Very Efficient Deep Learning in IoT

Teaching the IoT to learn



As the Internet of Things (IoT) continues to take shape, promising widespread automation and data exchange, one of the biggest challenges is to act on the data generated. The amount of data collected is huge, the computational power required for processing is high, and the algorithms are complex. The EU-funded VEDLIoT project develops an IoT platform that uses deep learning algorithms distributed throughout the IoT continuum. The proposed new platform with innovative IoT architecture is expected to bring significant benefits to a large number of applications, including industrial robots, self-driving cars, and smart homes. The project offers an Open Call at project midterm, incorporating additional VEDLIoT-related industrial use-cases in the project, increasing the market readiness of the VEDLIoT solutions.




Information to be added


Information to be added


Information to be added




The industrial use case employs DL-based solutions for motor condition monitoring and for arc detection, involving different challenges concerning the usage of DL models. In the former case it is necessary to comply with a ultra-low energy budget, and in the latter it is necessary to ensure a very low false-negative error rate.


​The automotive use-case focuses on increasing the processing efficiency DL tasks over the resources that are present in the traffic environment. This will be achieved through distribution of the processing tasks over resources such as the ego vehicle, cellular base station(s) in the close proximity,  as well as the cloud.


In the home domain, VEDLIoT will consider a virtual mirror application in which it will be necessary to deploy the software and execute DNN models on different kinds of hardware, namely at the edge (t.RECS) and on a specialized embedded platform (µ.RECS).


Promotion lecture for Professor by Pedro Trancoso

Promotion lecture for Professor by Pedro Trancoso

The work to be developed in the scope of VEDLIoT was mentioned in a promotion lecture for Professor at the Chalmers University, by Pedro Trancoso. The lecture, entitled "Game of Domains: accelerators are coming!", focused on the design of efficient, scalable, and...

read more
VEDLIoT presentation at HiPEAC

VEDLIoT presentation at HiPEAC

Pedro Trancoso gave a presentation at the 3rd Workshop on Accelerated Machine Learning (AccML), co-located with the HiPEAC 2021 Conference, in which he introduced VEDLIoT.

read more