Computer Sciences and Information Technology
Title : | Design of Index Codes for Relay-based Index Coding Problems |
Area of research : | Computer Sciences and Information Technology |
Focus area : | Information Theory, Network Communication |
Principal Investigator : | Dr. Chinmayananda A, Indian Institute Of Information Technology Dharwad, Karnataka |
Timeline Start Year : | 2023 |
Timeline End Year : | 2025 |
Contact info : | cachinmay4@gmail.com |
Details
Executive Summary : | In the multi-way relaying literature, only cases with full data exchange amongst a set of transceivers is considered. The case of partial data exchange amongst the terminals is not considered at all. Designing Instantly decodable codes is also not explored. The number of transmissions required for partial data exchange can be reduced by deploying intelligent techniques like coding at physical layer and network layer. This area is totally unexplored and requires fresh investigation. Scientific Objectives : To design codes which reduce the number of transmissions required for exchanging of messages amongst the given transceivers. This results in better bandwidth utilization, higher throughput and energy efficiency. Model / hypothesis to be tested : A central relay enables the exchange of desired messages amongst a set of transceivers which can not directly communicate amongst themselves. Relay does coding either at the network layer or physical layer leveraging the knowledge of messages present at various transceivers. Main experiments : First, Instantly decodable index codes are studied as they result in least decoding delay. After the reception of an instantly decodable coded packet, the concerned transceivers can decode requested messages, without waiting for the arrival of other coded packets. Then, the restriction of instant decodability is relaxed and codes are designed with different criteria of optimality. The above mentioned codes are designed at the network layer and bounds on the optimal code length will be provided using graph theory, linear algebra and information theory. Heuristic algorithms will be designed and performance will be compared in terms of different metrics including the number of transmissions or data exchange. Then, coding at physical layer will be considered (known as physical layer network coding), which further reduces the number of transmissions compared to only coding at the network layer. Finally, reinforcement learning based approaches will be explored to solve relay-based index coding problems. Thorough simulations will be done with different number of messages and transceivers. Different solution methodologies explored in this project will be compared with the state of the art algorithms. Significance to the field of research: The proposed research provides better multi-way relaying techniques (compared to the state of art ones) which can be used in 5G mmWave networks, Satellite, Maritime and Dense IoT networks, while solving coverage problem, making the network robust, incurring less deployment cost and providing more energy efficiency. |
Total Budget (INR): | 12,35,340 |
Organizations involved