Detection Algorithms for Communication Systems Using Deep Learning

• DOI: 10.48550/arxiv.1705.08044
• Publication Date: 2017-01-01
• Authors: Nariman Farsad; Andrea Goldsmith
• Keywords: Molecular Communication; Transfer of learning

The design and analysis of communication systems typically rely on the development mathematical models that describe underlying channel, which dictates relationship between transmitted received signals. However, in some systems, such as molecular where chemical signals are used for transfer information, it is not possible to accurately model this relationship. In these scenarios, because lack channel models, a completely new approach required. work, we focus one important aspect detection algorithms, demonstrate by borrowing tools from deep learning, train detectors perform well, without any knowledge models. We evaluate algorithms using experimental data collected platform, unknown difficult analytically. show learning significantly better than simple detector was previous works, also did assume channel.