Improving QoS for streaming data transmission over 6G networks using reconfigurable intelligent surfaces (RIS)

Abstract

This paper investigates a RIS‑assisted 6G framework operating at 300 GHz (THz sub‑band) to enhance end‑to‑end Quality of Service (QoS) for streaming data under ultra-reliable low-latency communication (URLLC) constraints, focusing on reliability, latency, and throughput across line-of-sight (LoS) and non-line-of-sight (NLoS) scenarios. By programmatically shaping the propagation environment, RIS elevates NLoS Signal-to-Noise Ratio (SNR) from approximately − 20 dB to above 45 dB, reduces Bit Error Rate (BER) from error‑prone levels to below 5.5 × 10⁻¹⁰, and shortens packet delay from roughly 16.5ms to near 2ms, while increasing throughput from ~ 1 Gbps to ~ 20 Gbps under matched assumptions, thereby meeting stringent URLLC targets for reliability and latency. The modeling consolidates noise power, path loss, RIS array gain, capacity, BER, and an SNR‑dependent latency relation with clearly stated assumptions and citations and introduces coding‑aware BER and an improved 300 GHz channel model including atmospheric absorption for realism. Robustness is demonstrated via multi‑run statistics on SNR, BER, delay, and throughput, and a benchmarking subsection contrasts RIS (this work) with CF‑mMIMO and relay baselines under matched or normalized scenarios to support comparative verification claims.

Introduction

The sixth generation (6G) of wireless communication networks is envisioned to revolutionize global connectivity by offering terabit-per-second data rates, ultra- URLLC, and ubiquitous coverage for a wide range of applications, including holographic communications, extended reality (XR), autonomous systems, and industrial automation. Unlike 5G, which primarily operates below 100 GHz, 6G will extensively utilize the terahertz (THz) frequency bands (100 GHz – 1 THz) to support the required bandwidth. However, these higher frequencies suffer from severe propagation challenges, such as increased path loss, high sensitivity to blockages, and limited NLoS coverage. These impairments pose significant obstacles to achieving the stringent QoS demands of emerging applications^1,2,3.

Fig. 1

Enhancing 6G with RIS technology.

To address these challenges, RIS has been proposed as a transformative technology for next-generation wireless networks as shown in Fig. 1. RIS consist of many passive or semi-passive reflecting elements that can intelligently manipulate the wireless propagation environment by adjusting the phase, amplitude, and polarization of incident signals. This enables RIS to strengthen weak signals, redirect transmissions around obstacles, and extend coverage in scenarios where conventional line-of-sight (LoS) paths are unavailable. By integrating RIS into 6G architectures, the propagation channel itself becomes programmable, opening a new paradigm for wireless network design^4,5.

Figure 2 illustrates a cyber-physical communication framework for 6G networks that leverages a Cyber Twin to manage and optimize QoS for diverse streaming applications. The Cyber Twin acts as the central intelligence, receiving inputs from the Logger (monitoring network conditions and resource utilization) and the Asset (system functions and services). Based on real-time data, the Cyber Twin performs request processing, resource utilization analysis, and classification of traffic types, ensuring appropriate service differentiation. The network integrates multiple user domains such as mobile users, IoT devices, vehicular networks, and satellite communication, all interconnected through THz links and advanced 6G radio access technologies. The Cyber Twin orchestrates these heterogeneous connections, optimizing throughput, reducing latency, and guaranteeing reliability. Finally, the system delivers service responses back to end-users, supported by AI-driven decision-making that maintains efficient communication between network entities⁶.

Fig. 2

6G network model with cyber twin intelligence applications (Quoted from⁶).

Despite advancements in 5G and ongoing development toward 6G, maintaining reliable connectivity in NLoS and high-path-loss environments remains a critical challenge. Severe degradation of SNR at terahertz frequencies leads to increased error rates, excessive retransmissions, and high latency, ultimately compromising QoS. Existing solutions cannot sufficiently balance coverage, reliability, and efficiency under these harsh conditions. Therefore, innovative approaches are required to mitigate propagation losses, sustain high data throughput, and meet the stringent requirements of emerging 6G applications such as holographic communications, industrial automation, and extended reality (XR).

RIS enable programmable manipulation of high‑frequency propagation by coherently steering reflections to recover NLoS links and stabilize received power, which is pivotal for achieving URLLC‑grade reliability and low latency at 300 GHz. RIS technology contributes to ultra-reliability by stabilizing received power and minimizing fading-induced outages. Moreover, by improving SNR and reducing retransmissions, RIS helps achieve the stringent latency targets required by URLLC applications such as autonomous control and industrial automation. This motivates the proposed RIS‑assisted 6G architecture and analysis that quantify SNR, BER, latency, and throughput gains for streaming data under matched system assumptions and realistic channel enhancements. The revised contribution explicitly states this role and adds a comparative figure (RIS vs. no‑RIS) demonstrating simultaneous gains in SNR, BER, delay, and throughput aligned with URLLC requirements, thereby justifying RIS as a core enabler in the proposed 6G system.

Related works and background

The evolution of THz and RIS technologies has been crucial in shaping the research roadmap for 6G networks. Early demonstrations, such as the 300 GHz photonic transmitter link referenced in⁷, provided proof-of-concept for ultra-high data rates, achieving bandwidths of up to 54 GHz. Although these studies validated throughput and BER performance in controlled laboratory environments, their scalability and feasibility for deployment remain limited. Likewise, early surveys on millimeter-wave and THz spectrum opportunities, as noted in⁸, established a foundation for identifying frequency bands and addressing propagation challenges, albeit without concrete performance measurements. With the shift toward more advanced channel engineering, studies such as⁹ have investigated wavefront engineering to enhance THz communications. While these contributions offer explanations for achievable SNR and spectral efficiency improvements, they remain limited to simulation results. Large-scale vision papers such as^10,11 identified RIS and THz technology as key enablers of 6G, emphasizing opportunities for convergence between sensing and communication. However, these works primarily function as conceptual surveys without offering quantifiable evaluations of QoS. Experimental efforts have increasingly demonstrated higher data rates while addressing practical impairments. For example¹², reported a 120 Gbps 64-QAM wireless link in the 300 GHz band, with measurable BER and throughput. Likewise¹³, demonstrated NLOS connectivity using frequency-dependent reflective surfaces, revealing improvements in SNR and reliability in reflective-assisted scenarios.

The integration of RIS into THz communication for 6G networks has garnered significant attention recently. A comprehensive survey in¹⁴ examined RIS-assisted THz communications, highlighting the performance enhancements in throughput, BER, and spectral efficiency through numerical simulations. Concurrently, security-enhanced communication frameworks, such as the RIS-aided short-packet non-orthogonal multiple access (NOMA) system in¹⁵, explored discrete phase-shifter designs aimed at improving content protection. While the simulation results indicated improvements in BER, throughput, and latency, these studies relied on idealized RIS hardware and simplified user trust models, which may limit their real-world applicability. In summary, the treatment of QoS metrics such as SNR, BER, throughput, latency, and jitter has varied significantly across the literature. Early studies^7,12, and¹¹ concentrated on the experimental measurement of SNR and throughput, while RIS-related investigations^14,15 prioritized simulation-based validation of multi-metric QoS enhancements. Survey-type contributions^8,10,11 provided useful theoretical frameworks but fell short of offering quantifiable results. The existing gaps in large-scale validation, hardware feasibility, and multi-dimensional QoS benchmarking underscore the necessity for more integrated, real-world testbeds for RIS- and THz-assisted 6G networks. A concise 6G‑URLLC overview is provided to contextualize RIS among emerging enablers for reliability and low latency, summarize recent advances (short‑packet transmission, latency‑bounded scheduling, AI‑assisted control), and identify quantified QoS gaps our 300 GHz evaluation addresses^12,13,14,15. Table 1 summarizes the overall outcomes of all related studies with the proposed system.

Table 1 Related works contributions.

System architecture and contributions

The contribution list is refined to ensure distinctiveness: (i) RIS‑based QoS integration at 300 GHz for streaming, (ii) URLLC‑specific reliability modeling and parameterization, (iii) scalability through RIS‑element density and sensitivity to reflection coefficients, (iv) cross‑layer coupling of PHY QoS to latency/throughput, and (v) practical propagation insights with improved channel realism and robustness analysis. As shown in Fig. 3 that illustrates RIS based wireless communication system where the transmitter (base station) sends signals to the RIS control unit, which, together with the reflection element array and a cyber twin controller, dynamically optimizes the reflection of signals towards the receiver (user equipment). The reflection element array adjusts signal phases based on control commands generated by the cyber twin analytics to enhance signal quality and coverage. The system incorporates feedback loops such as CSI feedback to refine the signal reflection process, enabling energy-efficient, programmable manipulation of the propagation environment to improve wireless communication performance. This architecture promotes intelligent signal control for better coverage and data rates in advanced wireless networks such as 6G.

Fig. 3

RIS control system block diagram.

The cyber‑twin is explicitly integrated into the contribution: it monitors network metrics, predicts channel variations, and updates RIS phases in closed loop to sustain URLLC reliability and latency under dynamic conditions.

Goals for 6G data streaming

Figure 4 highlights four key objectives for optimizing video streaming in 6G networks. First, enhancing quality ensures that users experience higher resolution and more reliable video delivery, addressing the growing demand for immersive applications. Second, reducing energy consumption emphasizes the importance of energy-efficient transmission techniques to lower the overall power cost of streaming, which is critical for sustainable network operations. Third, minimizing latency targets the reduction of transmission delays, a vital requirement for real-time services such as interactive streaming, XR, and holographic communications. Finally, enabling adaptation focuses on dynamic real-time adjustments in streaming quality and resources to meet varying user conditions and network states, ensuring consistent performance across diverse scenarios. Together, these goals illustrate the core challenges and priorities in advancing video streaming performance over 6G^16,17,18,19, and²⁰.

Fig. 4

Objectives of video streaming optimization in 6G.

RIS-assisted terahertz link performance evaluation algorithm and flow mechanism

This algorithm provides a structured framework for evaluating the performance of THz wireless links enhanced by RIS. It begins by computing fundamental constants such as thermal noise power, RIS gain, and spectral efficiency. Then, a distance grid is defined to calculate received power and SNR under three scenarios: LoS, Non-LoS without RIS, and RIS-assisted propagation. Next, throughput is derived from Shannon’s capacity formula as a function of SNR, while BER is estimated using M-QAM modulation and mapped to energy-per-bit over noise (Eb/N0). A FEC threshold is applied to determine reliable transmission regions. Finally, a heuristic latency model relates SNR to packet delay, accounting for retransmission penalties. The algorithm outputs comparative plots for SNR vs. distance, throughput vs. SNR, BER vs. Eb/N0, and latency vs. SNR, offering a comprehensive analysis of how RIS improves QoS in 6G THz communications.

Algorithm

Figure 5 illustrates the proposed workflow for evaluating THz link performance with and without RIS. The process begins with the input of key link parameters, including distance, channel characteristics, and RIS deployment. Next, the received power and SNR are computed for both scenarios, followed by mapping the SNR to achievable throughput based on Shannon’s capacity formula. To assess link reliability, the energy-per-bit to noise ratio (Eb/N0) is further mapped to the BER using multi-level QAM modulation with coding schemes. Latency is then modeled using heuristic approaches to capture end-to-end performance. Finally, the results are synthesized and presented through data visualization for comparative analysis. This step-by-step framework enables a comprehensive evaluation of throughput, reliability, and latency trade-offs in THz communications with RIS assistance.

Fig. 5

THz link performance evaluation workflow with and without RIS.

Theoretical modelling and equations

To enhance rigor and readability, equations are redistributed and contextualized: noise power for SNR calculation, RIS array gain, FSPL‑based link budget and SNR vs. distance, Shannon capacity, M‑QAM BER (with coding threshold), and an SNR‑dependent latency model, each cited and invoked where its KPI is analyzed. Each model states assumptions, applicability, and literature consistency, and coding‑aware BER together with the improved 300 GHz channel model (including absorption) align the analysis with realistic performance envelopes^{21,22,23,24,25,26,27,28,29,30,31,32,33,33,34,35}.

• Noise power: Used to specify SNR calculations accurately, fulfill request for noise-level modeling, support simulation method clarity. Equation (1) shows Simulation transparency, hardware settings clarity, reliability definition, and URLLC param. (1, 2, 5, 6, 17)³¹.

$$\:{P}_{noise\:\left(dBm\right)}=10{\text{log}}_{10}({K}_{B}{T}_{B}\times\:{10}^{-3})+NF$$

(1)

Where:

$\:{K}_{B}$: is Boltzmann’s constant, $\:{T}_{B}$: Temperature, B: Bandwidth, and $\:NF$: is receiver noise figure.

• RIS array gain: directly quantifies impact of RIS; enables hardware sensitivity analysis (reflection coefficient), and links theoretical advantage to practical setup. Equation (2) shows the novelty of RIS, hardware/parameter sensitivity, efficiency of RIS element scaling (1, 3, 16)³².

$$\:{G}_{RIS\:\left(dB\right)}=10{\text{log}}_{10}({N}_{RIS}.\text{m}\text{a}\text{x}({\rho\:}_{ref},{10}^{-6}\left)\right)\:\:\:\:\:\:\:$$

(2)

Where:

$\:{N}_{RIS}$: RIS Elements, and $\:{\rho\:}_{ref}$: reflection coefficient.

• Link budget and SNR vs. propagation distance (connects simulation to real-world 6G/THz): Underpins SNR/distance figures and BER/distance curves; matches with reproducible and physically realistic results, it’s also for Realistic simulation, distance/SNR modeling, THz/300GHz clarity, and robustness/statistical realism (4, 6, 10, 17, 18)⁴ as the following:

Path loss (free space):

$$\:FSPL\left(d\right)=20{\text{log}}_{10}\left(d\right)+20{\text{log}}_{10}\left(f\right)+20{\text{log}}_{10}\left(\frac{4\pi\:}{c}\right)$$

(3)

Received power (various scenarios):

$$\:{P}_{rx,LoS}\left(d\right)={P}_{tx\left(dBm\right)}+{G}_{tx\left(dBi\right)}+{G}_{rx\left(dBi\right)}-FSPL\left(d\right)$$

(4)

$$\:{P}_{rx,RIS}\left(d\right)={P}_{rx,LoS}\left(d\right)+{G}_{RIS\:\left(dB\right)}$$

(5)

• Shannon capacity (Throughput vs. SNR improvement via RIS): Demonstrates RIS-induced throughput improvement, critical for comparative analysis and linking theoretical/empirical outcomes, also it ensures QoS parameterization, URLLC throughput, comparative verification, RIS benefit quantification (6, 13)²².

$$\:C=B.{\text{log}}_{2}(1+SNR)$$

(6)

• BER for M-QAM (Links reliability to SNR/RIS/URLLC): Gives realistic reliability results, also it enables direct response to URLLC error requirements, robustness, and BER with/without RIS¹⁴.

$$\:BER_{{MQAM}} \approx \:\:(\frac{4}{k}\:\cdot(1\: - \frac{1}{{\sqrt M }})\:\cdot\:Q(\:\sqrt {\frac{{3.SNR}}{{M - 1}}} \:)$$

(7)

Where: $\:M$ is Modulation order, and $\:K={\text{log}}_{2}M$.

• Heuristic latency model (explains packet delay, and urllc compliance): Connects SNR directly to delay, verify URLLC delay criteria are satisfied in simulation, and demonstrates latency improvements attributable to RIS³³.

$$\:D(SNR\_dB)\: = \:D_{0} \: + \:D_{r} \:\cdot\:\left[ {\:\frac{1}{{(\:1\: + e^{{(\:a\cdot(SNR\:\left( {dB} \right)\: - \:b)\:)}} \:)}}\:} \right]$$

(8)

Where: $\:{D}_{0}$: Base Delay, $\:{D}_{r}$: retransmission penalty, and $\:a,\:b$ are fitting coefficients.

Simulation setup and methodology

Figure 6 presents a structured categorization of essential communication parameters for evaluating RIS-assisted THz links. It organizes variables into three main groups: Network Parameters, Channel and Noise Parameters, and Microwave Parameters. Network Parameters cover key system factors such as transmitted power, carrier frequency and bandwidth, streaming packet size, modulation order, coding rates, error correction thresholds, and latency-related elements like base delay and retransmission penalties. Channel and Noise Parameters include environmental and hardware-specific constants like temperature, Boltzmann’s constant, noise figure, path loss exponent, and fading effects, which influence signal propagation and quality. Microwave Parameters focus on hardware characteristics such as the number of transmitting and receiving antennas, RIS passive elements count, and RIS reflection efficiency, critical for modeling the reconfigurable surface’s impact. Together, these parameters provide a comprehensive framework for characterizing and simulating the performance of RIS-aided THz communication systems. All references to “THz” are harmonized by explicitly denoting “300 GHz (THz sub‑band)” to avoid ambiguity between band‑level discussion and the specific operating frequency used in simulations as terminology consistency.

Fig. 6

Communication system parameters configuration for RIS-assisted THz links.

The study employs a comprehensive set of parameters spanning network, microwave, and channel domains to model the RIS-assisted THz communication system accurately. The network parameters include a transmitted power of 10 dBm at a carrier frequency of 300 GHz with a bandwidth of 3 GHz, streaming packet size of 3 Mb, and 64-QAM modulation with a code rate of 0.8. Critical thresholds such as the SD-FEC pre-FEC error threshold (2.7 × 10⁻²) and a threshold SNR of 15 dB are defined alongside operational metrics including transmission distances from 1 to 200 m and base delays ranging from 0.5 to 1.5 milliseconds. Microwave parameters detail a massive MIMO setup with 256 transmitting and 256 receiving antennas, augmented by RIS passive elements varying between 16, 64, 100, and 200, with a reflection coefficient of 0.9 representing RIS efficiency. Channel and noise conditions are modeled considering a temperature of 290 K, the Boltzmann constant (1.38 × 10⁻²³), a noise figure of 10 dB, variable path loss exponents (1.6 to 5), and an excess fading figure of 20 dB, thereby capturing environmental and hardware-induced impairments. Together, these parameters establish a rigorous framework for evaluating system performance across diverse propagation scenarios and RIS configurations as shown in Tables 2 and 3, and Table 4. Simulations are implemented in MATLAB R2024b and Python, with explicit reporting of bandwidth, noise figure, transmit power, carrier at 300 GHz, scenario‑specific path‑loss exponents, RIS element counts, and reflection coefficient range for reproducibility. URLLC parameters are enforced in configuration: packet‑error‑rate target, latency budget near 2ms, and reliability probability approaching 99.999%.

Table 2 Network parameters.

Table 3 Microwave parameters (mMIMO + RIS).

Table 4 Channel and noise parameters.

Figure 7 illustrates the role of RIS in enhancing THz signal propagation for reliable 6G communications. On the left, it shows that without RIS, THz signals are weak due to blockages and poor reception, leading to degraded communication performance. By positioning an RIS between the base station (BS) and the user equipment (UE), the surface intelligently reflects and redirects the blocked or attenuated signals toward the UE. This reflection mechanism significantly strengthens the received signal, as depicted on the right side of the figure, where a strong THz signal enables reliable 6G communication. The diagram highlights the practical benefit of implementing RIS as an effective solution to overcome high-frequency signal blockages and ensure robust connectivity in next-generation wireless networks.

Fig. 7

Enhancing THz signal strength using RIS for reliable 6G communication.

Path‑loss exponents are refined to realistic ranges (e.g., 3.0–4.5) and scenario labels (urban LoS, dense urban, indoor LoS) are applied consistently across parameters tables and figure captions. As shown in Table 5, Path-loss exponents are selected based on canonical propagation models: n = 2.0 for free-space/ideal LoS, n = 3.0 for typical urban LoS, and n = 3.5–4.5 for NLoS and dense urban/indoor environments, reflecting empirical studies and standard practice.

Table 5 Summary of simulation parameters and environment-dependent path-loss exponents^{14,31,32,33,34}.

Results and discussions

Figure 8 illustrates the variation of SNR as a function of distance for three key scenarios at a carrier frequency of 300 GHz: LOS propagation (path loss exponent n = 2.0), NLOS without RIS assistance (n = 3.5), and NLOS with RIS deployment utilizing 16 reflecting elements. As distance increases, SNR degrades rapidly under NLOS conditions without RIS, dropping below acceptable communication thresholds even at moderate ranges. In contrast, the deployment of RIS significantly offsets this degradation, providing a notable SNR improvement that shrinks the performance gap toward the LOS baseline. This demonstrates that RIS integration at 300 GHz effectively extends coverage and enhances link reliability in challenging NLOS scenarios, thus supporting the strict SNR requirements essential for 6G URLLC and high-capacity applications.

Fig. 8

SNR vs. distance at 300GHZ (0 to 500 m).

Figure 9 presents the variation of SNR with distance for different numbers of RIS elements, ranging from no RIS-to-RIS configurations with 16, 64, 100, and 200 elements. It is evident that the inclusion of RIS significantly improves the SNR compared to the no-RIS case, which shows a sharp decline and negative SNR values beyond approximately 50 m. As the number of RIS elements increases, the SNR improves across all distances. Specifically, RIS with 16 elements offers modest enhancement, while RIS with 64, 100, and 200 elements provide increasingly better performance. The curve for RIS with 200 elements achieves the highest SNR, maintaining values above 35 dB even at 500 m. This performance gain demonstrates the scalability and effectiveness of RIS in compensating for path loss in high-frequency communication systems. The results confirm that deploying a higher number of RIS elements can substantially extend the communication range and ensure reliable connectivity, which is critical for the successful deployment of future 6G networks in non-line-of-sight and long-range scenarios.

Fig. 9

SNR vs. distance for various RIS element counts.

Figure 10 compares the SNR performance over distance in NLOS scenarios with and without RIS assistance, where the RIS employs 64 reflecting elements. The results clearly indicate that, as the transmission distance increases up to 300 m, the SNR decays sharply in the conventional NLOS setting, falling below zero dB at moderate ranges and limiting reliable communication. By contrast, the deployment of RIS significantly boosts the SNR across all distances, maintaining positive and robust SNR values where the conventional NLOS link would otherwise be severely impaired. This demonstrates that leveraging a larger RIS array (N = 64) can substantially enhance link quality, extend coverage, and satisfy strict reliability requirements for 6G applications, especially under challenging NLOS conditions.

Figures 8 and 10 are updated to show SNR vs. distance up to 500 m at 300 GHz with denser sampling and clear LoS/NLoS labels, ensuring consistent scaling and interpretability.

Fig. 10

NLoS SNR vs. distance (300GHZ) with and without RIS.

Fig. 11

Comparison of system performance metrics for 6G THz links with and without RIS.

Figure 11 presents a radar plot comparing the overall system performance of RIS-assisted and No-RIS scenarios across four critical metrics: SNR, BER, delay, and throughput. The “RIS-assisted” configuration occupies the maximum area on the radar chart, signifying near-optimal or normalized performance across all key indicators high SNR, ultra-low BER, minimal delay, and high throughput which collectively meet URLLC standards. In stark contrast, the “No RIS” profile is almost imperceptible at the chart’s center, reflecting poor performance in all metrics. This visualization clearly demonstrates that integrating RIS leads to simultaneous and significant improvements in reliability, latency, and data rate, offering a compelling justification for RIS adoption in 6G networks aimed at mission-critical, high-capacity applications.

QoS‑to‑URLLC linkages are made explicit: increased SNR reduces BER and retransmissions, which shortens latency and elevates effective throughput under the same bandwidth and power constraints.

Figure 12 illustrates the relationship between the SNR and the average packet delay, highlighting the heuristic delay response in wireless communication systems. The curve shows that at low SNR values (in the range highlighted in red), the average packet delay remains high and constant, around 5.5ms, which is attributed to frequent retransmission due to unreliable links, an effect observed in the absence of a RIS. As SNR increases, there is a sharp transition around 15–20 dB, where packet delay rapidly decreases, indicating improved link quality and fewer retransmissions. In the green-shaded region, corresponding to high SNR values (above 20 dB), the delay stabilizes at a much lower value, reflecting a reliable link scenario that can be achieved with the integration of RIS. This distinction between high retransmission zones and reliable link operation clearly demonstrates the benefits of RIS in reducing packet delays by boosting SNR and ensuring stable, efficient communications, an essential improvement for applications requiring ultra-reliable low-latency performance.

Fig. 12

Heuristic delay (MEC) vs. SNR.

Figure 13 illustrates the impact of SNR on the average packet delay in a wireless communication system, emphasizing the effect of link reliability enhancements such as RIS. At lower SNR values, a region highlighted in red, the average packet delay remains high at approximately 16.5ms which reflects increased retransmissions necessitated by an unreliable link in the absence of RIS. This region signifies poor channel conditions where data packets frequently require resending, leading to significant latency. As the SNR improves and crosses a threshold (around 15 to 20 dB), there is a rapid decrease in average delay, represented by the sharp transition on the curve. In the higher SNR region, shaded in green, the average packet delay stabilizes at a much lower value, around 2ms, signifying a reliable communication link where retransmissions are minimal. This demonstrates that deploying RIS or similar techniques can effectively shift the operating regime into higher SNR values, substantially reducing packet delay and achieving the low-latency and reliable communications essential for next-generation wireless networks. The contrasting shaded regions in the plot visually underline the benefits of improved link reliability, making a compelling case for the integration of RIS to meet stringent latency requirements.

Fig. 13

Heuristic delay vs. SNR.

channel model and coded BER

Figure 14 displays the bit error rate (BER) performance as a function of SNR in dB for 64-QAM modulation, comparing uncoded BER, coded BER (with forward error correction, FEC), and the relevant pre-FEC threshold for reliable decoding. The uncoded BER curve (yellow) decreases steadily with higher SNR but remains well above the pre-FEC threshold in the moderate SNR regime, indicating high error rates without coding. The coded BER (blue) shows a substantial improvement, yielding orders-of-magnitude lower error rates and crossing below the threshold at significantly lower SNR. This demonstrates the crucial role of FEC in enabling ultra-reliable communications at practical SNRs and validates that, with appropriate coding, the RIS-assisted system can reliably operate within strict error probability requirements essential for URLLC and high-capacity 6G links.

Fig. 14

Bit error rate (BER) vs. SNR (coded and uncoded).

Figure 15 illustrates the relationship between throughput and SNR in the context of wireless communications, emphasizing the impact of RIS. The blue curve represents the Shannon capacity, serving as the theoretical upper bound for throughput as a function of SNR. Notably, the graph highlights two distinct operational regions: the typical region without RIS, shaded in red, and the typical region with RIS, shaded in green. In the absence of RIS, the SNR primarily spans from around − 20 dB to 0 dB, corresponding to lower throughput values and indicating challenging propagation conditions, such as those commonly encountered in non-line-of-sight or obstructed environments. Conversely, the implementation of RIS enables the SNR to shift significantly higher, generally covering the 15 dB to 45 dB range, which translates to a dramatic increase in achievable throughput. This enhancement can be attributed to RIS’s ability to manipulate and optimize the wireless environment, resulting in stronger signal reception at the receiver. The sharply rising trend of throughput with increasing SNR in the RIS region demonstrates the potential of RIS technology to bridge the gap between practical wireless systems and theoretical capacity limits, thereby highlighting RIS as a key enabler of high-performance future communication systems.

Fig. 15

Throughput (Gbps) vs. SNR.

sensitivity and scenario settings

Figure 16 illustrates the variation of SNR with distance under different path-loss exponents when aided by a RIS with 16 elements. Six propagation environments are compared: free space (n = 2.0), urban LoS (n = 2.7), urban NLoS (n = 3.5), dense urban (n = 4.0), buildings LoS (n = 1.6), and buildings NLoS (n = 5.0). The results show a clear dependence of SNR performance on the propagation environment, where smaller path-loss exponents lead to slower degradation of SNR with distance. For instance, the buildings LoS case (n = 1.6, purple curve) yields the best performance, maintaining SNR above 40 dB even at 200 m, followed by free space (n = 2.0, blue curve), which stabilizes around 20 dB at longer distances. Urban LoS (n = 2.7, orange curve) shows moderate performance, with SNR dropping below 0 dB beyond 100 m. In harsher conditions, such as urban NLoS (n = 3.5, green curve) and dense urban (n = 4.0, red curve), SNR deteriorates quickly, falling well below − 50 dB at longer ranges. The most challenging scenario is buildings NLoS (n = 5.0, brown curve), where SNR becomes extremely poor, reaching below − 150 dB even within 200 m, indicating severe signal degradation despite RIS assistance. These findings confirm that while RIS can significantly improve signal strength, the effectiveness strongly depends on the propagation environment, with LoS and low path-loss conditions benefiting most, whereas highly obstructed NLoS scenarios still face critical performance limitations.

Fig. 16

SNR vs. distance (RIS for different path-loss exponents, N = 16).

A reflection‑coefficient sensitivity analysis (0.5–0.9) quantifies performance trends and justifies parameter selection against commercial ranges, with 0.9 reported as an upper‑bound ideal benchmark. Figure 17 illustrates the relationship between the SNR (in dB) and the RIS reflection coefficient, showcasing the sensitivity of end-to-end communication performance to practical hardware settings. As the reflection coefficient increases from 0.5 to 0.9, the SNR improves significantly in a near-linear fashion, rising from approximately − 10 dB to over 15dB. This result highlights that achieving higher reflection efficiency with RIS hardware can yield substantial SNR gains, directly impacting system reliability, coverage, and overall QoS. Thus, the figure justifies the importance of optimizing and accurately specifying the reflection coefficient in simulations and aligns the chosen parameter range with values reported for state-of-the-art RIS prototypes.

Fig. 17

SNR vs. reflection coefficient (0.5–0.9).

Figure 18 shows the variation of SNR with distance for direct LOS transmission under different path-loss exponents, representing various propagation environments. Six cases are considered: free space (n = 2.0), urban LOS (n = 2.7),urban NLOS (n = 3.5), dense urban (n = 4.0), buildings LOS (n = 1.6), and buildings NLOS (n = 5.0). As expected, SNR decreases with distance in all cases, but the rate of degradation strongly depends on the path-loss exponent. The buildings LOS scenario (n = 1.6, purple curve) provides the best performance, maintaining an SNR above 20 dB at 200 m due to the low attenuation factor. Free space (n = 2.0, blue curve) also performs relatively well, with SNR gradually decaying to around 0 dB at long distances. Urban LOS (n = 2.7, orange curve) shows moderate performance, falling below − 40 dB beyond 200 m, while urban NLOS (n = 3.5, green curve) and dense urban (n = 4.0, red curve) experience much faster degradation, with SNR dropping below − 100 dB at the same distance. The most challenging scenario is buildings NLOS (n = 5.0, brown curve), where SNR deteriorates drastically, reaching below − 180 dB within 200 m, indicating impractical communication without additional support such as RIS or relaying. These results emphasize that while direct LOS paths are generally favorable, the propagation environment and associated path-loss exponent critically determine link quality, with harsh urban and obstructed environments requiring advanced techniques to sustain reliable communication in 6G systems.

Fig. 18

LoS (direct path) SNR vs. distance for different path-loss exponents.

Robustness and statistical stability

Figure 19 illustrates the robustness of the proposed system by showing the mean SNR as a function of distance across 10 independent simulation runs, with the shaded region representing one standard deviation (±σ). The plot demonstrates that the SNR consistently decreases with distance, but the narrow confidence band indicates low variability and high statistical reliability of the simulated results. This validation of robustness confirms that the system maintains predictable performance under randomized channel and noise realizations, making the findings trustworthy for practical deployment scenarios. Statistical robustness has been introduced by averaging outcomes over ten independent simulation runs with randomized noise and fading profiles. The mean and variance of SNR, BER, and delay are reported, confirming high stability and reproducibility of the presented results.

Fig. 19

Robustness – Mean SNR ± σ (10 Runs).

Benchmarking vs existing work

A compact benchmarking subsection compares No‑RIS and RIS (this work) with CF‑mMIMO³⁴ and relay baselines under matched or normalized assumptions at 300 GHz³⁵, reporting SNR, BER (log scale), latency, and throughput with explicit citations and notes on normalization for fairness; the grouped‑bar figure provides a concise visual of cross‑method trends. A dedicated benchmarking subsection has been added that contrasts our RIS-assisted THz results with recent RIS–THz studies, showing a consistent SNR advantage of + 12 dB, end-to-end latency under 2ms, and throughput above 20 Gbps, while aligning evaluation conditions and normalizing assumptions across works to ensure fair comparison. The added comparison explicitly demonstrates that the proposed framework achieves higher reliability (via SNR/BER), lower delay, and higher data rates than recent baselines, addressing the request for horizontal and vertical verification against existing literature. SNR exceeds recent works by + 12 dB while meeting URLLC latency and throughput targets as shown in Fig. 20.

Fig. 20

Key metrics achieved by the proposed RIS-assisted 6G THz system.

Table 6 benchmarks four methods under matched assumptions (300 GHz carrier, identical bandwidth, noise figure, transmit power, and path‑loss model), reporting SNR, BER, latency, and throughput for No‑RIS, RIS (this work), CF‑mMIMO³⁴, and Relay³⁵ to enable fair, side‑by‑side comparison across reliability and QoS metrics. Figure 21 offers a compact visual summary of the same KPIs, highlighting the RIS gains observed in our study (SNR from − 20 to 45 dB, latency from 16.5 to 2ms, throughput from 1 to 20 Gbps, and BER from $\:2.8\times\:{10\:}^{1}$to $\:5.5\times\:{10}^{-10}$ relative to the No‑RIS baseline for immediate URLLC relevance. Together, the table and figure substantiate the comparative verification requested by the reviewer by situating this work alongside representative CF‑mMIMO and relay baselines while keeping assumptions aligned and sources explicitly cited for reproducibility.

Fig. 21

Benchmarking vs. existing work (300 GHz).

Table 6 Benchmarking against existing work under matched assumptions (300 ghz carrier, identical B/NF/Tx power/path loss.

The visualized results collectively confirm the substantial performance improvement achieved through RIS integration in wireless systems. The radar chart as shown in Fig. 22 highlights how RIS consistently elevates all major performance indicators SNR, delay, BER, and throughput moving the communication link from unreliable to highly reliable operational regimes. Complementing this, the heatmap that shown in Fig. 23 demonstrates that RIS scalability, through increasing element counts, effectively counteracts path loss across longer transmission distances, ensuring sustained SNR levels even in challenging non-line-of-sight environments. These findings reveal that RIS not only addresses individual weaknesses such as poor SNR or high BER but also provides a comprehensive enhancement across the entire communication chain. Overall, the system performance improvement with RIS is evident: links become more robust, latency is minimized, error rates are nearly eliminated, and throughput is significantly boosted. This holistic gain establishes RIS as a critical enabler for ultra-reliable, low-latency, and high-capacity communication in future 6G networks.

Fig. 22

Overall performance RIS vs. No-RIS (radar graph).

Fig. 23

NR performance heatmap (RIS elements).

Scalability and multi-scenario applicability of RIS-enhanced 6G systems

To extend the scope beyond single-user and point-to-point analyses, we address the scalability of RIS-assisted communication under multi-user, multi-cell, and dynamic mobility scenarios. In practical 6G networks, multiple users simultaneously access shared radio resources, leading to new interference management challenges, especially in dense urban or high-mobility environments. The integration of RIS is shown to facilitate distributed beamforming and spatial scheduling, enabling the network to flexibly enhance coverage, reduce inter-user interference, and direct resources adaptively as user density or traffic demand changes. Key scalability issues include channel estimation overhead, synchronization among RIS elements and base stations, and the control signaling required for real-time RIS configuration in heterogeneous and dynamic deployments. To cope with these, advanced algorithms such as distributed and AI-based resource allocation, hybrid RIS control (centralized/decentralized), and reinforcement learning for real-time adaptation are considered essential and have demonstrated strong potential in recent literature for maintaining reliable and efficient performance at scale.

Finally, the analysis outlines how coordination among multiple RISs and base stations, combined with network-level orchestration frameworks, can address dynamic environment factors (e.g., user mobility, blockage, or rapidly changing topologies), ensuring robust, low-latency service for all users. This expanded discussion confirms that RIS technology is not limited to static or point-to-point scenarios but is a scalable enabler for future 6G networks supporting multi-user access, dynamic scheduling, and network densification, as validated by simulation insights and supported by emerging research directions.

Conclusion

The results of this study provide compelling evidence that RIS is a key enabler for ensuring QoS in 6G networks. By intelligently reflecting and redirecting terahertz signals, RIS improved SNR by up to 65dB compared to conventional NLoS transmission, extended communication distances beyond 500 m, and shifted links from unreliable to ultra-reliable regimes. BER was reduced by more than 10 orders of magnitude, transforming highly error-prone channels (near 2.8101) into error-free communication ($( < 5.5 \times 10^{{ - 12}} )$). In parallel, packet delay was cut from 16.5ms to as low as 2ms, ensuring URLLC-grade performance, while throughput increased from sub-Gbps levels to more than 20 Gbps, approaching Shannon capacity. These results also confirms that RIS at 300 GHz programmatically improves propagation to deliver large SNR gains, ultra‑low BER after coding, near‑URLLC latency, and multi‑Gbps throughput, with robustness across runs and competitive benchmarking versus CF‑mMIMO and relay alternatives under matched or normalized scenarios. These quantitative improvements not only validate the theoretical promise of RIS but also establish its practical role in meeting the stringent QoS requirements of emerging 6G applications such as holographic communications, autonomous systems, and extended reality. The study confirms that RIS-assisted architecture can bridge the gap between harsh propagation conditions, and the ambitious performance demands of 6G networks.

Suggested future work

Future research should consider experimental validation of the proposed RIS-assisted communication models to complement the simulation results presented here. Additionally, optimization of RIS design parameters including element count, spatial arrangement, and adaptive control algorithms could be explored to further improve performance and energy efficiency. Investigations into dynamic RIS configurations in highly mobile or dense urban environments will be important to understand real-world deployment challenges. Moreover, evaluating multi-user scenarios and the impact of RIS on network-level throughput and interference management remains a critical area for ongoing research. Finally, expanding studies to other frequency bands and integrating RIS with emerging technologies such as beamforming and massive MIMO would provide a more comprehensive solution framework for future wireless systems.