Aleksandar Stevanovic

Prof. Aleksandar Stevanovic

University of Pittsburgh
orcid

Research Focus & Metrics

Academic Career Length

13Years (2014-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 April 2019 in Physics and Chemistry of the Earth, Parts A/B/C
Verified
Journal Article
Closed access
DBDebojit Biswas
Debojit Biswas
HSHongbo Su
Hongbo Su
CWChengyi Wang
Chengyi Wang
Aleksandar StevanovicAleksandar Stevanovic
Aleksandar Stevanovic
WWWeimin Wang
Weimin Wang

1 November 2019 in IEEE Transactions on Intelligent Transportation Systems
Verified
Journal Article
Closed access
NMNikola Mitrovic
Nikola Mitrovic
IDIgor Dakic
Igor Dakic
Aleksandar StevanovicAleksandar Stevanovic
Aleksandar Stevanovic
This paper presents a concept, herein called Combined Alternate-Direction Lane Assignment and Reservation-based Intersection Control (CADLARIC), for organizing directionally unrestricted traffic flows in automated vehicle environment. In the proposed concept, vehicles can use lanes traditionally reserved for the opposite direction of travel. This concept allows left and right turning vehicles to align themselves in an appropriate lane before reaching the downstream intersection, so that they can smoothly go through the intersection without having any conflicts with vehicles from the other movements. Conflicts between through movements are handled by a reservation-based algorithm. To simulate the proposed concept, i.e. allow flexibility of such driving maneuvers that cannot be accomplished in the other existing simulation tools, a new microsimulation platform is developed. The proposed CADLARIC control scheme is evaluated through a comparison with a conventional fixed-time signal control and Fully Reservation-Based Intersection Control (FRIC). The results show that CADLARIC: (i) significantly outperforms other control methods in terms of the traffic performance (i.e. delays and stops); and (ii) generates a reduction of conflicting situations at the network level when compared to FRIC.

30 June 2017 in Sensors
Verified
Journal Article
Open access
PDF file
DBDebojit Biswas
Debojit Biswas
Hongbo SuHongbo Su
Hongbo Su
CWChengyi Wang
Chengyi Wang
JBJason Blankenship
Jason Blankenship
Aleksandar StevanovicAleksandar Stevanovic
Aleksandar Stevanovic
Automatic car counting is an important component in the automated traffic system. Car counting is very important to understand the traffic load and optimize the traffic signals. In this paper, we implemented the Gaussian Background Subtraction Method and OverFeat Framework to count cars. OverFeat Framework is a combination of Convolution Neural Network (CNN) and one machine learning classifier (like Support Vector Machines (SVM) or Logistic Regression). With this study, we showed another possible application area for the OverFeat Framework. The advantages and shortcomings of the Background Subtraction Method and OverFeat Framework were analyzed using six individual traffic videos with different perspectives, such as camera angles, weather conditions and time of the day. In addition, we compared the two algorithms above with manual counting and a commercial software called Placemeter. The OverFeat Framework showed significant potential in the field of car counting with the average accuracy of 96.55% in our experiment.

Verified
Journal Article
Open access
PDF file
DBDebojit Biswas
Debojit Biswas
Hongbo SuHongbo Su
Hongbo Su
Chengyi WangChengyi Wang
Chengyi Wang
Aleksandar StevanovicAleksandar Stevanovic
Aleksandar Stevanovic
Speed detection of a moving object using an optical camera has always been an important subject to study in computer vision. This is one of the key components to address in many application areas, such as transportation systems, military and naval applications, and robotics. In this study, we implemented a speed detection system for multiple moving objects on the ground from a moving platform in the air. A detect-and-track approach is used for primary tracking of the objects. Faster R-CNN (region-based convolutional neural network) is applied to detect the objects, and a discriminative correlation filter with CSRT (channel and spatial reliability tracking) is used for tracking. Feature-based image alignment (FBIA) is done for each frame to get the proper object location. In addition, SSIM (structural similarity index measurement) is performed to check how similar the current frame is with respect to the object detection frame. This measurement is necessary because the platform is moving, and new objects may be captured in a new frame. We achieved a speed accuracy of 96.80% with our framework with respect to the real speed of the objects.

16 October 2020 in Transportation Research Record: Journal of the Transportation Research Board
Verified
Journal Article
Closed access
PDF file
NDNemanja Dobrota
Nemanja Dobrota
Aleksandar StevanovicAleksandar Stevanovic
Aleksandar Stevanovic
NMNikola Mitrovic
Nikola Mitrovic
Adaptive traffic control systems (ATCSs), have been in use since the late 1970s but have only received significant attention in the U.S. during the last 10 years. Since 2008, the number of these systems has increased by more than 600%. A brief literature review clearly shows a shortage of comprehensive studies that gather data from numerous field evaluations, categorize those evaluations based on several factors, summarize findings, and develop meaningful conclusions. This study addresses this gap by presenting a framework and a tool which enable comprehensive analysis of ATCSs deployed and evaluated in the U.S. Compared with some previous studies, where the focus was on brief and general experiences with ATCSs, the tool presented in this study allows detailed analysis of ATCS deployments and investigation of numerous criteria important for their evaluations. Relevant ATCS data from deploying agencies and literature are collected, discussed, and used to populate a database of the Assessment Tool for Adaptive Traffic Control ([AT]2C). The main purpose of (AT)2C is to help practitioners and researchers to identify, compare, and assess statistics of relevant ATCS technologies, mainly from the perspective of their field benefits.

22 September 2021 in Transportation Research Record: Journal of the Transportation Research Board
Verified
Journal Article
Closed access
PDF file
ZCZorica Cvijovic
Zorica Cvijovic
MZMilan Zlatkovic
Milan Zlatkovic
Aleksandar StevanovicAleksandar Stevanovic
Aleksandar Stevanovic
YSYu Song
Yu Song
Connected vehicle (CV) technologies enable safe and interoperable wireless communication among vehicles and the infrastructure with the possibility to run many applications that can improve safety, and enhance mobility. This paper develops CV-based algorithms which use transit vehicle speed and the estimated time that the vehicle needs to arrive at an intersection to trigger transit signal priority (TSP) initiation. This information is updated each second based on the traffic conditions such as speed, a current distance of a transit vehicle to the intersection, and queue conditions. The algorithm uses the actual speed of a transit vehicle and its latitude/longitude (lat/lon) coordinates to compute the time that the vehicle needs to reach the stop line. It was tested on a real-world network using VISSIM traffic simulation, but can easily be implemented in the field, since it is using world coordinates. The upgraded algorithm was applied to a future bus rapid transit (BRT) scenario, and included different levels of conditional TSP, which depend on three combined conditions: the time that a transit vehicle needs to reach the stop line, the number of passengers on board, and the lateness that the transit vehicle experiences. The test-case network used for model building is a corridor consisting of ten signalized intersections along State Street in Salt Lake City, UT. The CV algorithms coupled with TSP can yield notable delay reductions for both the regular bus and the BRT of 33% and 12%, respectively.