arXiv subjects

Himani Joshi

Publications and source records attributed to Himani Joshi.

3 recordsLinked to original sources

arXiv2020

Reconfigurable and Intelligent Ultra-Wideband Angular Sensing: Prototype Design and Validation

The emergence of beyond-licensed spectrum sharing in FR1 (0.45-6 GHz) and FR2 (24 - 52 GHz) along with the multi-antenna narrow-beam based directional transmissions demand a wideband spectrum sensing in temporal as well as spatial domains. We referred to it as ultra-wideband angular spectrum sensing (UWAS), and it consists of digitization followed by characterization of the wideband spectrum. In this paper, we design and develop state-of-the-art UWAS prototype using USRPs and LabVIEW NXG for the validation in the real-radio environment. Since 5G is expected to co-exist with LTE, the transmitter generates the multi-directional multi-user wideband traffic via LTE specific single carrier frequency division multiple access (SC-FDMA) approach. At the receiver, the first step of wideband spectrum digitization is accomplished using a novel approach of integrating sparse antenna-array with reconfigurable sub-Nyquist sampling (SNS). The reconfigurable SNS allows the digitization of non-contiguous spectrum via low-rate analog-to-digital converters, but it needs intelligence to choose the frequency bands for digitization. We explore the multi-play multi-armed bandit based learning algorithm to embed intelligence. Compared to previous works, the proposed characterization (frequency band status and direction-of-arrival estimation) approach does not need prior knowledge of received signal distribution. The detailed experimental results for various spectrum statistics, power gains and antenna array arrangements along with lower complexity validate the functional correctness, superiority and feasibility of the proposed UWAS over state-of-the-art approaches.

eess.SP↗

arXiv2020

Novel Deep Learning Framework for Wideband Spectrum Characterization at Sub-Nyquist Rate

Introduction of spectrum-sharing in 5G and subsequent generation networks demand base-station(s) with the capability to characterize the wideband spectrum spanned over licensed, shared and unlicensed non-contiguous frequency bands. Spectrum characterization involves the identification of vacant bands along with center frequency and parameters (energy, modulation, etc.) of occupied bands. Such characterization at Nyquist sampling is area and power-hungry due to the need for high-speed digitization. Though sub-Nyquist sampling (SNS) offers an excellent alternative when the spectrum is sparse, it suffers from poor performance at low signal to noise ratio (SNR) and demands careful design and integration of digital reconstruction, tunable channelizer and characterization algorithms. In this paper, we propose a novel deep-learning framework via a single unified pipeline to accomplish two tasks: 1)~Reconstruct the signal directly from sub-Nyquist samples, and 2)~Wideband spectrum characterization. The proposed approach eliminates the need for complex signal conditioning between reconstruction and characterization and does not need complex tunable channelizers. We extensively compare the performance of our framework for a wide range of modulation schemes, SNR and channel conditions. We show that the proposed framework outperforms existing SNS based approaches and characterization performance approaches to Nyquist sampling-based framework with an increase in SNR. Easy to design and integrate along with a single unified deep learning framework make the proposed architecture a good candidate for reconfigurable platforms.

eess.SP↗

arXiv2018

Throughput Optimized Non-Contiguous Wideband Spectrum Sensing via Online Learning and Sub-Nyquist Sampling

In this paper, we consider non-contiguous wideband spectrum sensing (WSS) for spectrum characterization and allocation in next generation heterogeneous networks. The proposed WSS consists of sub-Nyquist sampling and digital reconstruction to sense multiple non-contiguous frequency bands. Since the throughput (i.e. the number of vacant bands) increases while the probability of successful reconstruction decreases with increase in the number of sensed bands, we develop an online learning algorithm to characterize and select frequency bands based on their spectrum statistics. We guarantee that the proposed algorithm allows sensing of maximum possible number of frequency bands and hence, it is referred to as throughput optimized WSS. We also provide a lower bound on the number of time slots required to characterize spectrum statistics. Simulation and experimental results in the real radio environment show that the performance of the proposed approach converges to that of Myopic approach which has prior knowledge of spectrum statistics.

eess.SP↗