August 27–31, 2007, Kyoto, Japan | Sachin Katti, Shyamnath Gollakota, and Dina Katabi
This paper introduces Analog Network Coding (ANC), a novel approach to wireless network communication that leverages interference to increase network capacity. Unlike traditional methods that aim to avoid interference, ANC encourages senders to transmit simultaneously, allowing routers to forward the interfering signals without decoding. Receivers can then cancel the interfering signals and recover the intended packets using network-level information. The paper demonstrates that this approach theoretically doubles the capacity of a 2-way relay network and is practical, as shown through experiments with software radios. The implementation achieves significantly higher throughput compared to traditional wireless routing and prior work on wireless network coding. The paper also addresses practical challenges such as signal synchronization and channel distortion, providing a detailed algorithm for non-coherent demodulation of interfered signals. The authors evaluate their design in a testbed, achieving an average bit error rate of 2-4% and a throughput increase of 70% over traditional wireless routing and 30% over digital network coding.This paper introduces Analog Network Coding (ANC), a novel approach to wireless network communication that leverages interference to increase network capacity. Unlike traditional methods that aim to avoid interference, ANC encourages senders to transmit simultaneously, allowing routers to forward the interfering signals without decoding. Receivers can then cancel the interfering signals and recover the intended packets using network-level information. The paper demonstrates that this approach theoretically doubles the capacity of a 2-way relay network and is practical, as shown through experiments with software radios. The implementation achieves significantly higher throughput compared to traditional wireless routing and prior work on wireless network coding. The paper also addresses practical challenges such as signal synchronization and channel distortion, providing a detailed algorithm for non-coherent demodulation of interfered signals. The authors evaluate their design in a testbed, achieving an average bit error rate of 2-4% and a throughput increase of 70% over traditional wireless routing and 30% over digital network coding.