Demonstration of Massive Connectivity for the 6G Era

With the widespread adoption of drones, robots, and XR devices, next-generation wireless systems (6G) are expected to provide significantly enhanced uplink massive connectivity. Compared with current fifth-generation mobile communication systems (5G), device density in 6G networks is anticipated to increase by more than an order of magnitude. One promising technology for addressing this challenge is non-orthogonal multiple access (NOMA), which enables multiple devices to transmit simultaneously over the same time and frequency resources. In such scenarios, signals from multiple devices are superposed at the base station and must be individually detected. If the number of devices is denoted by 𝐾 and the modulation order by each device by 𝑀, the number of signal combinations grows exponentially as 𝑀^𝐾. Consequently, the computational complexity increases rapidly with the number of connected devices, leading to large processing latency and making real-time detection difficult.

Previously, we developed a hybrid signal processing method that combines a quantum annealing machine with classical computing (hereafter referred to as the “previous method”). In this framework, the quantum annealing machine efficiently explores candidate signal combinations, while a classical computer performs post-processing to estimate the probability distributions required for signal detection, thereby achieving both high detection accuracy and fast processing. However, the effectiveness of the previous method had only been demonstrated for limited communication scenarios. Its applicability to multi-antenna and multi-carrier transmission, which is essential component of both 5G and future 6G systems, remained unclear. To address this limitation, NICT developed a new hybrid signal processing method that integrates the quantum annealing machine with a classical computer and is applicable to multi-antenna and multi-carrier transmission (see Figure 1). The proposed method also incorporates essential components of modern mobile communication systems, such as channel estimation using reference signals, making it suitable for practical 6G deployment.

We evaluated block-error-rate performance of the proposed method through numerical simulations under the following conditions: four receive antennas at the base station, QPSK modulation (𝑀=4), and eight connected devices (𝐾=8). This setting corresponds to a combinatorial optimization problem involving approximately 4⁸≈60,000 possible signal combinations. In these simulations, simulated quantum annealing (SQA) was used as the annealing method. The results confirmed that the proposed method achieves higher detection performance than the widely used linear minimum mean square error (LMMSE) approach (see Figure 2).

Subsequently, the proposed method was implemented at a base station in a wireless communication experimental system, and outdoor over-the-air (OTA) experiments were conducted (see Figure 3). Using the same system parameters as in the simulations, performance was evaluated using both SQA and the D-Wave quantum annealing machine. The experiments demonstrated error-free signal detection for both annealing methods (see Figure 4). Further experiments confirmed successful simultaneous communication with up to 10 devices.

These results demonstrate that the proposed hybrid signal processing method can effectively support the massive connectivity expected in the 6G era, corresponding to a tenfold increase in device density compared with 5G systems.

This achievement represents a major step toward realizing the massive connectivity required in the 6G era and is expected to enable a wide range of machine-type communications involving drones, robots, and XR devices. Going forward, we will continue to advance experimental demonstrations aimed at supporting even larger-scale massive connectivity.

This work was supported by MIC Japan SCOPE (project number: JP235003004).