Perceptive, Resilient, and Efficient Networks assisted by Reconfigurable Intelligent Surfaces

In recent years, research in industry and academia has been directed towards identifying critical scenarios for future networks and devising the corresponding enablers to address the associated challenges. To take full advantage of the large bandwidth offered by high frequency bands, and to ensure high quality links, it is necessary to efficiently counteract the sources of blockage and attenuation that lead to signal degradation and, therefore, the necessity for new approaches to address such challenges becomes vital [1]. In this context, the capability to flexibly engineer the wavefront of beams that carry the information, taking into consideration environmental parameters and usage scenario characteristics, opens up a new ‘beyond communications’ playground, where communication nodes can be employed to identify the shape of surrounding obstacles, map the environment, detect the presence of objects, focus power, localize, track mobility and navigate [2].
It is gradually becoming common ground that future networks will be programmable, reconfigurable, with environmental awareness that maps the communication landscape, achieving unprecedented quality of service. As a result, they are expected to support a multitude of functionalities and, therefore, to bear multiple qualities; they are expected to be perceptive within the communication environment, resilient to communication outages and power efficient. Specifically, a perceptive network will be able to detect changes in the communication environment, such as the motion of the users or objects that block active links and even predict their motion. A resilient network will be able to restore quickly the connections in cases of outage from power loss or blockage and even avoid blockage. The link restoration will be seamless so as not to affect the user experience. A power efficient network will function with low energy requirements, and with as few corrective measures as possible.
These qualities can be enabled with beam profiles that offer advanced functionalities beyond conventional beam-forming. For example, beams that are able to redistribute the incident power into a focused area, or even bend or self-heal, are ideal candidates. They can be produced either directly from a specially designed emitter, e.g. a large antenna array, or a reflecting surface that is able to transform the incoming beam, such as a reconfigurable intelligent surface (RIS). Such beams usually involve rapidly varying phase profiles and, therefore, to produce them it is required that their wavefront is adequately sampled. As a result, many radiating elements are usually necessary, rendering large antenna arrays that require bulky hardware non-practical and inefficient.
Contrary to large antenna arrays, RISs are nearly passive surfaces, that is, they do not require active power amplifiers, digital signal processing, and multiple RF chains [3]. Owing to its passivity, the RIS may offer hundreds or thousands of elements at relatively low power consumption levels, making it ideal for advanced wavefront engineering operations that require dense wavefront sampling. Besides promising to restore the line-of-sight (LoS) of links interrupted by a blocker and to increase the channel gain via beam forming, the potentially large size of RISs enables a number applications, such as cost-efficient redirection of a beam, focusing the power of a beam to a small area with controllable size, generating beams that can bend to bypass obstacles, enabling them to reach users beyond the LoS of conventional beams. Overall, RISs can adjust the wireless medium, which was previously thought to be uncontrollable, by controlling the scattering, refraction and reflection characteristics of waves [4]. With RISs, the wavefront engineering also becomes programmable [5]. This enables the transmitter to sense the wireless environment and allows for integrated communication and sensing, thus enabling full advantage of the high bandwidth offered by high frequency bands, and promising high quality, perceptive, resilient and efficient links.
So far, communications have been driven by omnidirectional emitters and directional beamformers, while, in recent years, beam-focusing is gaining ground. Additionally, most applications in wireless communications have been designed for the far-field of transmitting antennas. However, with RISs it is now possible to adjust the near- and far-field regions [6], thus enabling new functionalities. In this work. In this paper, the potential of RISs for advanced wavefront engineering is investigated. The main contributions of this paper are as follows.

• •

  The concept of beam-forming, beam-focusing, self-accelerating and self-healing as enablers of perceptive, resilient, and efficient networks is proposed.

• •

  It is demonstrated how beams that are able to focus, bend and self-heal enable advanced functionalities necessary for future networks beyond typical beam-forming, and the design specifications to reach their full potential are discussed.

• •

  A hierarchical localization method based on a hybrid beam-forming/beam-focusing scheme is proposed.

• •

  New approaches to counteract the effect of blockage based on self-accelerating and self-healing beams are proposed.

• •

  Methods for achieving energy efficient networks are demonstrated. The concept of wireless power transfer and how the received power can be maximized with low transmitted power is discussed.

It is well known that the waves radiated by antennas evolve into the far-field roughly at the Fraunhofer distance, $2D^{2}/\lambda$ which is determined by the size of the antenna $D$ relative to the wavelength $\lambda$. For conventional antennas, e.g. phased arrays, the size of the radiating element (or radiating aperture) extends from a fraction of a wavelength up to only a few wavelengths and, therefore, the radiated waves quickly enter the far-field. With RISs, though, this intuitive picture can change. Owing to their large surface, usually extending up to several wavelengths, waves reflected by the RIS may evolve under a new aperture, i.e. a new Fraunhofer distance. For example, a wave propagating in the far-field of the transmitter and impinges on a RIS, illuminating the entire RIS surface, is steered under a new aperture determined by the size of the RIS (this is typically referred to as full illumination). On the other hand, a pencil beam that partially illuminates a large RIS area is steered with a new aperture determined by the size of the beam footprint on the RIS (this is typically referred to as partial illumination). As a result, the transition from the near- to the far-field can be much farther than what is intuitively expected at a certain frequency and can be controlled by the size of the RIS relative to the footprint of the incident beam [6]. This flexibility to adjust the Fraunhofer distance can enable new functionalities, such as focusing and generally advanced beams that require relatively large apertures with tailored phase response.
The RIS transforms the incident beam to a reflected beam with desired characteristics. The profile of the reflected beam is controlled by the phase, $\phi_{RIS}$, imposed by the RIS elements on the incoming wavefront, i.e. the incident field, $E_{i}$, is transformed to the reflected field as $E_{r}=E_{i}\exp(j\phi_{RIS})$. Depending on the functional form of $\phi_{RIS}$, different types of beams can be generated by the RIS [5]. In Fig. 1 four beam profiles are presented, corresponding to beamforming, beam-focusing, self-healing, and self-accelerating operations. These profiles can be advantageous to a number of applications, including but not limited to beam-tracking, blockage avoidance and power transfer.
Beamforming is the most fundamental of the functionalities of RISs [7] and is typically used to steer the incident beam towards the desired direction. This is achieved by imposing linear phase shifts on the RIS elements, i.e. $\phi_{RIS}=k(sin\theta_{i}-sin\theta_{r})x$, where $k$ is the free-space wavenumber, $\theta_{i}$, and $\theta_{r}$ denote the angles of incidence and reflection on the steering plane, respectively, and $x$ is the Cartesian coordinate; in this example the RIS is on the $xy$ plane, steering waves on the $xz$ plane. Beamforming with a RIS can be used to increase the received power at the direction of the user, establish an additional LoS link, or track the user. The size of the RIS affects the reflected beam in two ways. First, the beam-width of the reflected beam depends on the size of the footprint of the incident beam on the RIS or the size of the RIS, whichever is smaller, which in turn depends on the beam-width of the incident beam. As a result, as the footprint of the incident beam increases under partial illumination conditions, the beam-width of the reflected beam changes until, passing onto full illumination conditions, the footprint surpasses the RIS size and the reflected beam converges to a shape determined by the size and the shape of the RIS [6]. Second, the size of the RIS/footprint also affects the Fraunhofer distance. For example, a large footprint increases the Fraunhofer distance and can potentially “move” the user equipment (UE) in the near-field. On the other hand, a small footprint decreases the Fraunhofer distance and can “move” the UE to the far-field. The Fraunhofer distance can be effectively controlled through the gain of the transmitter, which tunes the size of the beam-footprint on the RIS. Controlling the Fraunhofer distance is fundamental as the optimization of beamforming in terms of received power depends on it. As shown in [6], maximizing the received power at a certain distance requires a specific RIS size or footprint that sets the Fraunhofer distance at the RIS-UE distance, i.e. that effectively places the UE at the near-to-far-field transition of the RIS. Moreover, under partial illumination, the beam in its near-field undergoes negligible diffraction, essentially propagating with constant peak power, which is ideal for power transfer applications. [8]
Beam-focusing is the ability of a RIS to focus the power of the reflected beam to a small area, similarly to how a lens focuses light onto a focal point. This can be achieved simply by imposing a parabolic phase profile, $\phi_{RIS}$, on the RIS elements, i.e. of the form $-kr^{2}/2f_{0}$, where $r=\sqrt{x^{2}+y^{2}}$ is the radial coordinate in the RIS $(x,y)$ plane, and $f_{0}$ is the focal distance. Beam-focusing can be used to estimate the location of the UE, to maximize the received power, as well as for wireless power transfer applications. The power is maximized at the focal point, and the locus of the spatial locations with half-maximum (or 3dB) power marks an ellipsoid centered at the focal point and oriented with its major axis along the beam propagation direction. Its size depends on the incident beam footprint, as well as the RIS-focal point distance (focal distance). In principle, the larger the footprint on the RIS, the smaller the size of the 3dB ellipsoid and, consequently, the higher the resolution in localization-related applications and the higher the maximum power at the focal point.
Self-healing beams are a type of diffraction-free beams with the ability to reconstruct themselves when faced with an obstacle [9]. This is possible, because rays from the edge of the beam replace the blocked rays as the beam propagates. Therefore, self-healing beams are ideal for counteracting the effect of blockage. Such beams require a power-law phase profile of the form $-kCr^{\gamma}$, where $\gamma$ and $C$ are design parameters (constants) [10]. Theoretically, although such beams require an infinitely large array to be generated properly [11], they can be efficiently generated with finite apertures, as shown in Fig. 1. The effectiveness, as well as the distance of beam reconstruction depend on the size of the obstacle relative to the RIS aperture. In principle, the larger the RIS, the faster and better the beam recovers.
Self-accelerating beams are also diffraction-free, but unlike self-healing beams, they bend with the propagation distance. They require a power-law phase profile of the form $-kCx^{\gamma}$, where $\gamma$ and $C$ are constants that control the bending radius. Such beams also possess the ability to reconstruct after an obstacle like all diffraction-free beams [10, 9] and, therefore, such beams are also ideal for counteracting the effect of blockage. In principle, the larger the RIS, the more the beam can be engineered to bend. With knowledge of the wireless environment, the location of the UE, and the beam propagation, self-accelerating beams can establish connection between points otherwise inaccessible with conventional beamforming.

To this day, communication networks are designed to operate in the far-field. The need for agility, programmability and recofigurability pushes the operating frequency to higher bands and imposes the need for large antenna arrays and RISs, consequently bringing communications in the near-field. The large area of the RIS and the multitude of scatterers it offers, renders the RIS ideal for advanced wavefront engineering, where the incident beam is transformed into a non-trivial reflected beam that is able to address the challenges of high frequencies more efficiently than conventional beam-forming. In this paper, the potential of advanced wavefront engineering was investigated for RIS-assisted networks, with the purpose of making the networks perceptive to changes in the wireless communication environment, resilient to outages, and energy efficient. Three advanced wavefront engineering beam profiles are demonstrated along with RIS-assisted beam-forming, the beam-focusing that concentrates the power at a small area, the self-healing beam profile that is resilient to obstacles, and the self-accelerating beam profile that can avoid obstacles. The design specifications to reach their full potential are discussed, and some techniques are demonstrated in an effort to make the wireless communications networks perceptive, resilient and efficient.

This work was supported by the European Commission’s Horizon Europe Programme under the Smart Networks and Services Joint Undertaking TERA6G project (Grant Agreement 101096949) and INSTINCT project (Grant Agreement 101139161).