Neuromorphic EMG Processing with Spiking Neural Networks

Informations projet

NEPSpiNN

N° de convention de subvention: 753470

DOI

10.3030/753470

Projet clôturé

Date de signature de la CE 14 Mars 2017

Date de début 1 Septembre 2017

Date de fin 31 Août 2019

Financé au titre de

• EXCELLENT SCIENCE - Marie Skłodowska-Curie Actions

Coût total

€ 175 419,60

Contribution de l’UE

€ 175 419,60

Coordonné par

UNIVERSITAT ZURICH
Switzerland

Periodic Reporting for period 1 - NEPSpiNN (Neuromorphic EMG Processing with Spiking Neural Networks)

Période du rapport: 2017-09-01 au 2019-08-31

Résumé du contexte et des objectifs généraux du projet

The NEPSpiNN project aimed to realize a neuromorphic event-based neural processing system
that can directly interface with a commercial surface selectromyography (EMG) for hand gesture
classification. The proposed neuromorphic event-based processing classified the hand gestures in
a single step through a full custom hardware implementation of a spiking neural network (SNN)
which extracts spatio-temporal information of EMG signals locally in real-time with very low
power consumption.

Travail effectué depuis le début du projet jusqu’à la fin de la période considérée dans le rapport et principaux résultats atteints jusqu’à présent

The general objective of the project was to exploit the low-power and low-latency features of a spiking neural network implemented in a mixed analog/digital neuromorphic chip for hand gesture classification.
Main results:
1) develop a spiking neural network able to process EMG signals
2) collection of 3 datasets (open access, zenodo) of EMG from the forearm
3) development and investigation of spiking recurrent neural network on chip
4) mobile application for sensor fusion (EMG+images) for neuroprosthetic control
5) spiking motor controller

Progrès au-delà de l’état des connaissances et impact potentiel prévu (y compris l’impact socio-économique et les conséquences sociétales plus larges du projet jusqu’à présent)

The project was the first using neuromorphic chip for biomedical applications. It was the first one that was using hardware spiking neural network for EMG classification.
The results of the project have high research and commercial exploitation potentialities. The research area of exploitation was primary neuromorphic engineering and neurorobotics. The project results are a step towards the applications of neuromorphic devices in real-world scenarios, in particular for biomedical applications. This will have a high impact on health and science. The commercial areas of exploitation will be information technology, biomedical applications, prosthetic devices, and real-time computing. During the project development, the researcher already established collaborations with industrial partners that can result in new patents and technology transfer to products that can be introduced in the market.

General project description