Key Results

This is the architecture of an RF-WPT system, which comprises i) energy transmitters (ETs); ii) wireless channels; and iii) energy receivers (ERs). 

On the ET side, the charging signal is generated in the digital domain and converted to analog by the digital-to-analog converter (DAC). This is followed by frequency upconversion to the RF domain, signal amplification supported by a high-power amplifier (HPA), and transmission over the wireless medium by one or multiple antennas. On the ER side, the impinging charging signal, distorted by the channel, is captured by the receive antenna(s) and converted to DC via a rectifying network, all composing the so-called rectenna(s). A matching network between the antenna(s) and the rectifying circuitry is needed to maximize the power transfer at a designed frequency band. The rectified signal goes through a power management unit (PMU), which may include a DC-to-DC converter and energy storage, interfacing with the energy demands of the device.

Charging based on RF-WPT is slower and less efficient than traditional wired and reactive wireless charging technologies. This is due to the many sources of energy consumption and loss in the system.

The main energy consumption sources are the baseband processing, including the digital signal generation and DAC controller, the carrier generation, and the HPA DC supply at the ET, while the signal processing and energy management may dominate on the ER side. Additional energy may be consumed by the antenna array controlling circuitry.

The main energy losses come from filtering processes, channel impairments, and the receive antenna aperture. Filtering may be applied i) after the DAC, to remove high-frequency components; ii) after upconversion, to remove parasite frequency components; iii) after the HPA, to remove frequency components introduced by the HPA non-linearity; iv) in the matching network, since frequency components outside its operation bandwidth are removed or attenuated; and v) after the rectifier, to remove the fundamental and harmonic frequencies of the rectified signal. Meanwhile, the channel impairments encompass path loss, fading, shadowing, and nonline-of-sight conditions, and cause the greatest efficiency loss in typical systems. Other energy losses may come from i) elements insertion in the case of not-integrated circuits; ii) transmit-receive polarization mismatches; iii) energy leakage in the energy storage element; and iv) energy dissipation at resistive elements, switching, conduction, reverse-recovery, and parasitic capacitance/inductance.

The efficiency of each block in the RF-WPT system has a multiplicative effect on the end-to-end PTE. The performance of the blocks may exhibit some dependence, so optimizing them independently is not optimal. For instance:

• DACs consume less energy as the required resolution decreases. However, lower DAC resolutions trigger the generation of high-frequency components that must be filtered out, leading to energy loss;

• In multi-tone charging waveforms, the more subcarriers, the greater the flexibility to optimize the waveform to drive the HPA and rectenna to high-efficiency regimes, but the greater the signal processing energy consumption;

• Reducing the number of transmit RF chains is appealing for reducing the transmit energy consumption, but might also limit the EBF gains of the transmit antenna array and the capability to charge multiple users simultaneously.

These fundamental trade-offs must be properly addressed at the design and operation phase. Regarding the latter, an efficient high-power RF-WPT system must operate in a closed loop. At least a reverse or forward communication link is needed to acquire channel state information (CSI) and combat channel impairments. More evolved optimizations require forward and reverse communication links supporting a control plane.

RF transmissions are non-ionizing, thus unable to directly alter the structure of molecules, which would lead to carcinogenic effects. However, this does not discard potential indirect links. The International Agency for Research on Cancer (IARC) classified RF emissions as "possibly carcinogenic to humans". This classification does not indicate that RF emissions definitively cause cancer, but that there is (limited) evidence suggesting a possible link. Nevertheless, it is well known that mild biological effects such as tissue heating may occur when exposed to very strong EMFs at microwave frequencies or higher.

EMF exposure limits have been set by regulatory bodies, such as the United States Federal Communications Commission (FCC) and Food & Drug Administration (FDA), and the International Telecommunication Union (ITU), to mitigate the potential adverse effects on human health. The regulations are intended to balance the potential risks and benefits of using RF technologies and are regularly reviewed and updated. There are also standard bodies, such as the American National Standards Institute (ANSI), the Institute of Electrical and Electronics Engineers (IEEE), and the International Commission on NonIonizing Radiation Protection (ICNIRP), which can not legally enforce their recommended limits or protocols but whose specifications are usually adopted by regulators.

EMF exposure metrics such as specific absorption rate (SAR, in W/kg), specific absorption (SA, in kJ/kg), absorbed power density (APD, in W/m2), and absorbed energy density (AED, in kJ/m2) correspond to physical quantities that are closely related to RF-induced biological effects. SAR and SA are valid for systems operating below 6 GHz, where EMFs penetrate deep into the tissue, while APD and AED are more useful above 6 GHz where penetration depth is less relevant. However, they may not be easily measured/estimated in practice. Other related quantities that are more easily evaluated are the incident power density (in W/m2), incident energy density (kJ/m2), electric field strength (in V/m), and magnetic field strength (in A/m). EMF exposure limits are usually conservatively set in terms of them such that SAR, SA, APD, and AED values remain below the safety limits with extremely high reliability.

High-power RF charging may be possible only for short ranges as the channel losses increase significantly with the distance. In short-range scenarios, precise (highly focused) spatial charging is needed to avoid increased EMF exposure levels in the surroundings of the ERs. This may be only achieved by operating at relatively large frequencies and in the massive MIMO regime, but also in the near-field region of the ETs.  

The above table compiles the EMF exposure limits for the general public according to ICNIRP. The EMF limit values decrease slowly with the operation frequency above 6 GHz in both incident power and energy densities. Moreover, the incident power density limit is far more stringent for local exposure zones than for the whole body, for which the averaging time is also greater.

The key components and vision for a competitive indoor WPT cyber-physical system for two key scenarios are illustrated in the above figure. A proper holistic design of these components, although extremely challenging, is critical to unleashing the full potential of RF-WPT and turning it into a sustainable wireless charging technology, bringing the attention of the market, propitiating more standardization attempts and commercial products.

Massive MIMO deployments can compensate for the extremely large channel losses, promote near-field conditions, and support ubiquitous energy access. However, the required high manufacturing and operating expenses, as well as the increased power consumption, make the development and deployment of truly large-scale antenna arrays challenging, motivating the research on more affordable and low-power MIMO architectures, potentially powered by ambient EH sources, that can scale with the number of antennas more sustainably.

Hybrid analog-digital MIMO implementations require a reduced number of power-consuming RF chains, while further power reductions can be achieved by using low-resolution DACs. In the analog processing domain, one can rely on antenna selection architectures, parasitic, load-modulated, and lens arrays. Metasurface-aided architectures, as reflective intelligent surfaces (RIS) and large intelligent surface (LIS), constitute another key research front to support low-power MIMO. RISs can efficiently assist already-deployed MIMO networks by on-the-fly and opportunistic reconfiguration” of the wireless medium, directing the power towards the intended ER(s), by means of dynamically adapting phase, amplitude, frequency, and polarization of transmit signals. Conventional RISs are composed of fully passive reflecting elements and the only active power consumption comes from the RIS controller, allowing low-power and energy-efficient implementations that do not incorporate additional RF radiation into the environment. However, the potential performance gains may be limited in practice if the number of reflective elements is not sufficiently large. This has motivated the research on active RIS alternatives, where the reflecting elements incorporate reflection-type amplifiers implemented without high-cost and power-hungry RF chains. Such active circuitry allows simultaneously altering the phase and amplitude to enhance the signal power at the receiver. Nevertheless, they incur modestly higher power consumption and require more evolved CSI acquisition procedures. Identifying passive vs active RISs performance trade-offs in the context of RF-WPT, where noise power is not an issue, while including overhead and protocol issues, are open research directions.

Meanwhile, LISs are low-power/cost massive MIMO metasurfaces. They are equipped with RF circuits and signal processing units and composed of a virtually infinite number of elements to form a spatially continuous transceiver aperture. LISs can be realized by using an RIS illuminated by a collocated digital beamforming-based transmitter. The reconfigurability of the RIS provides extra degrees of freedom in this implementation, allowing it to scale the number of passive elements without requiring a larger number of active antennas. Meanwhile, DMA, another LIS implementation, is comprised of multiple waveguides, e.g. microstrip, and each may embed a large set of radiating metamaterial elements whose frequency response can be externally and individually adjusted. One or several RF chains can feed the waveguides, while the input signals are radiated by all the metamaterial elements

The above figure illustrates the power consumption of RIS and DMA based ETs as a function of frequency for a single-ER setup and ET form factor. Herein, we also plot the resulting power density at 15 cm from the ER. Observe that both architectures exhibit a similar performance, especially as the frequency increases. Still, a RIS-based ET seems preferable at relatively low frequency, although this architecture may be more costly than DMA. Nevertheless, all this ultimately depends on the scenario geometry and the number of ERs, which are simplified herein and require further investigation. Interestingly, the performance of RIS-based ETs with only two control bits per element is already close to the optimal, obtained from assuming infinite-resolution RIS. Meanwhile, it is observed that given a fixed form factor and charging distance, increasing the operation frequency is appealing for satisfying the EMF radiation exposure constraints. Indeed, ICNIRP regulations are met herein for frequencies above 7.5 GHz in both RIS and DMA based setups. A critical LIS optimization challenge is encountered as the frequency increases, since the number of variables, i.e., the phase shifts configuration of the antenna elements, becomes massive, for which conventional optimization solvers are inefficient. Two aggravating factors are that: i) in the case of RIS, the phase shifts are discrete; ii) in the case of DMA, more antenna elements can be equipped given a certain form factor and they have a Lorentzian-form response. Based on the results illustrated here and after efficiently dealing with the previous challenges, one can speculate that end-to-end PTEs in the order of 10% may be realized when using greater form-factor LIS based ETs.

The most investigated metasurfaces are those for passive beamforming/reflection applications. Other metamaterial functions such as refraction, absorption, collimation, polarization change, splitting, and analog processing can be software-engineered using accurate physics-compliant models to promote more flexible and dynamic metasurface-based WPT implementations. Moreover, one would ideally distribute the WPT antennas and metasurfaces over the entire charging area, what is not aesthetically or economically attractive in most cases. It is more appealing favoring imperceptible installation via hybrid deployments of distributed and centralized antenna arrays, e.g., using radio stripes, together with scenariospecific metasurface assemblage. The use of hardware-flexible ETs such as motor-equipped and flying ETs may be attractive.