FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3004537488> ?p ?o ?g. }

• W3004537488 abstract "The capability of having human and robots cooperating together has increased the interest in the control of robotic devices by means of physiological human signals. In order to achieve this goal it is crucial to be able to catch the human intention of movement and to translate it in a coherent robot action. Up to now, the classical approach when considering physiological signals, and in particular EMG signals, is to focus on the specific subject performing the task since the great complexity of these signals.This thesis aims to expand the state of the art by proposing a general subject-independent framework, able to extract the common constraints of human movement by looking at several demonstration by many different subjects. The variability introduced in the system by multiple demonstrations from many different subjects allows the construction of a robust model of human movement, able to face small variations and signal deterioration. Furthermore, the obtained framework could be used by any subject with no need for long training sessions. The signals undergo to an accurate preprocessing phase, in order to remove noise and artefacts. Following this procedure, we are able to extract significant information to be used in online processes. The human movement can be estimated by using well-established statistical methods in Robot Programming by Demonstration applications, in particular the input can be modelled by using a Gaussian Mixture Model (GMM). The performed movement can be continuously estimated with a Gaussian Mixture Regression (GMR) technique, or it can be identified among a set of possible movements with a Gaussian Mixture Classification (GMC) approach. We improved the results by incorporating some previous information in the model, in order to enriching the knowledge of the system. In particular we considered the hierarchical information provided by a quantitative taxonomy of hand grasps. Thus, we developed the first quantitative taxonomy of hand grasps considering both muscular and kinematic information from 40 subjects. The results proved the feasibility of a subject-independent framework, even by considering physiological signals, like EMG, from a wide number of participants.The proposed solution has been used in two different kinds of applications: (I) for the control of prosthesis devices, and (II) in an Industry 4.0 facility, in order to allow human and robot to work alongside or to cooperate. Indeed, a crucial aspect for making human and robots working together is their mutual knowledge and anticipation of other’s task, and physiological signals are capable to provide a signal even before the movement is started. In this thesis we proposed also an application of Robot Programming by Demonstration in a real industrial facility, in order to optimize the production of electric motor coils. The task was part of the European Robotic Challenge (EuRoC), and the goal was divided in phases of increasing complexity. This solution exploits Machine Learning algorithms, like GMM, and the robustness was assured by considering demonstration of the task from many subjects. We have been able to apply an advanced research topic to a real factory, achieving promising results." @default.
• W3004537488 created "2020-02-14" @default.
• W3004537488 creator A5075092636 @default.
• W3004537488 date "2018-11-30" @default.
• W3004537488 modified "2023-09-27" @default.
• W3004537488 title "Subject-Independent Frameworks for Robotic Devices: Applying Robot Learning to EMG Signals" @default.
• W3004537488 hasPublicationYear "2018" @default.
• W3004537488 type Work @default.
• W3004537488 sameAs 3004537488 @default.
• W3004537488 citedByCount "0" @default.
• W3004537488 crossrefType "journal-article" @default.
• W3004537488 hasAuthorship W3004537488A5075092636 @default.
• W3004537488 hasConcept C115961682 @default.
• W3004537488 hasConcept C119857082 @default.
• W3004537488 hasConcept C120665830 @default.
• W3004537488 hasConcept C121332964 @default.
• W3004537488 hasConcept C127413603 @default.
• W3004537488 hasConcept C145460709 @default.
• W3004537488 hasConcept C154945302 @default.
• W3004537488 hasConcept C177264268 @default.
• W3004537488 hasConcept C192209626 @default.
• W3004537488 hasConcept C199360897 @default.
• W3004537488 hasConcept C201995342 @default.
• W3004537488 hasConcept C2779038628 @default.
• W3004537488 hasConcept C2779843651 @default.
• W3004537488 hasConcept C2780451532 @default.
• W3004537488 hasConcept C34413123 @default.
• W3004537488 hasConcept C34736171 @default.
• W3004537488 hasConcept C41008148 @default.
• W3004537488 hasConcept C61224824 @default.
• W3004537488 hasConcept C90509273 @default.
• W3004537488 hasConcept C99498987 @default.
• W3004537488 hasConceptScore W3004537488C115961682 @default.
• W3004537488 hasConceptScore W3004537488C119857082 @default.
• W3004537488 hasConceptScore W3004537488C120665830 @default.
• W3004537488 hasConceptScore W3004537488C121332964 @default.
• W3004537488 hasConceptScore W3004537488C127413603 @default.
• W3004537488 hasConceptScore W3004537488C145460709 @default.
• W3004537488 hasConceptScore W3004537488C154945302 @default.
• W3004537488 hasConceptScore W3004537488C177264268 @default.
• W3004537488 hasConceptScore W3004537488C192209626 @default.
• W3004537488 hasConceptScore W3004537488C199360897 @default.
• W3004537488 hasConceptScore W3004537488C201995342 @default.
• W3004537488 hasConceptScore W3004537488C2779038628 @default.
• W3004537488 hasConceptScore W3004537488C2779843651 @default.
• W3004537488 hasConceptScore W3004537488C2780451532 @default.
• W3004537488 hasConceptScore W3004537488C34413123 @default.
• W3004537488 hasConceptScore W3004537488C34736171 @default.
• W3004537488 hasConceptScore W3004537488C41008148 @default.
• W3004537488 hasConceptScore W3004537488C61224824 @default.
• W3004537488 hasConceptScore W3004537488C90509273 @default.
• W3004537488 hasConceptScore W3004537488C99498987 @default.
• W3004537488 hasLocation W30045374881 @default.
• W3004537488 hasOpenAccess W3004537488 @default.
• W3004537488 hasPrimaryLocation W30045374881 @default.
• W3004537488 hasRelatedWork W128262646 @default.
• W3004537488 hasRelatedWork W1505492451 @default.
• W3004537488 hasRelatedWork W2049844716 @default.
• W3004537488 hasRelatedWork W2069324174 @default.
• W3004537488 hasRelatedWork W2284097119 @default.
• W3004537488 hasRelatedWork W2771549791 @default.
• W3004537488 hasRelatedWork W2784033568 @default.
• W3004537488 hasRelatedWork W2796176989 @default.
• W3004537488 hasRelatedWork W2808875021 @default.
• W3004537488 hasRelatedWork W2886296216 @default.
• W3004537488 hasRelatedWork W2887234572 @default.
• W3004537488 hasRelatedWork W2889168904 @default.
• W3004537488 hasRelatedWork W2908543926 @default.
• W3004537488 hasRelatedWork W2954908482 @default.
• W3004537488 hasRelatedWork W2955818852 @default.
• W3004537488 hasRelatedWork W2982252571 @default.
• W3004537488 hasRelatedWork W3007715492 @default.
• W3004537488 hasRelatedWork W3010650220 @default.
• W3004537488 hasRelatedWork W3120428840 @default.
• W3004537488 hasRelatedWork W3101614029 @default.
• W3004537488 isParatext "false" @default.
• W3004537488 isRetracted "false" @default.
• W3004537488 magId "3004537488" @default.
• W3004537488 workType "article" @default.