User Tools

Site Tools


Action unknown: copypageplugin__copy
projects:year7:7a.028.tut

7a.028.TUT - Co-Botics - Intelligent Cooperating Robots and Humans - Phase-II

Project - Summary

In systems of collaborative robotics, the success of the decision-making process is often based on the ability to efficiently utilize the data coming from multiple sources. This is due to the fact that various sensors are usually utilized in such environment, resulting in the need of non-trivial combination of different signals (visual, audial, etc.). Such problems are referred to as multimodal or multi-view problems, and the success of the solution generally relies on finding a common representation space for different data modalities. Besides, an important research direction in the context of human-machine interaction lies in the detection of unexpected (anomaly) events. Such problems can be solved by means of one-class classification. Another obstacle comes from the requirement of a fast speed of the developed algorithms. In this project, we proposed several solutions to the problems of anomaly detection and fast analysis of multi-modal data in the context of collaborative robotics environment.

Project - Team

Team Member Role Email Phone Number Academic Site/IAB
Moncef Gabbouj PI moncef.gabbouj@tuni.fi +358 (400) 736613 Tampere University
Alexandros Iosifidis Co-PI ai@ece.au.dk +45 9350 8875 Tampere University
Jenni Raitoharju Co-PI Jenni.raitoharju@tut.fi +358 50 447 8418 Tampere University
Fahad Sohrab Student/Researcher fahad.sohrab@tuni.fi 46 962 9962 Tampere University
Kateryna Chumachenko Student/Assistant Researcher Not available Not available Tampere University
Peter Matthews Project Mentors Not available Not available CA Technologies
Steven Greenspan Project Mentor Not available Not available CA Technologies
Matti Vakkuri Project Mentor matti.vakkuri@haltian.com 358 40 512 6894 Tieto

Project - Novelty of Approach

  • This second year of the project will allow us to visualize this imaginary world in order to better understand the way each robotic unit perceives its environment and lead to better decision-making methodologies.
  • We aim at creating an augmented world representing the objects appearing in the direct environment of the robotic unit and the cooperating person(s).

Project - Deliverables

Deliverables
1 Advanced multi-modal analysis methodologies
2 Decision strategies (Model)
3 Visualizations of combined data identities in latent space
4 Efficient implementation and integration to prototype

Project - Benefits to IAB

We have submitted two papers: “Multimodal Subspace Support Vector Data Description” [9] to “Speed-up and multi-view extensions to Subclass Discriminant Analysis” [10] to Pattern Recognition for this project.

[9] F. Sohrab, J. Raitoharju, A. Iosifidis, and M. Gabbouj, 2019. Multimodal Subspace Support Vector Data Description. arXiv preprint arXiv:1904.07698.

[10] K. Chumachenko, J. Raitoharju, A. Iosifidis, and M. Gabbouj, 2019. Speed-up and multi-view extensions to Subclass Discriminant Analysis. arXiv preprint arXiv:1905.00794.

Project - Presentation Video

Project - Documents

FilenameFilesizeLast modified
7a.028.tut_ip_info_sheet.docx120.2 KiB2019/08/20 11:47
7a.028.tut_final_report.pdf509.8 KiB2019/08/20 09:14
7a.028.tut_confluence_project_page.pdf142.9 KiB2019/08/13 15:07
7a.028.tut_quad_chart_2018_spring_meeting.pptx518.3 KiB2019/08/13 15:07
7a.028.tut_cvdi-mid-year-report_cobotics.pdf311.9 KiB2019/08/13 15:07
7a.028.tut_executive_summary.pdf120.1 KiB2019/08/13 15:07
projects/year7/7a.028.tut.txt · Last modified: 2021/06/02 15:15 by sally.johnson