Frequent asked Questions

Fluid antenna system (FAS) efers to any software-controllable fluidic, dielectric or conductive structures, including but not limited to liquid-based antennas, pixel-based antennas, and metasurfaces, that can dynamically reconfigure their shape, size, position, length, orientation, and other radiation characteristics. Of particular interest in physical layer design is the reconfiguration capabilities, such as the positional flexibility that FAS can offer. By near-continuously switching the antenna’s position within a predefined area to a position where the channel has desirable conditions, FAS can obtain spatial diversity and in multiuser communications, naturally make interference disappear by exploiting opportunities in fading. 

In the context of FAS, a port means a flexible position that the radiator can reach. When the radiator is operating at a specific position for communications, the port is regarded as an active port (or an activated port). The active port has to be connected to an RF chain for digital signal processing, which may or may not be the same RF chain depending on the considered techniques. In other context, an antenna port usually refers to an antenna with an individual radiator and RF chain. As a result, multiple antenna ports have multiple radiators and RF chains. By contrast, FAS can have much fewer RF chains than ports. In its simplest form, FAS only has one active port requiring only one RF chain but a more advanced FAS design can have several active ports.

No. FAS was initially inspired by Bruce Lee's combat philosophy "Be water, my friend". Thus, the term "fluid" is used to emphasise the flexible nature of the antenna and it does not necessarily mean a real "fluid" material such as liquid metal for antennas. In fact, FAS formally refers to any software-controllable fluidic, conductive or dielectric structure that can change its shape and position to reconfigure the radiation characteristics. Therefore, it can be realised using "non-fluidic" materials such as RF pixel-based switchable antennas, metasurface, or mechanical movable antennas.

Traditional multiple-input multiple-output (MIMO) systems are based on fixed-configuration antennas. FAS on the other hand is an emerging communication system where the configurations such as the positions of the antennas can be flexibly changed. Evidently, the range of reconfigurability in FAS can even include shape and other radiation characteristics, which is not covered in traditional MIMO systems. It is worth pointing out that FAS and MIMO can combine to form a MIMO system with flexible-position antennas. This is referred to as MIMO-FAS, or flexible MIMO.

Antenna Selection is a technique of selecting a subset of antennas in traditional MIMO systems for communications while FAS wishes to select a subset of positions (or ports) for communications. From the signal processing viewpoint, they are indeed similar. However, it is worth noting that traditionally, MIMO is based on fixed-position antennas which are usually placed with at least half wavelength apart. This means that Antenna Selection normally deals with a not-so-large number of independent signals for selection (e.g., say <100). By sharp contrast, FAS is designed to switch the antenna’s position very finely within a given space, typically handling an enormous number of correlated signals for selection (e.g., >100 or even in the thousands). The very fine resolution of the antenna’s position in FAS enables a new way to mitigate interference which is not feasible using Antenna Selection using fixed-position antennas. This technique is referred to as Fluid Antenna Multiple Access (FAMA).

Fluid Antenna Multiple Access (FAMA) is a FAS-aided interference mitigation scheme. The key concept of FAMA is to find an opportunity in the spatial domain for the transmitter and receiver to communicate while the interference suffers from a deep fade. FAMA can be classified into Slow FAMA and Fast FAMA. In Slow FAMA, a FAS receiver finds and switches to the position where the average signal-to-interference plus noise ratio (SINR) is maximised. This is regarded as “Slow” because the antenna position is only switched when the channels change. On the other hand, in Fast FAMA, a FAS receiver finds and switches to the position where the ratio of the instantaneous energy of the desired signal to that of the sum-interference plus noise signal is maximised. Fast FAMA switches the antenna’s position on a symbol-by-symbol basis because the instantaneous energy depends on the data. Existing research on FAMA has been limited to selecting only one port at each FAS receiver. Future work may look at the potential performance gains if more than one ports are selected.

Channel state information (CSI) holds significant importance in fully realising the potential of FAS. Full CSI can be obtained by transmitting pilots at all the ports. Considering that a FAS usually has a large number of ports, it would be too expensive to estimate the full CSI if all the ports are involved in the estimation. Fortunately, in the case of finite scattering channels, as in millimetre-wave or terahertz systems, the channel has a sparse scattering characteristics. It is therefore possible to transmit a small number of pilots and estimate the sparse channel parameters, e.g., angles of arrival, angles of departure, path gains, etc., from a few selected ports, and then the full CSI can be reconstructed based on the geometric model. Even in the case of rich scattering, given the finite space of FAS, there can only be a finite number of possible independent channels, which depends on the numerical rank of the spatial correlation matrix. For FAS with size 0.5𝛌x0.5𝛌, the complexity of estimating the FAS channel is the same as that with 4 independent paths. If the size is increased to 2𝛌x2𝛌, then the complexity will be equivalent to the case with 25 independent channel paths.

No. For movable antennas, there are a few designs that enable FAS to move the radiators in a 2D space. For example, one may consider using liquid metal as a droplet on a 2D space. The 2D position of the droplet can be controlled using the electrowetting-on-dielectric (EWOD) technology. Multiple droplets are also possible. Using this technology, the positions of multiple droplets can be independently controlled and surface wave technology can be used to flexibly connect the droplets (i.e., radiators) to the RF chains.