Signal Recovery For Massive Carrier Aggregation Via Non-Linear Compressive Sensing
Feibai Zhu,An Liu,Vincent Kin Nang Lau
DOI: https://doi.org/10.1109/ICC.2018.8422123
2018-01-01
Abstract:Due to the demand for higher throughput, there is need for aggregating more carriers to serve one user equipment (UE). Massive carrier aggregation (MCA) may help as it aggregate a large number of potentially non-contiguous carriers spanning a wide bandwidth. However, implementing MCA brings challenges to the design of the receiver and corresponding data recovery algorithms. For example, if we assign a separate receiver chain for each carrier, the number of receiver chains will be large, which imposes a huge cost. If we use a single receiver chain for all non-contiguous carriers, an expensive high rate analog-to-digital converter (ADC) is required to sample the entire span of the carriers. To reduce the cost, we propose a receiver architecture that employs only one receiver chain with a non-uniform ADC whose sampling rate is much smaller than the Nyquist rate, and a low cost power amplifier with small dynamic range. Under such architecture, the received signal suffers from non-linear distortion and interference, and the resulting data recovery is a challenging non-linear compressive sensing problem. We propose an algorithm to jointly mitigate the interference and recover the data, which is proved to have theoretical performance guarantees and verified advantageous over baselines in simulations.