Loading...

Application of Artificial Intelligence for Encoding and Decoding in the Internet of Things Communication

Hassanpour, Mohammad Amin | 2021

224 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 54531 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Farhadi, Alireza
  7. Abstract:
  8. One of the important issues in the field of telecommunications is the optimization of coding algorithms. Especially, in the field of the Internet of Things (IoT) due to the low processing power and the limitation in the energy storage of devices, the needs for such optimized algorithms are more essential. One effective method to improve coding algorithms in the transmission of messages is obtained by paying more attention to the amount of information contained in each part of the message and transmit information accordingly. Some coding methods have addressed this issue indirectly, but they do not pay enough attention to this practical issue. Based on the simple but practical idea mentioned above, we propose two methods that can have a significant impact on the performance of the existing coding algorithms. Such coding algorithms must be able to store or send a message with different lengths of data. Many artificial intelligence-based coding schemes have such structures, which are very relevant to the methods proposed in this dissertation. The performance of these schemes can be significantly increased by combining them with the above proposed practical idea. In order to evaluate the performance of such combined coding schemes, artificial intelligence-based joint source-channel coding schemes are used for the transmission of pictures/frames online. It is observed that the proposed schemes can reduce the transmission rate by 25%. The first proposed scheme is based on a complete simulation of the telecommunication system and the second proposed scheme is based on the predictions obtained from a neural network. The first scheme leads to better results than those of the second scheme due to its higher accuracy, but the second scheme has less computational complexity. By conducting numerous experiments, the desirable performance of the first scheme is observed in various conditions, including different sizes of image patches, changes in compression rate, and changes in the noise power of the telecommunication channel. Also, the two factors of the importance of sending symbols (bits) and the amount of freedom given to this method are observed to be effective on the performance of the first scheme
  9. Keywords:
  10. Internet of Things ; Artificial Intelligence ; Joint Source and Channel Coding ; Picture Transmission ; Decoding Algorithm

 Digital Object List

 Bookmark

No TOC