On the Road to 6G: Visions, Requirements, Key Technologies, and Testbeds

This paper does not test an intervention. It is a **survey and vision paper** that reviews and synthesises hundreds of existing research papers, white papers from industry (e.g., Huawei, Nokia, Samsung, Qualcomm), and standards body documents (ITU-R, 3GPP) to create a structured overview of:

The **vision** for 6G: what services and experiences it should enable by ~2030

The **technical requirements** (e.g., peak data rate, latency, reliability, energy efficiency)

The **application scenarios** (e.g., holographic communications, digital twins, pervasive AI, integrated sensing and communication)

The **key enabling technologies** (e.g., terahertz communications, reconfigurable intelligent surfaces, cell-free massive MIMO, AI-native air interface, integrated sensing and communication)

The **testbeds and verification platforms** that have been built to prototype these technologies

There is no comparator group, no outcome measure, no effect size, and no statistical analysis. The paper is a qualitative synthesis of expert opinion and early-stage research.

No human participants were studied. The paper reviews:

**Published research papers** (the reference list contains approximately 300 citations)

**Industry white papers** from Huawei, Nokia, Ericsson, Samsung, Qualcomm, and others

**Standards documents** from ITU-R (International Telecommunication Union – Radiocommunication Sector) and 3GPP (3rd Generation Partnership Project)

**Testbed descriptions** from universities and research labs in China, Europe, the US, South Korea, and Japan

The "population" is the global research community working on 6G, not a sample of human subjects.

No measurements were taken. The paper uses:

**Qualitative comparison** of different proposed architectures and technologies

**Quantitative target values** drawn from industry and standards documents (e.g., "peak data rate of 1 Tbps", "latency of 0.1 ms", "positioning accuracy of 1 cm")

**Timeline analysis** of when different technologies are expected to mature (e.g., "terahertz components expected to be commercially viable by 2028–2030")

**Testbed performance metrics** reported by other researchers (e.g., "achieved 100 Gbps at 300 GHz using a photonics-based transmitter")

These are not experimental results from a controlled study. They are aspirational targets and early prototype demonstrations.

**Study design:** This is a **comprehensive survey paper** — a type of secondary research that systematically collects, organises, and critiques existing primary research and industry documents. It is not a systematic review or meta-analysis (no formal search strategy, no inclusion/exclusion criteria, no quality assessment of included studies). It is closer to a **narrative review** with elements of a **technology roadmap**.

**How the authors gathered information:**

They reviewed published academic papers from IEEE journals and conferences (the primary source)

They reviewed industry white papers from major telecommunications companies

They reviewed standards body documents from ITU-R and 3GPP

They reviewed testbed descriptions from university and industry labs

They organised the information into thematic categories: vision, requirements, scenarios, technologies, testbeds, open challenges

**What this design can and cannot prove:**

**Can prove:** Nothing in the experimental sense. The paper can show what the current consensus is among researchers and industry players about where 6G is heading. It can identify which technologies are receiving the most research attention and which have been demonstrated in prototype form.

**Cannot prove:** That any specific technology will actually work at scale, meet its performance targets, or be commercially viable. The paper cannot establish cause-and-effect relationships. It cannot provide effect sizes or confidence intervals because no experiments were conducted.

**Major methodological weaknesses:**

**Selection bias:** The authors are based at Chinese universities (Southeast University, Nanjing University of Posts and Telecommunications) and may over-represent Chinese research efforts. The paper does not disclose a systematic search strategy, so it is unclear how comprehensive or balanced the coverage is.

**No quality assessment:** The paper does not evaluate the quality of the studies it cites. A testbed demonstration from a well-funded lab is treated similarly to a theoretical simulation from a small group.

**No quantitative synthesis:** There is no meta-analysis or statistical pooling of results. The paper reports "peak data rates of 100 Gbps" from one testbed and "1 Tbps" as a target, but does not explain the gap or the uncertainty.

**Industry influence:** Many of the cited white papers come from companies with commercial interests in shaping the 6G standard. The paper does not discuss potential conflicts of interest.

**Rapidly changing field:** The paper was published in 2023, but 6G research is moving quickly. Some of the technologies described may already be obsolete or superseded.

Because this is a survey paper, the "findings" are the authors' synthesis of the current state of the field. Here are the main points:

**6G Vision and Requirements:**

6G is expected to be commercially deployed around **2030**, with research and standardisation happening now (2020–2025) and prototype development in 2025–2028

6G targets a **peak data rate of 1 Tbps** (1000 Gbps), compared to 5G's target of 20 Gbps — a **50× increase**

**Latency target: 0.1 ms** (one-way), compared to 5G's 1 ms — a **10× reduction**

**Positioning accuracy: 1 cm** indoors, compared to 5G's ~10 cm — a **10× improvement**

**Energy efficiency: 100× improvement** over 5G (bits per joule)

**Connection density: 10⁷ devices/km²**, compared to 5G's 10⁶ — a **10× increase**

**Reliability: 99.99999%** (seven nines), compared to 5G's 99.999% (five nines)

**Key Application Scenarios:**

**Holographic communications:** Real-time 3D hologram transmission for telepresence, education, entertainment

**Digital twins:** Real-time digital replicas of physical systems (factories, cities, human bodies) for monitoring and simulation

**Pervasive AI:** AI embedded throughout the network, enabling autonomous decision-making at the edge

**Integrated sensing and communication (ISAC):** Using the same radio signals for both communication and radar-like sensing (detecting objects, tracking movement)

**Extended reality (XR):** High-fidelity augmented and virtual reality with tactile feedback

**Space-air-ground-sea integrated networks:** Seamless connectivity across satellites, drones, ships, and underwater sensors

**Key Enabling Technologies:**

**Terahertz (THz) communications:** Using frequencies from 100 GHz to 3 THz to achieve ultra-high data rates. Current testbeds have demonstrated **up to 100 Gbps at 300 GHz** over short distances (1–10 metres)

**Reconfigurable intelligent surfaces (RIS):** Flat panels with thousands of tiny, electronically controllable elements that can reflect, focus, or steer radio waves to improve coverage and capacity. Simulations suggest **2–10× improvement in signal strength** in indoor environments

**Cell-free massive MIMO:** Hundreds or thousands of distributed antennas working together to serve users, eliminating cell boundaries. Simulations show **5–10× improvement in spectral efficiency** compared to traditional cellular architectures

**AI-native air interface:** Embedding machine learning directly into the physical layer (channel estimation, modulation, coding) rather than using it only for higher-level tasks. Early prototypes show **10–20% improvement in throughput** under realistic channel conditions

**Integrated sensing and communication (ISAC):** Sharing hardware and spectrum between communication and radar functions. Prototypes have demonstrated **simultaneous 10 Gbps data rate and 1 cm ranging accuracy** at 28 GHz

**Visible light communications (VLC):** Using LED lights to transmit data. Laboratory demonstrations have achieved **up to 10 Gbps** over short distances

**Orbital angular momentum (OAM):** Using twisted radio waves to encode additional data channels. Early experiments have shown **2–4× spectral efficiency improvement** in controlled settings

**Testbeds and Verification Platforms:**

The paper catalogues **over 20 testbeds** worldwide, including:

- **China:** The "6G Test Network" at Southeast University (terahertz, RIS, cell-free MIMO)

- **Europe:** The "Hexa-X" project testbed (AI-native air interface, ISAC)

- **US:** The "PAWR" platform (mmWave and sub-terahertz)

- **South Korea:** The "6G R&D Center" at KAIST (terahertz, OAM)

- **Japan:** The "Beyond 5G/6G" testbed at NICT (terahertz, photonics)

Most testbeds are **indoor, short-range (1–50 metres)**, and use **laboratory-grade equipment** (not commercial hardware)

**No testbed has demonstrated all 6G requirements simultaneously** — each testbed focuses on one or two technologies

**Open Challenges and Future Directions:**

**Hardware limitations:** Terahertz components (amplifiers, antennas, mixers) are still expensive, inefficient, and difficult to manufacture at scale

**Channel modelling:** The propagation characteristics of terahertz and sub-terahertz frequencies are not well understood, especially in outdoor and dynamic environments

**AI complexity:** Training and deploying AI models at the network edge requires significant computational resources and energy

**Security and privacy:** New attack surfaces emerge with pervasive AI, integrated sensing, and massive connectivity

**Standardisation:** ITU-R is expected to define the 6G vision by **mid-2023**, but detailed technical specifications will not be finalised until **2028–2029**

**Energy consumption:** While 6G targets 100× better energy efficiency, the absolute energy consumption of dense antenna arrays and terahertz circuits may still be very high

There are no effect sizes from a controlled experiment. However, the paper reports **target performance improvements** relative to 5G:

**Peak data rate:** 1 Tbps vs. 20 Gbps — roughly **50× faster** (like going from a 4K video stream to a holographic video stream)

**Latency:** 0.1 ms vs. 1 ms — roughly **10× faster** (like going from the blink of an eye to a fraction of a blink)

**Positioning accuracy:** 1 cm vs. 10 cm — roughly **10× more precise** (like going from knowing which room someone is in to knowing exactly where they are standing)

**Energy efficiency:** 100× improvement — roughly **two orders of magnitude** more data transmitted per unit of energy

**Connection density:** 10⁷ vs. 10⁶ devices per km² — roughly **10× more devices** (like going from every person having 10 connected devices to every person having 100)

For the specific technologies:

**RIS:** 2–10× improvement in signal strength (simulation-based)

**Cell-free massive MIMO:** 5–10× improvement in spectral efficiency (simulation-based)

**AI-native air interface:** 10–20% improvement in throughput (prototype-based)

**ISAC:** Simultaneous 10 Gbps data rate and 1 cm ranging accuracy (prototype-based)

**OAM:** 2–4× spectral efficiency improvement (laboratory-based)

These are **not statistically validated effect sizes** — they come from simulations, prototypes, and idealised laboratory conditions. Real-world performance will likely be lower.

The authors acknowledge several limitations, and a critical reader would note additional ones:

**What the authors acknowledge:**

The 6G vision is still evolving and "wide open" — there is no global consensus yet

Many key technologies are at an early stage and face significant hardware and theoretical challenges

Testbed results are from controlled environments and may not generalise to real-world deployments

The paper cannot cover all proposed technologies in depth due to space constraints

**What a critical reader would note:**

**No systematic methodology:** The paper does not describe a search strategy, inclusion/exclusion criteria, or quality assessment. It is a narrative review, not a systematic review. This means the selection of papers and technologies may be biased toward the authors' own research interests and geographic region.

**Industry influence:** Many of the cited white papers come from companies (Huawei, Nokia, Samsung, Qualcomm) that have commercial interests in shaping the 6G standard. The paper does not discuss potential conflicts of interest or how industry funding might affect the reported targets.

**Geographic bias:** The authors are based at Chinese universities, and the paper gives disproportionate attention to Chinese research efforts (e.g., the "6G Test Network" at Southeast University is described in detail, while European and US testbeds receive less coverage).

**No quantitative synthesis:** The paper reports performance numbers from individual testbeds but does not pool results across studies or provide measures of uncertainty (confidence intervals, standard deviations). It is impossible to know how reliable or reproducible the reported numbers are.

**Rapid obsolescence:** The paper was published in 2023, but 6G research is moving quickly. Some of the technologies described (e.g., specific terahertz components, AI algorithms) may already be outdated. The paper's value is as a snapshot of the state of the field in early 2023.

**No human subjects:** The paper does not study human users, so it cannot speak to user experience, adoption barriers, or real-world usability of 6G technologies.

**No cost analysis:** The paper does not discuss the economic feasibility of 6G — how much it will cost to deploy, who will pay for it, and whether the benefits justify the investment.

**No comparison to alternative futures:** The paper assumes a particular trajectory for 6G (higher data rates, lower latency, more devices) but does not consider alternative scenarios (e.g., a focus on reliability and security over raw speed, or a shift toward decentralised mesh networks).

**Important caveat:** This paper is a survey of future technologies, not a guide for running a personal experiment. You cannot test "6G" as an intervention because it does not exist yet. However, you can use the paper to identify **specific technologies or concepts** that you could test in your own life or work.

For someone running their own n=1 experiment:

**What to test (specific intervention and dose):**

**Terahertz signal propagation:** If you have access to a terahertz testbed (unlikely for most individuals), you could test how different materials (walls, furniture, clothing) affect signal strength at 100–300 GHz. Dose: vary the distance (1–10 metres) and the material type (drywall, glass, wood, metal).

**Reconfigurable intelligent surface (RIS) placement:** If you have a programmable metasurface (available as a research kit from some universities), you could test how the orientation and position of the RIS affects Wi-Fi signal strength in your home. Dose: place the RIS at different angles (0°, 30°, 60°, 90°) and distances (1–5 metres) from the router.

**Visible light communication (VLC):** You could build a simple VLC system using an LED light and a photodiode (available as hobbyist kits for ~$50–100). Test how data rate varies with distance (0.5–5 metres), ambient light level (dark room vs. sunny window), and LED colour (white, red, blue, green).

**AI-native signal