Wireless telemetry and deep learning models can significantly improve wireless performance. The following papers discuss these technologies in detail and provide techniques that can help advance wireless to provide superior receiver decoding and enhanced Multi-User MIMO performance.
Deep Learning-Based Receiver Design for 802.11ax Systems
A traditional processing pipeline typically involves a sequence of signal processing-based algorithms, including channel estimation, common phase error correction, sampling rate offset correction, and equalization. In DeepWiPHY, we introduce a deep learning-based alternative that can replace the sequence of operations with a single network. Experimental results show that DeepWiPHY can achieve comparable and sometimes superior decoding compared to conventional WLAN receivers. Read the paper where Cisco employees Dan Wai-tian Tan and Rob Liston introduce the concepts in IEEE Transactions on Wireless Communications.
Advances in Multi-User MIMO performance
New research demonstrates that downlink MU-MIMO performance in a practical network not only depends on the client’s channel but is also influenced by factors that are not captured by conventional models, such as client motion and device type. A proposed data-driven algorithm with low computational complexity determines whether a client should operate in MU mode and the MU-MIMO group for clients in MU mode. Experimental results show throughput improvement of up to 35%. Read the paper where Cisco employees Dan Wai-tian Tan and Rob Liston introduce the concepts in IEEE Transactions on Communications.
Check out our Cisco Networking video channel