Intelligent reflective surfaces in 5G and beyond: optimizing uplink satellite connectivity for IoT

The proposed architecture

Within the context of the architectures developed by previous scholars, notable emphasis has been placed on the downlink transmission aspect, encompassing signal coverage and transmission capabilities. However, it is imperative to direct attention toward the equally significant uplink transmission. This entails a comprehensive analysis of aspects such as signal coverage, interference levels, and other pertinent metrics. While the downlink transmission is undoubtedly a critical component, the uplink transmission holds equal weight in shaping the overall communication landscape. By delving into the uplink dimension, we open doors to a more holistic understanding of the architecture’s capabilities and limitations. Analyzing factors like signal coverage ensures that the proposed architecture is conducive not only for transmitting signals but also for receiving them effectively; because due to the substantial surge in the quantity of users and IoT devices, there is a growing need for inventive communication solutions.

Furthermore, satellite communication offers a unique and valuable approach to address the requirements of widespread connectivity. Unlike traditional terrestrial communication infrastructure, satellites can cover vast geographic areas, including remote and underserved regions, making them an ideal solution for achieving global connectivity. In remote areas where laying down conventional communication infrastructure is impractical or cost-prohibitive, satellite communication can bridge the gap effectively. This is particularly crucial for supporting critical applications such as emergency services, disaster response, and healthcare delivery in areas where reliable communication is paramount. Satellite networks can play a pivotal role in managing and transmitting the data generated by these devices, enabling industries to leverage IoT capabilities even in remote or mobile environments. In conjunction with the progress made in 5G technology, integrating satellite communication can lead to a comprehensive communication framework that offers both high-speed urban connectivity and wide-reaching coverage through satellites. This synergy can pave the way for innovative services and applications that were previously unattainable. Further reviewing the integration of IoT devices for satellite transmission, coupled with the assistance of IRS technology in the uplink, the architectural framework comprises several key components. At its core, this integration seeks to capitalize on the capabilities of IoT devices, which generate and transmit diverse streams of data, and leverage the efficiency of satellite communication while incorporating innovative IRS technology to enhance transmission in the uplink phase. The resulting architecture is a synergy of cutting-edge technologies that promise enhanced connectivity and communication efficiency.

In the IoT-satellite communication landscape, IRS integration is achieved through the strategic placement of reflective surfaces that manipulate electromagnetic signals. The placement strategy we adopted involves positioning the IRS panels on both the satellite (SAT-side) and the ground station (GN-side) to ensure comprehensive coverage and improved signal reflection. The height and orientation of these surfaces have been optimized to enhance satellite communication, considering line-of-sight (LOS) requirements for efficient signal phase manipulation. The reflection matrix is dynamically adjusted based on the IRS’s feedback loop, aligning with the satellite communication paradigm.

In this study, we adopted the SAT-side IRS placement approach. The IRS panels are mounted on low Earth orbit (LEO) satellites, as this configuration facilitates a more effective reflection path to enhance uplink communication. The GN-side IRS approach is under consideration for future work, particularly in scenarios where additional ground-based relays are necessary.

Simulink section

The utilization of the MATLAB tool facilitates the translation of the theoretical concepts of the system model into a practical simulation setup. MATLAB feature “Simulink” which is a block diagram environment for multidomain simulation and Model-Based Design, was adopted. It supports system-level design, simulation, automatic code generation, and continuous testing and verification of embedded systems (MathWorks, 2023b). This section primarily dealt with the creation of a visual representation of the system model and the simulation of its dynamics.

Smart IoT element block

From the system model, the smart IoT device acts as a signal generator that communicates with the satellite by signal exchange. To properly represent this on MATLAB tool, we introduce an object called a “signal generator” block (from the communication toolbox) which generates various waveforms. Figure 3 shows the parameters used for the IoT block.

Signal generator properties are listed below:

• i.

  Waveform: Specifies the waveform to be generated. Options include sine, square, sawtooth, and random. For this simulation, the waveform was set to be ‘sine’.

• ii.

  Time: Specify whether to use ‘simulation time’ as the source of values for the time variable, or an ‘external source’ (MathWorks, 2023b). Time was set to ‘simulation time’.

• iii.

  Amplitude: specify the amplitude of the output sine wave signal. A default value of ‘1’ was used.

• iv.

  Frequency: specifies the frequency of the generated waveform. This is set to ‘0.3’.

• v.

  Units: specifies the signal units as Hertz or rad/sec. This is set to ‘Hertz’, indicating cycles per second (MathWorks, 2023a).

MATLAB code section

The MATLAB code section entailed scripting in the MATLAB programming language to implement various simulation-related functionalities. This section encompassed the implementation of algorithms, performance metrics calculations, and comparative analyses. To bolster the foundation of our work, we built upon the simulation code provided by Ardavan (2023) regarding IRS implementation on IoT devices. This code, tailored for IRS integration within IoT, served as a cornerstone upon which we expanded our simulation framework. The IRS optimization is handled through a reinforcement learning algorithm. This algorithm continuously updates the phase shifts of each IRS element to maximize signal reception at the satellite. The phase shifts introduced by the IRS elements are typically computed based on the geometric relationship between the transmitter (e.g., IoT device, satellite), the IRS, and the receiver. To optimize the reflected signals, the phase shifts for each IRS element are designed to ensure constructive combination at the receiver. A closed-form analytical expression was employed to determine these phase shifts, as it provides a simplified and efficient method for calculation. This approach considers the distances between the IoT devices, the IRS, and the satellite, thereby maximizing the performance metric effectively.

The phase shift introduced by each IRS element θ_n (where n represents the nth element of the IRS) can be expressed as:

$${\theta }_n=-(2\pi /\lambda )(d_{\text{IRS-to-Device}}+d_{\text{IRS-to-Receiver}})+{\psi }_n$$where:

• λ is the wavelength of the signal.

• d_{IRS-to-Device} is the distance between the transmitting device and the nth IRS element.

• d_{IRS-to-Receiver} is the distance between the nth IRS element and the receiver.

• ψ_n is an additional phase shift that can be applied for optimizing the system’s performance (e.g., maximizing received signal strength).

Since different devices will have varying distances to the IRS and different angles of incidence/reflection, the location of the IoT devices are considered. For each IoT device, the distances d_{IRS-to-Device} and angles of incidence (from the device to each IRS element) are calculated. The goal is to ensure that the reflected signals from the IRS align constructively at the receiver. The algorithm is based on real-time feedback from the received signal quality, dynamically adjusting reflection coefficients to enhance beamforming and reduce interference. This dynamic adjustment is essential to ensure that the IRS continuously optimizes the reflected signals towards the intended receiver, adapting to changing conditions such as device movement, environmental changes, or signal variations. By incorporating Ardavan’s insights, we gained valuable insights into the intricacies of implementing IRS technology in the IoT context.

Ardavan (2023) defined several functions that served as a base for the simulation framework as listed below:

• i.

  Function to define parameters for Transmitters and IRS locations:

• params.Tx = [0, 0, 0]; // x, y, z coordinates of Tx (transmitter)

• params.IRS_location = [10, 10, 0]; // x, y, z coordinates of IRS

• ii.

  Frequency and wavelength definition:

• params.f = 2.4e9; // frequency in Hz (2.4 GHz is a common frequency for Wi-Fi)

• params.lambda = physconst(‘LightSpeed’)/params.f; // wavelength in meters (calculating based on the speed of light)

• iii.

  Functions to defined IRS parameters:

• params.N = 8; // number of elements in the IRS

• params.R = ones(1,params.N); // initial reflection coefficients of the IRS elements

• params.delta = zeros(1,params.N); // initial phase shifts of the IRS elements

• params.R_max = 0.8; // maximum reflection coefficient magnitude

• iv.

  Functions to define IoT area and devices.

• params.IoT_area = 20; // side length of the IoT area in meters (square area)

• params.num_IoT_devices = 10; // number of IoT devices

• v.

  Function to define antenna signal and channel parameter, noise power

• params.path_loss_exponent = 2; // path loss exponent (affects the rate of signal attenuation with distance)

• params.Pt = 1; // transmitted power in Watts

• params.Gtx = 1; // transmitter gain (dimensionless)

• params.Grx = 1; // receiver gain (dimensionless)

• params.ht = 1.5; // transmitter antenna height in meters

• params.hr = 1.5; // receiver antenna height in meters

• params.Prx_ref = 1e-3; // reference received power at 1 meter

• params.noise_power = -100; // noise power in dBm

• vi.

  Function to Randomly distributed IoT devices within the IoT area

• IoT_positions = params.IoT_area * rand(params.num_IoT_devices, 2)

• vii.

  Function to calculate Two-Ray Ground Reflection Model path loss function (a function that calculates path loss based on distance, heights, and wavelength)

• Two_ray_ground_reflection = @(d, ht, hr, lambda) (ht * hr)^2 / (d^4 * (1 - exp(-1i * 2 * pi * d / lambda))^2)

• viii.

  Function to calculate the signal gain due to the reflection from the IRS

• Gamma = R_opt .* exp(1j * delta_opt) .* exp(1j * 2 * pi * (d / params.lambda) * cos(theta_i) * (0:params.N-1)); // reflection coefficient with phase shift

• g = abs(1 + Gamma).^2

• ix.

  Function to calculate the path loss using the Two-Ray Ground Reflection Model

• pl = 10 * log10(1 ./ abs(two_ray_ground_reflection(d, params.ht, params.hr, params.lambda)))

• x.

  Function to calculate the total received signal power

• Ptot(idx) = params.Pt * params.Gtx * params.Grx * params.Prx_ref * 10^(-pl / 10) * sum(g)

• xi.

  Function to calculate the interference power

• I_interf(idx) = calculate_interference_power(IoT_positions, idx, params.path_loss_exponent, params.Pt, params.Gtx, params.Grx)

• xii.

  Function to calculate SNR for each IoT device

• SNR = 10 * log10(Ptot ./ (10^(params.noise_power / 10) + I_interf));

To further broaden Ardavan’s code, several functions and satellite parameters were introduced and defined to facilitate our simulation. These functions and parameters were defined in detail below:

Function to define satellite parameters:

% Satellite properties

satellite_params.Latitude = 35; // latitude of satellite (deg)

satellite_params.Longitude = −40; // longitude of satellite (deg)

satellite_params.Altitude = 2000; // Altitude of satellite (Km)

% Satellite transmitter properties

satellite_params.TxFeederLoss = 2; // Satellite transmission feeder loss (dB)

satellite_params.OtherTxLosses = 1; // Satellite transmission other loss (dB)

satellite_params.TxHPAPower = 17; // Satellite high-power amplifier (HPA) transmit power (dBW)

satellite_params.TxHPAOBO = 6; // Satellite output back-off (OBO) of a high-power amplifier (dB)

satellite_params.TxAntennaGain = 38; // Satellite transmit antenna Gain (dB)

% Satellite receiver properties

satellite_params.InterferenceLoss = 2; // Satellite Interference loss (dB)

satellite_params.RxGByT = 25; //ratio of receiver gain to system noise temperature (dB/K)

satellite_params.RxFeederLoss = 1; // Satellite receiver feeder loss (dB)

satellite_params.OtherRxLosses = 1; // Satellite receiver other loss (dB)

% Satellite link properties

satellite_params.Frequency = 14; // Frequency (GHz)

satellite_params.Bandwidth = 6; // Bandwidth (MHz)

satellite_params.BitRate = 10; // Bitrate (Mbps)

satellite_params.RequiredEbByNo = 10; // energy per bit to noise power spectral density ratio (dB)

satellite_params.PolarizationMismatch = 45; // polarization mismatch (deg)

satellite_params.ImplementationLoss = 2; // implementation loss dB

satellite_params.AntennaMispointingLoss = 1; // antenna mispointing loss (dB)

satellite_params.RadomeLoss = 1; // radome loss dB

% Signal and channel parameters

satellite_params.path_loss_exponent = 2; % path loss exponent (affects the rate of signal attenuation with distance)

satellite_params.Pt = 1; % transmitted power in Watts

satellite_params.Gtx = 1; % transmitter gain (dimensionless)

satellite_params.Grx = 1; % receiver gain (dimensionless)

% Antenna heights

satellite_params.ht = 1.5; % transmitter antenna height in meters

satellite_params.hr = 1.5; % receiver antenna height in meters

satellite_params.Prx_ref = 1e⁻³; % reference received power at 1 m

The choice of N = 8 elements was made to simplify the simulation and reduce computational complexity. We acknowledge that larger IRS panels, such as 30 × 30-element arrays, are typically used in practice. Future work will extend this model to larger IRS panels to assess scalability and performance improvements more comprehensively. The simulation considers three IRSs to evaluate the performance of a multi-IRS system, offering a broader understanding of IRS-assisted communication in complex scenarios. Covering an area of 1,000 m², the inclusion of multiple strategically placed IRSs ensures adequate coverage and performance. This approach mirrors real-world scenarios where deploying multiple IRSs is both feasible and practical, especially in urban settings or large indoor environments. Simulating multiple IRSs aims to represent realistic deployment strategies while assessing the benefits of inter-IRS coordination and resource optimization. This configuration ensures the simulation results are comprehensive and applicable to scenarios requiring multiple IRSs for reliable and efficient communication over large coverage areas. In the initial phase of simulations, the Gain Block was set to one to facilitate simplified testing. However, for the final simulations, realistic values were applied based on standard satellite communication models, including a Tx gain of 38 dB and an Rx gain of 25 dB, with path loss calculations adjusted to reflect actual signal attenuation over distance.

The communication frequency between the IRS and satellite in the simulation is set to 14 GHz, typical for satellite communications, particularly in the Ku-band.

SNR (improves by increment)

The plot below in Fig. 8 highlights the scattered distribution of SNR across different IoT devices while the plot in Fig. 9 shows the aggregated SNR values of the IoT devices. It shows the SNR performance ‘With IRS (right scatter plot)’ and ‘Without IRS (left scatter plot)’. More devices record improved average values from 24 dB to 26 dB after IRS Integration. This increase in SNR signifies a substantial boost in the quality and reliability of the transmitted signal. The introduction of the IRS has played a pivotal role in achieving this advancement as it directly translates to better signal strength and reduced noise interference.

Channel capacity (improves incrementally)

The plot below in Fig. 10 highlights the scattered distribution of channel capacity across different IoT devices while the plot in Fig. 11 shows the aggregated channel capacity values of the IoT devices. Improved average values from 3.6 bps/Hz to 3.65 bps/Hz after IRS integration were recorded. The depicted plots vividly present the performance variation in terms of channel capacity, drawing a clear distinction between scenarios involving the presence and absence of IRS. Such a transformation bears great significance, as it inherently signifies a notable enhancement in the data transmission rate that can be sustained within the given channel.

Interference (improves by decrement)

The plot below in Fig. 12 highlights the scattered distribution of interference levels across different IoT devices while the plot in Fig. 13 shows the aggregated interference values of the IoT devices. The plot visually portrays the performance disparity concerning interference, drawing a clear contrast between scenarios involving the presence and absence of the IRS. The initial level of interference, quantified at 0.66, exhibits a discernible decrement to 0.24 after the integration of IRS. This notable reduction in interference level stands as a testament to the efficacy of integrating IRS technology. The inclusion of the IRS has contributed to a reduction in interference, thereby fostering a more conducive environment for effective signal transmission.

Pathloss (improves by decrement)

The plot below in Fig. 14 highlights the scattered distribution of pathloss across different IoT devices while the plot in Fig. 15 shows the aggregated pathloss values of the IoT devices. The plots provide a visual representation of the improvement of pathloss performance in scenarios both “With IRS” and “Without IRS by reducing from 55 dB to 52 dB”. Notably, the primary focus of this illustration is the quantifiable difference in path loss resulting from the integration of IRS technology. The application of IRS technology has led to an attenuation in the loss of signal strength as it traverses the communication channel. This outcome shows the efficiency of the IRS in effectively manipulating the propagation environment, thus resulting in improved signal propagation characteristics.

Signal quality (improves incrementally)

The plot below in Fig. 16 highlights the scattered distribution of Signal quality across different IoT devices while the plot in Fig. 17 shows the aggregated Signal quality values of the IoT devices, With IRS and Without IRS. The simulation shows an improved signal quality from 0.8 to 2.0 after IRS integration. This improvement in signal quality bears testament to the efficacy of IRS technology in augmenting the dependability and reliability of the transmitted signal. Such improvements are a direct result of the IRS’s capacity to actively manipulate the propagation conditions, thus mitigating detrimental effects and distortions that can compromise signal integrity.

Coverage area (improves by increment)

The plots in Figs. 18 and 19 visually present a comparative analysis of coverage area performance as a scatter distribution and as an averaged value respectively, distinguishing between scenarios involving the presence and absence of IRS. It conveys, in a concise graphical format, the quantitative shift in coverage area attributable to the utilization of IRS technology. Specifically, the plot shows a substantial expansion in coverage area facilitated by the integration of the IRS. The numerical depiction is striking: the coverage area experiences a significant rise, soaring from an initial expanse of 446 m² to a notably larger 889 m². This empirical evidence underscores the potency of the IRS in augmenting the reach and extent of the covered area within the given communication milieu.

Signal strength (improves by increment)

The plot below in Figs. 20 and 21 shows the signal strength performance distribution and aggregated bar plot respectively, ‘with IRS’ and ‘without IRS’. Specifically, the plot showcases a notable enhancement in signal strength facilitated by the incorporation of IRS. The signal strength demonstrates a substantial increase, surging from an initial intensity of 0.9 mW to a notably higher 1.82 mW. This empirical data effectively emphasizes the efficacy of IRS in bolstering the magnitude of signal strength within the given communication environment.

Figure 21 illustrates the distribution of signal strength across IoT devices in two scenarios: without IRS (left panel) and with IRS (right panel). The signal strength is markedly improved in the presence of IRS, with more IoT devices exhibiting higher signal power, as indicated by the warmer colors (yellow and orange) in the right panel. The right plot shows an overall shift in the distribution of IoT devices towards higher signal strengths, suggesting that the IRS enhances the received power by improving the signal propagation path. Specifically, devices within the reflective field of the IRS experience significant signal boosts, demonstrating the system’s efficiency in signal transmission and reception. This improvement is reflected in the wider range of signal strengths (up to 1.8 units with IRS), compared to the maximum of 0.9 units without IRS.

Spectral efficiency (improves by increment)

The plots in Figs. 22 and 23 show the IoT Spectral distribution and aggregated bar plot respectively, ‘with IRS’ and ‘without IRS’. With IRS, the spectral efficiency shows an increase from an initial value of 1.67 bits per second per Hertz (bps/Hz) to 1.8 bps/Hz. This data-driven observation robustly emphasizes the efficacy of IRS in enhancing spectral efficiency within the uplink satellite communication for IoT.

Figure 22 shows the spectral efficiency of the communication system both without IRS (left panel) and with IRS (right panel). With IRS, there is a noticeable increase in spectral efficiency across the IoT devices, as seen by the broader range of higher spectral efficiency values (up to 1.8 bps/Hz) compared to the scenario without IRS (maxing out at 1.67 bps/Hz). The IRS-enhanced setup allows for more efficient utilization of available bandwidth, which leads to a higher bits-per-second-per-Hertz ratio. This improvement is vital in bandwidth-constrained environments, where efficient use of spectrum is crucial for maintaining high data throughput. The devices situated closer to the IRS also show more pronounced spectral efficiency gains, underlining the role of IRS in optimizing communication for IoT networks.

Link budget (improves by increment)

Figures 24 and 25 provide a comprehensive illustration of the link budget’s status through a scatter plot and aggregated bar plots, respectively. These visual representations distinctly capture the conditions both before and following the integration of IRS technology. A noticeable shift is evident in the scatter plot depicted in Fig. 26, showcasing a clear before-and-after comparison. The accompanying aggregated bar plots in Fig. 27 further consolidate this comparison. Notably, the numerical data points toward a significant increase in the link budget value, specifically ascending from an initial 105 dB to a revised 116 dB. This alteration serves as a testament to the demonstrable advantages brought about by the integration of IRS technology.

Throughput (improves by increment)

Figures 26 and 27 provide a visual representation of the throughput performance, showcasing a scatter plot and aggregated bar plots, respectively. These graphs illustrate a clear comparison between the conditions before and after the integration of IRS technology. In Fig. 26, the scatter plot vividly demonstrates the contrast between these two scenarios. Moving to Fig. 27, the aggregated bar plots offer a consolidated view showing a remarkable surge in throughput performance from an initial 1.1 bits per second (bps) to a notably improved 8.3 bps. This illustrates the fact that the integration of IRS technology holds the potential to significantly amplify the speed of data transmission.

BER (improves by decrement)

Figure 28 presents a scatter plot depicting the BER performance, comparing scenarios both “With IRS” and “Without IRS”. Additionally, Fig. 29 showcases a corresponding bar plot for a comprehensive view of the data. Both plots illustrate the evident enhancement brought about by IRS integration. Specifically, the BER exhibits a substantial improvement, decreasing notably from an initial value of 4.3 to an impressive 0.38 when IRS technology is integrated. This outcome underscores the significant reduction in error noise rate achieved through the incorporation of IRS which further supports the notion that IRS technology holds the potential to considerably enhance the quality and accuracy of data transmission.