Özlem Durmaz Incel

Özlem Durmaz Incel

Associate Professor, Pervasive Systems Research Group, University of Twente

Coordinated Projects

  • On-device and Distributed Machine Learning with Resource-Constrained Mobile and Wearable Devices for Sensing Applications, (principal investigator (PI), funded by Bogazici University Research Fund (2022-ongoing): Although running deep learning algorithms is challenging due to resource constraints on mobile and wearable edge devices, they improve performance compared to lightweight or shallow architectures. This project aims to explore efficient on-device deep learning and distributed learning for mobile and wearable devices, particularly from the sensor data analytics perspective. Check our review paper about this project: On-Device Deep Learning for Mobile and Wearable Sensing Applications: A Review

  • DAKOTA: Behavioral Pattern Based Authentication, PI, funded by the Scientific and Research Council of Turkey (Tubitak), 1505 Program (Industry-University Collaboration programme) (2018-2020): DAKOTA proposes a behavioral biometrics-based continuous authentication system, which is mainly designed for a mobile banking application. The DAKOTA system records data from the touch screen and the motion sensors on the phone to monitor and model the user’s behavioral patterns. We collected a dataset and applied machine learning and deep learning models for the authentication process. For further information, check our paper: DAKOTA: Sensor and Touch Screen-Based Continuous Authentication on a Mobile Banking Application

  • Efficient and Dynamic Resource Management in Mobile Devices, PI, funded by the Scientific and Research Council of Turkey Tubitak, 1002 Program (2018-2019) We proposed a dynamic parameter selection algorithm, which is called Conawact, for on-device human activity recognition. We tested its performance on the Android platform regarding resource usage such as battery, CPU, and memory. We have published the findings of the study in Context-aware and dynamically adaptable activity recognition with smart watches: A case study on smoking

  • Efficient Resource Management in Mobile and Wearable Devices, PI, funded by Galatasaray University Research Fund (2017-2020) Our aim in this project was to analyze the performance recognition of an activity recognition system in terms of recognition accuracy and resource utilization.

  • Context Recognition from Mobile Devices, PI, funded by Galatasaray University Research Fund (2015-2017) The project’s objective was to semantically classify places visited by smartphone users utilizing the data collected from sensors and wireless interfaces available on the phones as well as phone usage patterns, such as battery level, and time-related information, with machine learning algorithms. Check out our paper: Semantic place prediction from crowd-sensed mobile phone data

  • Activity-Based Crowd-sourced/Participatory Sensing, PI, funded by the Scientific and Research Council of Turkey Tubitak, 2013-2016 (National Young Researchers Career Development Program (3501-CAREER)) One of the objectives of this project was to design and develop an activity based crowdsourced sensing platform where the activities of the individuals related to movement, are recognized with the sensors on the phones and the findings are transmitted to the backend servers where the behaviors of the crowds are analyzed. the second objective of the project was to create a large-scale dataset which will be shared with other researchers working in the domain and hence constitute as a benchmark platform. Check the summary here: Arservice: a smartphone based crowd-sourced data collection and activity recognition framework

Participated Projects

  • SARAS: Sensor-Based Augmented Reality Application, funded by the Turkish Ministry of Science, Industry and Technology under the SAN-TEZ program, 2014-2015 (as a researcher).
  • Green Dynamic Base Station Planning with Power Adaptation for Wireless Cellular Networks, funded by Turk Telekom, 2012-2013 (as a researcher).
  • Fall Detection Using Wearable Acceloremeters, funded by Turk Telekom, 2011-2012 (as a researcher).
  • Intelligent Home Gateway, funded by Tubitak Teydeb and co-funded by Ericsson Turkey, 2011-2012 (as a researcher)
  • COST Action IC0906 - WiNeMO - Wireless Networking for Moving Objects, funded by COST- European Cooperation in Science and Technology, 2011-2012 (as a researcher).
  • FireSense - Fire and Smoke Detection through a WSN and Estimation of Fire Propagation for the Protection of Cultural Heritage Sites, funded by European Commission FP7, 2009-2012 (as a researcher).
  • SmartSurroundings, funding by Dutch Ministry of Economic Affairs, 2005-2009 (as a research assistant)
  • WiSeNt: Wireless Sensor Networks: Omnipresent Embedded Systems for Exploration and Control, funded by European Commission FP6, 2004-2005 (as a research assistant).
  • Service Discovery in Ad Hoc Networks, funded by the Scientific and Research Council of Turkey (Tubitak), 2002-2004 (as a research assistant).