lifetime of electric motors which are typically embedded within larger systems such as used in transportation or medical. The reduction of energy consumption through improved performance of leads to more ‘ecological’ systems.
The intersection of embedded systems and machine learning as it is happening today with AI plays an important role in this context. The so-called Tiny Machine Learning (TinyML) is still an emerging field. It brings machine learning inference to edge devices. To run the inference of a neural network, energy efficiency is crucial.
A recent scientific paper from our consortium member TU Munich has researched around this issue:
“Towards Rapid Exploration of Heterogeneous TinyML Systems using Virtual Platforms and TVM’s UMA”
Workshop on Compilers, Deployment, and Tooling for Edge AI (CODAI), September 21st 2023
Authors: Samira Ahmadifarsani, Rafael Stahl, Philipp van Kempen, Daniel Mueller-Gritschneder, Ulf Schlichtmann, TU Munich
How can TinyML solutions benchmarked to estimate the final performance? Virtual prototyping is key to estimate and explore the performance, enabling the exploration and simulation of new hardware architectures as well as profiling and debugging of software stack.
Online Paper Presentation – Watch here to the presentation
This research paper has been presented at CODAI 20203 (Workshop on Compilers, Deployment, and Tooling for Edge AI) in September 2023.