Sustainable transportation systems are necessary to accommodate the rising need for mobility while minimizing the possible adverse social, economic, and environmental effects during a time of fast urban development. In this work, autonomous vehicles (AVs) are examined as a potential mode of transportation for sustainable and intelligent development. First, we examine the conventional approaches for object tracking and then the deep learning-based approaches for object tracking. The development of an object-tracking model based on two approaches such as traditional models like model-based, sensor fusion-based, stereo vision-based, grid-based, etc. was then discussed, along with the deep learning-based approaches like online vs offline and detection-based vs detection-free models. To completely test the tracking capability of the suggested algorithm on various objects in difficult scenarios, pertinent experiments based on real driving movies and public datasets were completed. The findings proved that when compared to the existing algorithms, the suggested approach had high tracking speed and accuracy, showed higher robustness and anti-interference skills, and could still be tracking object accurately in real time without the addition of complicated structures. These methods can be used to provide driving assistance for intelligent vehicles and will support the advancement of computer vision technologies in environmental perception.
Multiple Object Tracking of Autonomous Vehicles for Sustainable and Smart Cities
Contributions to Environmental Sciences & Innovative bus. technol.
2023-09-26
19 pages
Article/Chapter (Book)
Electronic Resource
English
Connected and autonomous vehicles in smart cities
TIBKAT | 2021
|A Novel Multiple Object Tracking Algorithm for Autonomous Vehicles
British Library Conference Proceedings | 2020
|A Novel Multiple Object Tracking Algorithm for Autonomous Vehicles
Springer Verlag | 2020
|Unmanned Aerial Vehicles in Sustainable Smart Cities
Springer Verlag | 2023
|