< Back to previous page


EmbedML: Accelerate the integration of MachineLearning in products with Embedded Hardware (EmbedML)


Making products smarter with digital technology is one of the crucial digital transformations of companies in the manufacturing industry, and machine learning (ML) is a key technology in this respect. Thanks to recent evolutions in development platforms and embedded hardware, ML can now also be applied in the microcontrollers found in a wide range of products for various applications. Although embedded ML is coming within reach of the broader industry, the target group of companies developing smart products still needs to make a big innovation leap to integrate this technology into their product development in the right way and at the right cost.

Overall goal

EmbedML develops a set of supporting tools that guide SMEs in making their products smarter with embedded machine learning. This allows them to realize products with higher added value and accelerate their development. The tools bring insight into the problem-solution fit of their intended solution, how to quickly develop proof-of-concepts (PoCs) and what the required development steps are from concept to supporting products with end users.

EmbedML targets about 150 companies (85% SMEs) that develop smart products and are driven as early adopters to integrate new digital technology into products. Firstly on builders of smart sensor and monitoring solutions, and companies that provide development services for product companies in the form of electronic modules and/or software. Secondly, also on product builders who want to add smart features to their system and previously outsourced this to service providers.

Concrete goals

To help accelerate the integration of machine learning into products of the target group, EmbedML is pursuing the following project deliverables and KPIs:

·         Online casebook of industrial applications of embedded ML that helps SMEs identify and evaluate opportunities.

·         Tools to support companies for the in-house development of PoCs, providing on the one hand design rules for feasibility evaluation and on the other hand practical methods for the optimal use of existing development platforms.

·         Integrated development approaches and guidelines that clarify the requirements, techniques and expertise needed from initial feasibility to product follow-up with end users.

·         Industrial example cases that validate and support the development approach and guidelines. For at least 8 cases (4 in Flanders) the complete concept will be worked out and for 4 cases (2 in Flanders) PoCs will be built.

Date:1 Jul 2021 →  Today
Keywords:machine learning, AI, embedded
Disciplines:Machine learning and decision making, Embedded and real-time systems
Project type:Service project