AIoT Research Lab

Academic Projects

---

Design and Development of IoT System Architecture with QoS Support

supported by TUBITAK, 119N049, 2020-2022

This project addresses the end-to-end Quality of Service (QoS) issue in IoT systems. Its main objective is to design a scalable multilayer architecture (with edge/fog and cloud services) to satisfy the QoS requirements of real world IoT applications. To achieve this objective, end-to-end QoS models and a set of novel dynamic resource allocation schemes will be developed for IoT networks.

---

Improving WebRTC Security via Blockchain Technology

supported by ASELSAN-SSB, 2019-2020, Principal Investigator

In this project, we aim to strengthen the security infrastructure provided in the Command Control software during the transmission of real-time audio, video and application data using WebRCT application technology with a blockchain-based approach.

---

Fog Computing Based Complex Event Processing and Content Curation for Internet of Things

supported by TUBITAK, 118E212, 2018-2021, Principal Investigator

In this project, we address the problem of collecting IoT data using a fog (edge) computation based distributed architecture and content curation via complex event processing techniques. By using fog computing concept, the big data analysis problem is handled in layered distributed approach. We perform content curation at the fog layer to extract local valuable information with a minimal latency. Later, in cloud layer, we perform stream data mining algorithms to analyze this data again to achieve global information extraction. Hence, by using distributed fog servers, IoT data requirements such as volume, variety, latency, time and location dependency are satisfied.

---

Reconfigurable Network-on-Chips

supported by TUBITAK, 117E130, 2017-2020, Researcher

Network-on-Chip (NoC) is a communication infrastructure for chips consisting too many processing elements. It has been created as a better alternative to classical bus-based communication method and it inherits most of the computer network concepts. When a NoC architecture is designed, the designer should consider performance, cost, fault-tolerance and energy consumption criteria as the chips have limited resources on them. The NoC topology plays an important role to meet the aforementioned criteria of the final network design. For example, a regular topology (such as mesh) might be preferable due to its scalability, fault-tolerance, and reusability for different applications whereas an irregular topology can be favorable due to its huge optimization space for performance, energy, and cost. If we can merge the benefits of these two different topology types, we can have higher chances to meet the required criteria for the applications. Motivated by this observation, in this project, we aim to add reconfigurability to both mesh-based and irregular topology-based NoC designs.

---

The Effects of Eye-Tracking Based and Parent Mediated Joint Attention Eary Intervention Program on Social Communication Skills of Children with ASD

supported by TUBITAK, 115K459, 2015-2018, Researcher

The main purpose of the proposed project is to examine eye tracking skills of children with ASD and determine atypical face processing patterns observed in young children with ASD. A technology based assessment protocol will be developed to determine eye tracking skills specific to ASD. Data mining techniques will be used to discover the embedded but useful patterns and relations in the data that is collected from children with ASD and typically developing children. The data analysis of the project will be realized in KAVEM Lab.

---

Energy Efficient Connected Dominating Set Based Virtual Backbone Construction in Wireless Sensor Networks via Multi Objective Evolutionary Algorithms

supported by TUBITAK, 113E328, 2014-2016, Principal Investigator.

Wireless sensor networks have become increasingly popular in recent years for controlling systems where human intervention is undesirable or impossible. However, there is no central management or any fixed infrastructure for these networks. Therefore, Virtual Backbone (VB) is currently used to support topology control, efficient routing and broadcast communication in wireless sensor networks. A Connected Dominating Set (CDS) is a promising candidate formulation for constructing a VB stimulated by different characteristics of wireless sensor networks. Although, a minimum sized CDS serves as a less maintenance overhead for virtual backbone, finding such CDS is an NP-hard problem. Although there are several factors (load balancing, coverage, latency, energy efficiency etc.) that affect the VB design to the best of our knowledge, all related works consider heuristic and/or meta-heuristic optimizations for formulating different single objective CDSs in wireless sensor networks. The VB constructed using a single objective CDS can only improve the design objective, the other objectives (constraints) of the wireless sensor network cannot be satisfied by this VB. In this proposal, different from the previous work, VB construction is achieved using a multi objective based CDS. In order to construct the CDS multi objective evolutionary optimization algorithms are used.

---

Development of a Sensor Network and Camera Based Hybrid Critical Area Surveillance System, supported by HAVELSAN A.Ş. and Ministry of Science, Industry and Technology

0211.STZ.2013-1, 2013-2015, Principal Investigator

In this project, a reliable hybrid critical area surveillance system that will mitigate the shortcomings of the camera based and sensor network based surveillance systems will be developed. The proposed system will be composed of two major subsystems. The first subsystem is wireless sensor network that is deployed to the area to be monitored and the second subsystem is the video camera subsystem that can move its cameras based on the information received from the wireless sensor network. The proposed hybrid system’s main tasks can be listed as (i) target detection, (ii) target identification and (iii) target tracking. Target detection task is performed by the wireless sensor network. The location information of the detected target is forwarded to the camera subsystem. Then, the camera subsystem positions one of the cameras to the target and identifies the target. If the identified target something that needs to be tracked, the target will be tracked by both the camera subsystem and the wireless sensor network. The novel aspect of the proposed project is as follows: if the camera subsystem looses the target during the tracking, the wireless sensor network will provide necessary location information of the target so that camera can be re-positioned for target tracking.

---

National Broadband Communication Infrastructure Planning Project

supported by Ministry of Transport, Maritime and Communications, 2012-2013, Principal Investigator

There are more than 50 countries commit to invest more than 100 billion US dollar on the National Broadband Network construction. The “2015 Broadband Targets” announced by the Broadband Commission on October 2011 also inspire some developing countries to start their national broadband initiatives. As a fast growing country in economic development and telecommunicaton services, Turkey is the most prominent country to start the National Broadband Network initiative. In this project, economic and technological models that will promote the Turkish National Broadband Network project are investigated. This project is completed in TUBITAK-ULAKBIM.

---

Energy Efficient Coverage in Wireless Sensor Networks using Genetic Algorithm

supported by Gazi University SRP, 06/2012-49, 2012-2014, Principal Investigator

Wireless Sensor Networks (WSN) are one of the recent hot research areas that attracts researchers. WSNs have a wide range of application areas that include military, chemical, environmental and industrial applications. Due to the physical properties of WSNs to ensure energy efficiency is the most important issue dealt with. For this reason, the method will be developed on the target coverage problem aims to provide energy efficiency at the same time. Sensor nodes are distributed in a specific area to cover the entire area is intended to minimize the energy consumption. By using multi-objective genetic algorithm for both cases the results obtained using the technique will be compared with the other results in the literature.

---

User Profile Analysis in Social Networks Using Data Mining

supported by Gazi University SRP 06/2011-41, 2011-2012, Principal Investigator

Recently, data mining in social networks has received great attention and researchers started to investigate the issues in this area. For example, a well known social network Twitter includes opinions of millions of people that can be a good information source. In this project, we will use a Twitter dataset to classify the opinions in a large stream of data. Data mining techniques will be used for data preprocessing and classification. In particular, we will classify the users’ short text messages into several emotional classes thereby analyzing their profiles. The extracted information can be used for commercial and advertising purposes.

---

In-Network Outlier Detection Based Fault Tolerant Data Aggregation Protocols for Wireless Sensor Networks

supported by Gazi University SRP, 06/2011-09, 2011-2013

Data aggregation protocols are essential for wireless sensor networks to prolong network lifetime by reducing energy consumption of sensor nodes. For mission critical wireless sensor networks, however, not only the energy consumption of sensor nodes but also the correctness of the data aggregation results is critical. As wireless sensor networks are usually deployed in harsh and hostile environments, malfunctioning and/or compromised sensor nodes negatively affect the correctness of the data aggregation results. In this project, we will develop a set of fault tolerant data aggregation protocols that eliminate the false data sent by malfunctioning and compromised sensor nodes. To conserve energy while eliminating false data, an in-network outlier detection technique that is based on Locality Sensitive Hashing (LSH) scheme will be employed.

---

Wireless Sensor Network Based Border Surveillance System for Turkey's Border Security

supported by Gazi University SRP, 06/2008-44, 2008-2010, Principal Investigator

Border security is one of the most important issues of any country due to smuggling, illegal immigration and terrorist threats. Today’s border surveillance and protection systems include border patrol vehicles, video surveillance system, control centers, permanent and mobile observation posts, and therefore incur high deployment and operational costs. Moreover, the recent incidents show that the current border surveillance and protection systems are not able to provide a complete border security. To this end, this project proposes wireless sensor network based real time border surveillance and protection system that has low deployment and operational cost compared to current order surveillance systems. The proposed project includes a novel target detection and tracking algorithm that takes advantage of context information gathered by sensor nodes. Due to employment of context information during target detection the precision of the system is expected to increase which results in less number of “false alarms”.

---

Design of National IPv6 Infrastructure and Transition to IPv6 Protocol

supported by TUBİTAK-KAMAG, Researcher

In this project, a road map for the IPv6 transition process for Turkey is drawn. As a reseacher, security aspects of the IPv6 transition are investigated and reported. More specifically, my research was focused on worm propagation in IPv6 networks.

Industry Projects