Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Slađana Đurašević has a total of 27 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Smartphone Application for Cycling Cadence Estimation Using Wireless IMU Sensor

    by Uroš Pešović, Slađana Đurašević Pešović, Saša Cvetković, Pietro Picerno

    Cycling cadence or pedaling rate is one of the most commonly used parameters for analyzing cycling performance. This paper presents a smartphone-based application which measures the cycling cadence using the wireless internal measurement unit (IMU) sensor. Cadence is measured from the foot acceleration using a CC2650 SensorTag wireless sensor placed onto the rider’s ankle and acquired data is transmitted to a smartphone via Bluetooth Low Energy (BLE). A custom experiment for Phyphox smartphone-based application is developed which provides data acquisition and visualization, peak detection, and cadence estimation. Proposed system provides a simple and low-cost solution for real-time cadence monitoring and storage of acceleration data for offline analysis and is suitable for educational demonstrations, preliminary cycling analysis, and further development of smartphone-based cycling measurement systems.

  • Open Access

    Solid-State Drive Failure Prediction Using Anomaly Detection

    by Vanja Luković, Željko Jovanović, Slađana Đurašević Pešović, Uroš Pešović, Borislav Đorđević

    Solid-State Drives (SSDs) enabled the implementation of real-time cloud services, with a primary focus on high performance and high availability. SSD failure prediction can improve overall system availability by preventing data loss and service interruption. SSDs employ a built-in SMART (Self-Monitoring, Analysis, and Reporting Technology) system to predict failures when certain operating parameters exceed predefined thresholds. Such univariate SMART-based models can predict a limited set of drive failures. Research in SSD failure prediction is focused on multivariate models, which can exploit the complex interactions between SMART attributes that lead to drive failure in order to detect a much larger set of failures. This paper presents an anomaly detection model, based on the Mahalanobis distance measure, which is used for the failure prediction of SSD drives. The model is able to rank the features according to their influence on failure prediction by using a forward feature selection algorithm. The proposed model is tested on a publicly available Alibaba SSD dataset, where the six highest-ranked SMART features were identified. Using this subset of SMART features, our model was able to detect 64% of failures with 81% accuracy while keeping a high precision of 96%.


  • Detection of Broadcast Storms in Local Area Networks

    by Uroš Pešović, Slađana Đurašević Pešović, Biljana Savić, Dušan Marković

    Broadcast storms represent events in the local area network (LAN) which is the result of excessive amounts of packet retransmissions by network switches which drastically reduce network performance and even overload network infrastructure which can result in that network becoming inoperational. Broadcast storms originate from broadcast and multicast packets which are targeted for all or groups of stations in LAN. Since these packets are addressed to multiple destinations, they are usually forwarded by network switches on multiple ports to reach all targeted destinations. This can become a serious problem in institutional LANs in which broadcast storms can create a temporary or permanent overload of the entire network infrastructure. This is most notably observed during the videoconferencing calls which could cause packet loss or connection interruptions. In this paper, we analyzed the influence of broadcast storms on network performance in institutional LAN and proposed a machine learning algorithm for the detection of broadcast storms based on the network traffic data collected by the Packet Capture (PCAP) Wireshark software.


  • Open Access

    Algorithm for optimal path planning of a robotic lawnmower

    by Uroš Pešović, Dušan Marković, Ranko Koprivica, Slađana Đurašević, Biljana Veljković

    Horticultural lawns include areas of land that are covered with sown grass in urban areas and require periodic mowing in order to best preserve their aesthetic and functional purpose. This activity may require significant involvement of human and material resources, especially in large public areas such as sports fields, parks and picnic areas. Thanks to the progress of computer technology, this tiresome routine work is gradually being taken over by robotic mowers that can bring significant savings in human and material resources. In this paper, the problem of robotic lawnmowers related to path planning in the working environment is presented. Commercial solutions to date are mostly based on choosing a random straight line path until the mower reaches a lawn boundary or obstacle, after which it turns in the opposite direction and continues mowing in an arbitrary direction. This paper presents a path planning method based on the Chinese postman's algorithm, which enables the planning of the most optimal path that the mower will take so that it mows the entire lawn in the most economical way.

    • 13 October 2022

    Evaluation of LoRa wireless technology for military applications

    by Uroš Pešović, Sladjana Djurasevic, Mihailo Knežević

  • LORA (LONG-RANGE) TEHNOLOGIJA U PRECIZNOJ POLJOPRIVREDI

    by Dušan Marković, Uroš Pešović, Slađana Đurašević, Mihailo Knežević, Dalibor Tomić, Vladeta Stevović

    Precise agriculture, with the support of modern information and communication technologies, enables significant optimization of the agricultural production process. Thanks to the concept of the Internet of Things (IoT), the management of individual operations during the production process can be achieved remotely. The paper presents LoRa (Long Range) technology that enables wireless communication over long distances with low energy consumption, as well as the methodology of LoRa network design. The use of LoRa technology in the open is very important in order to collect data from agricultural fields that spread out on a larger area.

  • Anomaly Detection Model for Predicting Hard Disk Drive Failures

    by Sladjana M. Djurasevic, Uros M. Pesovic, Borislav S. Djordjevic

    The electromechanical design of the HDD (Hard Disk Drive) renders it more susceptible to failures than other components of the computer system. The failure of HDD leads to permanent data loss, which is typically more expensive than HDD itself. The SMART (Self-Monitoring, Analysis and Reporting Technology) system warns the user if any HDD parameter has exceeded the predefined threshold value needed for safe HDD operation. Machine learning methods take advantage of dependence between multiple SMART parameters in order to make failure prediction more precise. In this paper, we present a failure prediction model based on the anomaly detection method involving an adjustable decision boundary. SMART parameters are ranked by the importance and the 13 most significant ones are used as the initial feature set in our model. In the following stage, we optimized the feature set by removing those that have no major contribution to the anomaly detection model, forming the final feature set comprising seven features only. The proposed anomaly detection model achieved 96.11% failure detection rate on average, with 0% false detection rate in ten random tests. The proposed model predicted more than 80% of failures 24 hours before their actual occurrence, which enables timely data backup.


  • PRAĆENJE AKTIVNOSTI PČELA PRIMENOM RAČUNARSKE VIZIJE

    by Slađana Đuraševic, Uroš Pešovic, Dejan Vujičic, Dušan Markovic, Snežana Tanaskovic, Dalibor Tomic, Vladeta Stevovic

    Computer vision as a tool enables automated processing of visual information and provides the possibility of significant improvement of the agricultural production process. This paper presents the results of the application of the YOLO algorithm for monitoring bees at the entrance to the hive. The applied model achieved a detection accuracy of 92.86% and was implemented on a Raspberry PI 4 computer system. This small computer system can be used for further field testing, where the activity of bees at the entrance to the hive is monitored via video recording.

  • Interference classification for IEEE 802.15.4 networks

    by Uros Pesovic, Sladjana Djurasevic, Vanja Lukovic, Peter Planinsic

    Wireless sensor networks most commonly operate in 2.4 GHz band using the IEEE 802.15.4 standard for wireless transmission between sensor nodes. This frequency band is also used by IEEE 802.11 and IEEE 802.15.1 networks which operate with higher transmitting powers and could cause significant inter-protocol interference, especially in highly urbanized areas. To improve coexistence in 2.4 GHz band, IEEE 802.15.4 networks must to be able to identify and evade these interferences. In this paper, we present practical implementation of interference classification algorithm based on the k-means clustering which could be used by low-cost commercially available IEEE 802.15.4 transceivers.


  • Open Access

    Remote monitoring of beehive activity

    by Uroš Pešović, Dušan Marković, Slađna Đurašević, Siniša Ranđić

    Colonies of the European honeybee, the most important pollinator insects, are subjected to many negative impacts of modern society, in form of pollution, pests, invading species and colony collapse disorder (CCD). The newly emerging technology of the Internet of Things (IoT), enables remote monitoring using wireless sensors inside beehives. In this paper, we present the model of the system for remote temperature monitoring on a various points inside the beehive colony using IoT devices. Measured data is transmitted to a remote server where received data streams are processed in real-time using Complex Event Processing (CEP) which enables detection of critical events and report it to the beekeeper. These data streams are compared with reference temperature patterns using machine learning algorithms, which give computers the ability to learn to detect events without being explicitly programmed. This system can significantly reduce the beekeeper reaction time and increase chances for a beehive colony overcoming certain types of anomalies with human intervention.