Machine Learning Techniques for Equalizing Nonlinear Distortion (2020)

By Rob Thompson, Comcast; Xiaohua Li, State University of New York at Binghamton

Since the early 2000s, the cable television (CATV) industry has been playing its part in the Artificial Intelligence (AI) community by deploying equalization technology to enable its digital signals to survive varying frequency response conditions within its cable plants. Simon Haykin describes how the perceptron and the adaptive filter using the least mean squares (LMS) algorithm are naturally related.

Equalization has evolved into a powerful tool, enabling the CATV industry to achieve communication efficiencies once thought impossible -- but that story is not quite complete. The limits of equalization may extend beyond the linear frequency response, and cancel the nonlinear responses commonly associated with nodes and other active devices which use power amplifiers (PAs). Achieving nonlinear equalization requires new equalization methods, like receiver post-distorter equalization, where techniques include AI models, such as deep neural networks (DNNs). Furthermore, researchers have been advancing nonlinear distortion cancellation via other methods, including peak-to-average-power ratio (PAPR) reduction, and digital pre-distortion (DPD). These technologies are beginning to show up in newer generation devices, where demands for radio frequency (RF) output power is high, while keeping power consumption low, like the full duplex DOCSIS (FDX) remote PHY device (RPD) nodes. DPD technologies cancel the contribution of the transmitting device only. More aggressive nonlinear distortion cancellation methods may be accomplished by advanced DNN approaches, such as incorporating input features derived from Volterra series models, which has become a popular model for no linear distortion that can be used to describe multiple nonlinearity orders and memory. Then efficiencies across the CATV network could be considered, either by higher node RF output power, or more efficient PA architecture/bias within the node, amplifier, and/or customer premise equipment (CPE). This paper will propose how current CATV equalization systems could be enhanced to cancel severe nonlinear distortion based on some of these novel approaches to nonlinear equalization.

By clicking the "Download Paper" button, you are agreeing to our terms and conditions.

Similar Papers

CATV System Performance -- The New York Experience
By Kenneth L. Foster, New York State Commission on Cable Television
1975
Simplifying Field Operations Using Machine Learning
By Sanjay Dorairaj, Bernard Burg & Nicholas Pinckernell, Comcast Corporation; Chris Bastian, SCTE
2017
Applications of Machine Learning in Cable Access Networks
By Karthik Sundaresan, Nicolas Metts, Greg White, Albert Cabellos-Aparicio, CableLabs
2016
A Machine Learning Pipeline for D3.1 Profile Management
By Maher Harb, Jude Ferreira, Dan Rice, Bryan Santangelo & Rick Spanbauer, Comcast
2019
Machine Learning: The Past, Present and the Future
By Narayan Srinivasa, Intel Corporation
2016
Operationalizing Streaming Telemetry and Machine Learning Model Serving: Customer Experience Automation
By Nick Pinckernell & Scott Rome, Comcast
2020
Optimizing Video Customer Experience with Machine Learning
By Mariela Fiorenzo, Claudio Righetti, María Cecilia Raggio, Fernando Ochoa & Gabriel Carro, Telecom Argentina S.A.
2019
Field Measurements of Nonlinear Distortion in Digital Cable Plants
By Tom Williams and Belal Hamzeh, CableLabs and Ron Hranac, Cisco Systems
2014
Network Capacity and Machine Learning
By Dr. Claudio Righetti, Emilia Gibellini, Florencia De Arca, Carlos Germán Carreño Romano, Mariela Fiorenzo, Gabriel Carro & Fernando Rodrigo Ochoa, Cablevisión S.A.
2017
Detecting Video Piracy with Machine Learning
By Matthew Tooley & Thomas Belford, NCTA – The Internet & Television Assocation
2019
More Results >>