Stetsenko Inna, I. Sikorsky National Technical University of Ukraine «Kyiv Polytechnic Institute» (37 Pobeda Av., 03056 Kyiv, Ukraine)

Sukhaniuk Maryna , I. Sikorsky National Technical University of Ukraine «Kyiv Polytechnic Institute» (37 Pobeda Av., 03056 Kyiv, Ukraine)

Shyshkin Vladyslav , I. Sikorsky National Technical University of Ukraine «Kyiv Polytechnic Institute» (37 Pobeda Av., 03056 Kyiv, Ukraine)

Language: ukrainian


Relevance of research topic. Nowadays, smart devices are very popular, namely modern devices with built-in information technologies. This article presents the development of a smart DVR that supports the ability to recognize Ukrainian automobile numbers and their further analysis. Such a tool can speed up the search of stolen cars, which is a topical issue.

Target setting. During the development of the hardware-software complex, issues related to its architecture and data processing are solved, namely, which methods and algorithms are used for the transformation, transmission, receiving, consolidation of data and storing in the database. It should be kept in mind that during construction of device, the physical features of the modules and their processing and transmission capabilities, such as processor speed, memory size have to be considered.


Analysis of recent research and publications. The latest technologies in the field of data processing (serialization and deserialization libraries), automobile number recognition algorithms and the database with the possibility of accelerated text search are analyzed.

Uninvestigated parts of general matters defining. The architecture and software modules of smart DVR is proposed for the first time, the problem of data transmission to the cluster under the conditions of high load and intermittent Internet communication is solved; full-text and partial search of automobile numbers in the database is proposed; algorithm of automobile numbers recognition in motion is developed.

The research objective. Identify the required combination of methods and algorithms for the implementation of the smart DVR information system, which ensures the reliable and efficient execution of its functional capabilities.

The statement of basic materials. The article presents the description of modules, which represent the smart DVR, and the purpose of each module is taken into consideration. The scheme of the system as a whole is presented and the algorithm of real-time search of automobile numbers based on the YOLO neural network is developed. The basic principles of communication between servers and devices are considered.

Conclusions. The architecture of a new smart device is proposed and the main algorithms that implement its functionality are developed. Identified problems that may arise in further development and identified ways to resolve them.

Key words:

DVR; smart device; ASUS Tinker Board; GPS; NMEA; GPSD; serialization; YOLO.


1. Statystyka vykradennia avtomobiliv [Car Jacking Statistics]. (n.d.). Retrieved from [in Ukrainian].

2. Chan, J. W. (2018). Learn 5 Single Board Computer: Raspberry Pi, Asus Tinkerer Board, Banana PI M2, Pine A 64, Chip, Rock 64. Amazon Digital Services LLC [in English].

3. El-Rabbany, A. (2006). Introduction to GPS: The Global Positioning System. Boston: Artech House [in English].

4. Kaplan, E., Hegarty, C. J. (2006). Understanding GPS principles and applications. Boston: Artech House [in English].

5. Kleppmann, M. (2010). Designing data-intensive applications: The big ideas behind reliable, scalable, and maintainable systems. Sebastopol, CA: OReilly Media [in English].

6. Patel, S. K. (2013). Instant GSON. Packt Publishing [in English].

7. Welty, E. (2001). The Jackson cookbook. Jackson, MS: The League [in English].

8. Protocol Buffers | Google Developers. (n.d.). Retrieved from [in English].

9. YOLOv3: An Incremental Improvement. (n.d.). Retrieved from [in English].

10. Rother, C. (2016). Computer Vision I - Image Processing. Computer Vision Lab [in English].

11. Redmon, J., Angelova, A. (2014, December 01). Real-Time Grasp Detection Using Convolutional Neural Networks. Retrieved from 2014arXiv1412.3128R. [in English].

12. Smith, R. (2007). An Overview of the Tesseract OCR Engine. IEEE [in English].