Name: Embedded Platform for EMG acquisition with Artificial Intelligence software toolkit (EMGP-AI)
Project number: PN-III-P2-2.1-PED-2019-2392
Contract number: 425PED / 2020
Period: 27.10.2020 - 22.10.2022
Research fields: Machine Learning, Biomedical Signal Processing
General Objective
The goal of this project is to develop an intelligent platform (hardware and software) capable of both real-time recording of the electrical activity of the muscles, produced in response to the motor nerve stimulus (Electromyography - EMG),
as well as providing a necessary set of software tools based on Artificial Intelligence (AI) for the development of applications in various fields: medicine, security, entertainment, etc. The platform will be structured in two components -
the EMG signal acquisition module and the software tools for Artificial Intelligence. The EMG signal acquisition module will be a non-invasive acquisition device that can record and monitor EMG activity from the forearm. The architecture of
the device will consist of a network of EMG sensors and a central unit (Microcontroller/Microcomputer) responsible for synchronously recording the signals and communicating with the software tools for Artificial Intelligence. We will create an
EMG dataset corresponding to 15 hand gestures, which we will use for extensive analysis of all the methods and algorithms that we will include in the AI software toolkit. For validation, we will develop an Automatic Gesture Recognition (AGR) system
using the previously described platform. This system will allow the real-time classification of 15 gestures (identical to those in the dataset). More precisely, the system will be responsible for acquiring the EMG signal, processing
it, and finally associating the input signal with the corresponding gesture (the output of the system will be represented by the gesture executed by the user).
Towards this goal, we have set three objectives:
O1. Design and develop an EMG Acquisition Module
O2. Collect an EMG dataset for gesture classification
O3. Design and develop an Automatic Gesture Recognition (AGR) framework
Motivation
In recent times, a significant amount of research has been focused on human-computer interaction (HCI), augmented reality (AR) and virtual reality (VR). Hand gestures represent a very natural and easy way
to interact with the devices around us, taking the IoT experience to a different, more organic level.
In this project we aim to develop an intelligent platform (hardware and software) able to capture in real time electrical activity in response to a nerve's stimulation of the muscle (Electromyography - EMG)
and offer the necessary Artificial Intelligence software tools for developing applications in various domains: healthcare, security, entertainment, etc. The platform will be structured in two components -
EMG Acquisition Module and Artificial Intelligence software tools. The EMG Acquisition Module will be a non-invasive acquisition device that can collect and monitor EMG activity at the forearm level.
The architecture of this device will be composed of a network of several EMG sensors and a central unit (Microcontroller/Microcomputer) responsible for captured signals synchronization and communication
with Artificial Intelligence software tools. We will create an EMG dataset with 15 hand gestures followed by an extended analysis of every method and algorithm included in the Artificial Intelligence software tools.
For validation purposes, we will develop an Automatic Gesture Recognition (AGR) framework using the platform. The AGR will be a complex system that enables real-time labeling of the 15 gestures of the dataset.
More precisely, this framework will incorporate the acquisition of the signal to-be labeled, the pre-processing of the signal, and finally the classification (the system's output is the gesture that the user has performed).
Working Plan
The project is structured in 3 stages, corresponding to the reporting stages. Each stage is based on the results of the previous stages or studies in the project, as follows:
Stage 1: Project planning and hardware component design for the data acquisition module
Stage 2: Development of the acquisition module and creation of EMG resources for the training and testing the gesture classification module
Stage 3: Complete implementation of the automatic gesture recognition system based on biometric data (EMG)
Expected Results
The expected results in each phase of the project are as follows:
Stage 1:
Period: 27.10.2020 - 31.12.2020
Design of the acquisition module
Stage 2:
Period: 01.01.2021 - 22.11.2021
Development of the data acquisition hardware module
Development of the data acquisition software module
Acquisition of the EMG dataset for 15 gestures
Data segmentation software
Stage 3:
Period: 23.11.2021 - 22.10.2022
Study on the state of the art in the field of gesture recognition
The concept and software implementation of the classifier
Automatic gesture recognition system, based on Artificial Intelligence techniques
Dissemination in 2 international conferences, 1 publication at a specialized conference
Final Results
In this project a complete system for classifying EMG signals was created. In the first stage, a device for recording EMG signals
was developed, in the form of a portable bracelet. Using this device, a dataset of electromyographic signals was created, acquired
from 50 subjects, both men and women. Each participant was recorded 2 times while performing a gesture, for a total number of 15 gestures.
In order to create machine learning models to classify gestures based on EMG signals, the acquired dataset was filtered to reduce any possible noise.
Each signal was afterwards segmented in small time-segments, which were used to extract representative features of the EMG signal. Methods to increase
the robustness of the automatic gesture recognition system have also been experimented with.
The final system is capable of a classification accuracy of 98.67% for 15 gestures. Therefore, the system offers high performance and can be included
in smart prostheses that can help people who have lost an upper limb. Examples of the final model are illustrated below:
Romanian Academy Research Institute for Artificial Intelligence “Mihai Drăgănescu” (ICIA) was set up in 1994 as a centre of competence and active dissemination of knowledge in the domain of Artificial Intelligence. ICIA has a core of permanent research positions and a small number of consultant researchers. Besides them, a variable number of contract-based positions, especially students, participate in the R&D activities of ICIA. Every year, since its foundation, ICIA was evaluated as an excellency institution by the Romanian Academy. In 2001 ICIA won the competition for Centres of Excellency (in Information Science and Technology) organized by the Ministry of Education and Science. In 2002 it was rated the best research institute of the Romanian Academy (from 60 institutes and centres). In 2002 ICIA was granted the right to organize doctoral studies in the Romanian Academy system. In 2008 ICIA was accredited as a research institution and as a research and development, member of the research and development system of national interest, in conformity with HG no. 551/2007.
In 1947, following a memorandum addressed to the Ministry of National Education on the establishment of a Polytechnic in Cluj with three faculties: Constructions, Electromechanics and Silviculture, through the provisions of the Law on Education Reform of August 1948, the Institute of Mechanics from Cluj was established, having a faculty with two sections: Thermotechnics and Work Machines. The increase in the need for technical staff made the Institute of Mechanics in 1953 transform into the Polytechnic Institute of Cluj.
After the revolution in 1989, the Romanian higher education returned to the Romanian tradition correlated with the Western system. Starting with 1992, the Polytechnic Institute changed its name to the Technical University of Cluj-Napoca, and from the three faculties that existed at that moment, through restructuring, seven faculties were created: Automation and Computers, Electronics, Telecommunications and Information Technology, Engineering Electrical Engineering, Construction, Machine Building, Mechanics, Materials Science and Engineering (becoming in 2011 Engineering Materials and Environment), as well as the Technical, Economic and Administrative University College. In 1998, the structure of the Technical University of Cluj-Napoca complemented with the Faculty of Architecture and Urbanism and in 2007, with the Faculty of Installations.
Since 2012, the Technical University of Cluj-Napoca has thirteen faculties following the merger with the North University of Baia-Mare, which became the North University Center in Baia Mare, comprising the Faculty of Engineering, the Faculty of Letters, the Faculty of Mineral Resources and Environment and the Faculty of Sciences. At present, the Technical University of Cluj-Napoca prepares specialists through bachelor, master, doctorate and postgraduate studies, with the number of students exceeding 20,000. The fundamental, or applicative scientific research is an essential preoccupation of the teaching staff and researchers at UTCN. The scientific potential of the University has enabled it to organize or be involved in the organization of large-scale scientific events with a large participation of Romanian and foreign specialists.
University POLITEHNICA of Bucharest is the oldest and most prestigious engineer school in Romania. Its traditions are related to the establishment, in 1818, by Gheorghe Lazar, of the first higher technical school with teaching in Romanian language, in the Saint Sava Monastery in Bucharest. In 1832, it was reorganized into the Saint Sava College.
.
On October 1, 1864, the School of Bridges and Roads, Mine and Architecture was established, and on October 30, 1867, it became the School of Bridges, Roads and Mine, with studies duration of 5 years. Under the leadership of Gheorghe Duca, on April 1, 1881, the institution acquires a new structure, under the name of "The National School of Bridges and Roads"; on June 10, 1920, the Politehnica School of Bucharest was founded, with four sections: Electromechanics, Construction, Mining and Metallurgy, Industrial Section.
From November 1920 the name changes to POLITEHNICA of Bucharest.
On August 3, 1948, the Polytechnic Institute of Bucharest was founded, which initially included 4 faculties and in which, since 1950, most of the today faculties have appeared. Based on the resolution of the Senate of November 1992, the Polytechnic Institute of Bucharest became University POLITEHNICA of Bucharest.
At University POLITEHNICA of Bucharest, people are trained without whom the society, as we know it today, could not work - the engineers. We are a nation with inherited technical abilities, with native engineering inclination, and many times Romanian engineers have added their name to the pantheon of world science.
A history of the University of Iaşi takes us through the lessons of the history book of the Romanians from the 16th century. If, in general, the existence of a university is linked to the historical context, in Romania, the University of Iasi has often played a role as a creative historian. In the 17th century, for example, the university was a shelter of Byzantine culture; later it was one of the most important factors in spreading the Romanian language and indigenous culture; after its establishment as the first modern higher education institution of Romania, the university provided the state with the most important thinkers and actors in the public life so that, during communism, the humanist spirit could be kept alive.
