Hardware Evaluation of Interference Alignment Techniques Under Different Channel State Information Updating Rates

• Published in: IEEE open journal of vehicular technology Vol. 6; pp. 1760 - 1773
• Main Authors: Villalonga, David Alejandro Urquiza, Barrios, Alejandro Lopez, Morales-Cespedes, Maximo, Garcia, M. Julia Fernandez-Getino
• Format: Journal Article
• Language: English
• Published: IEEE 2025
• Subjects: Bit error rate; Channel estimation; Channel state information; Interference alignment (IA); Interference channels; multi-user interference channels; Precoding; Receivers; Signal to noise ratio; Spectral efficiency; Ultra-dense networks; universal software radio peripheral (USRP); Wireless networks
• ISSN: 2644-1330; 2644-1330
• DOI: 10.1109/OJVT.2025.3581878

Wireless networks are evolving to provide high data rates, ultra-low latency, reliable communications, and the connectivity of multiple devices in a reduced area. However, massive densification of networks leads to an increase in interfering signals. In this context, interference alignment (IA) algorithms have been proposed to manage interference while increasing the achievable degrees of freedom. However, the practical implementation of IA algorithms faces several issues such as the lack of perfect channel state information (CSI), network synchronization, or modeling a highly heterogeneous signal-to-interference-plus-noise (SINR) distribution. In this work, we propose an experimental evaluation of IA emulating an interference-limited network but focusing on the user perspective. In contrast to previous works, a hardware testbed with universal software radio peripherals (USRPs) is implemented to model heterogeneous SINR networks. The role of both closed and open loops for providing CSI is evaluated. Then, the impact of CSI updating on the spectral efficiency and also on the bit error rate (BER) is analyzed. Furthermore, precoding techniques such as zero-forcing (ZF) or singular value decomposition (SVD) are also considered for comparison purposes. All the results are based on real measurements providing valuable insights into the performance of IA algorithms in real wireless networks.