Abstract
Short-and long-range wireless communication is imperative for many remote monitoring applications such as industrial automation, digital health, and smart cities. Remote monitoring is one of the components of complex systems that can be optimized with the use of machine-learning-based techniques. In this article, an intelligent parking system is presented. The system exploits the benefits of a synergy between licensed and unlicensed wireless technologies, IoT, computer vision, and artificial intelligence. We discuss how an end user can interact with the system and manage it with a user interface. The user interface provides a range of dashboard facilities that provide various statistics about car park utilization and optimization facilities. Our proposed solution comprises a range of sensors including high-resolution cameras, which are distributed throughout a car park. The cameras are connected using different types of wireless communication. The data is processed in real time by the sensors and sent to a cloud server, where it is analyzed, stored, and presented to the end user. The object detection, tracking, and optical character recognition system are executed by the vision sensor, which utilizes a combination of convolutional neural networks and a proprietary deep-learning-based object detection algorithm. This solution has achieved high levels of classification accuracy (99.97 percent) and high image processing speeds (66 ms).
Original language | English |
---|---|
Article number | 8782872 |
Pages (from-to) | 23 - 29 |
Number of pages | 7 |
Journal | IEEE Network Magazine |
Volume | 33 |
Issue number | 4 |
DOIs | |
Publication status | Published - 31 Jul 2019 |
Keywords
- Smart parking
- computational intelligence
- Deep learning
- unlicensed spectrum