Intellectual Merit

Precoding refers to the processing a signal undergoes prior to wireless transmission to counteract the effects of multipath propagation and interference. Conventional precoding methods are designed to completely remove interference at the intended receivers, while counteracting the effects of the propagation channel. However, due to the randomness of the transmitted data and the correlations that exist between the propagation channels for different receivers, at any given time some of the interference can actually be beneficial, and add constructively to the decoding of the transmitted data. Since the transmitter is aware of the propagation channels of the receivers and also the data that is being sent to them, it can predict when such constructive interference will occur. “Symbol-level” or “interference exploitation” precoding refers to methods that take advantage of this constructive interference rather than eliminating it. Since the potential constructive interference changes every time the transmit data changes, a non-linear optimization must be implemented at the same rate the data is being transmitted, which requires a higher complexity than conventional methods that employ linear processing. While this increases the precoder complexity, it can dramatically improve the reliability of the wireless link and reduce the amount of transmit power required to achieve a certain level of performance. For this reason, interference exploitation precoding is well-suited for relatively low-rate communication scenarios that demand high reliability, such as sensors or internet-of-things devices.

This project has focused on several new aspects related to symbol-level precoding. In one of these, it was shown how such precoding methods can enable a secondary network to co-exist with a conventional legacy network on the same spectrum. If the primary legacy network is willing to share its data and propagation channel information with the secondary network, the secondary network can exploit symbol-level precoding to not only co-exist side-by-side with the primary network, but also improve the robustness and resiliency of the primary network at the same time. This win-win cooperation motivates the primary network to share its resources with the secondary network, thus helping to alleviate spectrum congestion, which is a significant problem for wireless systems today. The project demonstrated that the benefits of this spectrum sharing via symbol-level precoding is effective even when the propagation channel information that is shared by the primary network is imperfect due to quantization.

Another new aspect studied under this grant is the impact of external interference on both symbol-level and conventional block-level precoding. In particular, the effect of “circular” versus “non-circular” interference was studied, where “circularity” refers to whether or not some correlation exists between the in-phase and quadrature components of a given signal. Conventional methods have typically only studied the case of circular interference, where no such correlation exists. We showed that this is a reasonable assumption for standard block-level precoding methods, since circular interference is in fact the worst-case scenario for such methods, so focusing on this case provides robustness in cases where the circularity of the interference is unknown. However, the opposite is true for symbol-level precoding. We demonstrated that in fact fully non-circular or 100% correlation is the worst-case scenario for symbol-level precoding, and that the precise phase of this correlation has a significant effect on symbol-level precoding performance. Consequently, we derived a modified symbol-level precoding design targeted for the worst-case non-circularity in order to achieve the most robust performance. It was demonstrated that even with highly non-circular interference, the symbol-level precoding approach still maintains a significant advantage over conventional block-level precoding in terms of reliability and reduced transmit power.

Broader Impacts

Two Ph.D. students funded by the grant have graduated and joined the U.S. high-tech workforce in the area of wireless communications and localization. As mentioned above, symbol-level precoding has considerable potential to address the problem of spectrum congestion that is a critical issue that limits the growth of wireless technologies. Another application studied under this grant is the use of symbol-level precoding for systems that integrate both wireless communications and sensing (e.g., localization or radar). Integrated sensing and communication (ISAC) systems will be key drivers in emerging 6G wireless systems. They find applications in vehicular systems, health and well-being monitoring, localization, internet-of-things sensing, smart homes, security, gesture recognition, and other short range and high-resolution sensing tasks. Moreover, there is an urgent need for ISAC technologies in sub-6GHz frequencies such as the 3.5 GHz band, where radars and many commercial communication systems have to share and operate in a congested spectrum. ISAC is also a significant driver of RF convergence where the same transceiver and antenna structures are used for a variety of sensing and communication tasks. Radio systems will be able to perform multiple functions simultaneously with the same device and operate in a shared spectrum. The science developed under this project will thus address a broad collection of use cases across multiple domains.

Last Modified: 01/06/2025
Modified by: Arnold L Swindlehurst