Asymmetric access point deployment system for non-cellular large-scale MIMO system

Abstract

The invention discloses an asymmetric access point deployment system for a non-cellular large-scale MIMO (Multiple Input Multiple Output) system. The asymmetric access point deployment system comprises L full-function access points TRP, M receiving dedicated access points RP, K user equipment UE and a CPU (Central Processing Unit), the TRP and the RP are connected to a central processing unit (CPU) through a forward transmission network, in an uplink transmission stage, all the TRP and the RP jointly receive uplink signals of the UE to form L + M paths of received signals, and the CPU jointly processes the uplink received signals from the L + M access points (AP); and in the downlink transmission stage, only the L TRP carries out downlink data transmission, and the CPU only distributes downlink transmission data to the L TRP. According to the method, the uplink capacity and coverage are remarkably improved with extremely low hardware and energy consumption cost while the downlink performance is guaranteed.

Description

Technical Field

[0001] This invention relates to the field of wireless communication technology, and in particular to an asymmetric access point deployment system for non-cellular massive MIMO systems. Background Technology

[0002] With the commercial deployment of fifth-generation mobile communication (5G) and the in-depth research on sixth-generation mobile communication (6G), wireless networks are evolving towards ultra-high spectral efficiency, ultra-high connection density, and extremely uniform coverage. Cellular-free massive multiple-input multiple-output (CF-mMIMO) technology, as a key candidate solution to overcome the bottlenecks of traditional cellular architecture, has received widespread attention in recent years. This technology significantly improves system capacity and edge user experience by distributing a large number of low-cost access points (APs) in the service area and coordinating their control with a central processing unit (CPU).

[0003] In traditional CF-mMIMO systems, each access point (AP) typically employs a full-function transceiver design, meaning each AP simultaneously possesses both a radio frequency transmit link and a receive link. In Time Division Duplex (TDD) or Frequency Division Duplex (FDD) modes, it participates in both downlink data transmission and uplink signal reception. However, this symmetrical deployment architecture faces significant challenges in practical applications: uplink performance often becomes the system bottleneck. The fundamental reason is that user units (UEs) (such as mobile phones and IoT terminals) are limited by size, battery capacity, and RF safety regulations, with their maximum transmit power typically only 10-20 dBm. During downlink transmission, the AP can transmit at high power, resulting in uplink spectral efficiency being far lower than downlink. This problem of insufficient uplink coverage and limited data rates is particularly prominent in cell edge environments, high-density user scenarios, or industrial IoT scenarios.

[0004] To improve uplink performance, existing solutions typically enhance spatial diversity and receive gain by increasing the number of access points (APs). However, if the newly added APs are also designed to be fully functional, it will bring several drawbacks:

[0005] (1) Significantly increased cost and power consumption: Each AP needs to be equipped with a complete RF transmission link, including high-cost and high-power components such as power amplifiers, up-conversion mixers, and filters;

[0006] (2) Increased downlink interference: Additional AP participation in downlink cooperation may introduce unnecessary co-channel interference and asynchronous problems, especially under non-ideal channel state information conditions.

[0007] (3) Waste of resources: In scenarios where uplink is limited but downlink capability is sufficient, the transmission capability of the newly added AP is not fully utilized, resulting in hardware redundancy.

[0008] While existing research has proposed energy-saving strategies such as partial AP hibernation and selective activation, these methods are still based on symmetric hardware architectures and do not fundamentally solve the link asymmetry problem of "weak uplink and strong downlink". Furthermore, in emerging application scenarios such as Industrial IoT, intelligent transportation, and high-definition video backhaul, user services exhibit a clear "uplink-intensive" characteristic (e.g., sensor data uploads, vehicle video streaming, UAV telemetry), further highlighting the urgent need for efficient uplink enhancement mechanisms.

[0009] Therefore, there is an urgent need for a new type of non-cellular system architecture that can decouple uplink and downlink hardware resource configurations, and specifically enhance uplink reception capabilities without increasing downlink complexity and interference. By introducing dedicated access points (RPs) with only reception functions, the uplink spatial freedom and reception signal-to-noise ratio can be effectively improved, while avoiding the cost, power consumption, and interference overhead of full-function access points (TRPs), thereby achieving a more efficient, economical, and scalable non-cellular network deployment. Summary of the Invention

[0010] Purpose of the invention: This invention provides an asymmetric access point deployment system for non-cellular massive MIMO systems, which significantly improves uplink capacity and coverage while ensuring downlink performance at extremely low hardware and energy consumption costs.

[0011] Technical Solution: The present invention discloses an asymmetric access point deployment system for non-cellular massive MIMO systems, comprising: L full-function access points (TRPs), M dedicated receiving access points (RPs), K user equipment (UEs), and a central processing unit (CPU). The TRPs and RPs are connected to the CPU via a fronthaul network. During the uplink transmission phase, all TRPs and RPs jointly receive uplink signals from the UEs, forming L+M receiving signals. The CPU performs joint processing on the uplink received signals from the L+M access points (APs). During the downlink transmission phase, only the L TRPs perform downlink data transmission, and the CPU distributes downlink transmission data only to the L TRPs.

[0012] Furthermore, it also includes edge distributed units (EDU) and user-centric distributed units (UCDU). EDUs are deployed near APs to perform local joint signal processing, enhance coverage, and alleviate the computational burden on the CPU. UCDUs perform data stream merging and distribution, as well as resource scheduling and allocation, further supporting the scalability of the system.

[0013] Furthermore, TRP and RP achieve time synchronization through a common clock source.

[0014] Furthermore, the CPU performs joint processing on the uplink received signals from L+M access points (APs), receives uplink baseband signals from L+M APs, and performs joint detection based on the uplink channel state information of L+M APs using maximum ratio combining, zero-forcing reception, or minimum mean square error reception algorithms, demodulating and decoding to obtain the UE's uplink data; or performs local signal detection locally on the L+M APs, sends the detection results to the edge distributed unit (EDU) for joint detection, and sends the results to the upper layer for merging after detection.

[0015] Furthermore, the UE's uplink data is an uplink pilot signal. L TRPs and M RPs receive the pilot signal, and L+M APs estimate the uplink channel state information based on the received pilot signal and upload it to the CPU for joint processing.

[0016] Furthermore, during the downlink transmission phase, only L TRPs transmit downlink data, and the downlink precoding matrix is ​​calculated using zero-forcing precoding, regularized zero-forcing precoding, or maximum ratio transmission algorithm.

[0017] Furthermore, in a time-division duplex system, downlink channel state information is obtained by estimating the uplink pilot signal using channel reciprocity; in a frequency-division duplex system, the UE feeds back downlink channel state information to L TRPs.

[0018] Furthermore, RPs are deployed in coverage blind spots, edge areas, or areas with high uplink traffic demand to achieve efficient uplink coverage enhancement.

[0019] Furthermore, the fronthaul network uses fiber optic fronthaul and supports eCPRI or CPRI protocols; the wireless fronthaul uses millimeter wave or terahertz frequency bands; and the hybrid fronthaul uses fiber optic connections for TRPs and wireless connections for some RPs.

[0020] Beneficial Effects: Compared with existing technologies, this invention has the following significant advantages: Through an asymmetric AP deployment architecture, this invention achieves the design goal of "enhanced uplink and simplified downlink," significantly improving uplink capacity and coverage with extremely low hardware and energy consumption while ensuring downlink performance. By introducing a dedicated access point (RP) with only receiving capabilities, the system can utilize L+M channels to jointly receive signals and enhance the signal-to-noise ratio during the uplink phase, while the downlink phase relies solely on the full-function access point (TRP) for collaborative transmission, avoiding interference and resource waste. This solution is compatible with TDD / FDD systems, supports flexible deployment and dynamic services, and is particularly suitable for typical 6G scenarios with intensive uplink services such as industrial IoT, intelligent transportation, and high-definition video backhaul, effectively alleviating the uplink performance bottleneck caused by limited transmit power of user equipment. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the system structure of the present invention.

[0022] Figure 2 The figure shows the Monte Carlo simulation results of the distributed MMSE receiver used in this invention, comparing the cumulative distribution of uplink spectral efficiency with different numbers of RPs.

[0023] Figure 3 The figure shows the Monte Carlo simulation results of the centralized MMSE receiver used in this invention, comparing the cumulative distribution of uplink spectral efficiency with different numbers of RPs. Detailed Implementation

[0024] In traditional CF-mMIMO systems, each access point (AP) typically features a full-featured transmit / receive design, meaning each AP simultaneously supports both RF transmit and receive links. Whether in Time Division Duplex (TDD) or Frequency Division Duplex (FDD) mode, the AP can simultaneously transmit downlink data and receive uplink signals. However, this symmetrical deployment architecture faces significant challenges in practical applications, particularly as uplink performance often becomes the system bottleneck. The root cause is that user equipment (UEs), such as smartphones and IoT terminals, are limited by their size, battery capacity, and RF safety regulations, typically limiting their maximum transmit power to only 10-20 dBm. In contrast, APs usually have higher transmit power for downlink transmission. This power asymmetry results in significantly lower uplink spectral efficiency compared to downlink. Specifically, APs can transmit downlink signals at high power, achieving higher spectral efficiency and wider coverage on the downlink. However, the limited transmit power of UEs leads to greater constraints on uplink transmission, especially in high-density user areas, industrial IoT scenarios, and cell edges, often resulting in insufficient uplink coverage and limited data rates.

[0025] Furthermore, with the increasing number of users and the diversification of terminal devices, traditional symmetrical deployment methods not only fail to effectively alleviate uplink performance bottlenecks but may also lead to a decrease in total system throughput and service quality, an increase in cost and power consumption, and a waste of resources. Therefore, how to efficiently and specifically improve uplink performance has become the key to enhancing the overall performance of CF-mMIMO systems.

[0026] like Figure 1As shown, an asymmetric access point deployment system for a cellular-free massive MIMO system includes: L full-function transceiver access points (TRPs), M dedicated receiver access points (RPs), and K user equipment (UEs). All TRPs and RPs are connected to a central processing unit (CPU) via a fronthaul network. Each TRP is configured with a complete RF transmit and receive link, including a power amplifier, transmit mixer, low-noise amplifier, receive mixer, digital-to-analog converter (DAC), analog-to-digital converter (ADC), and baseband processing unit. Each RP is configured only with an RF receive link, including a low-noise amplifier, receive mixer, ADC, and baseband receive processing unit, but does not include a power amplifier, transmit mixer, or DAC, and therefore does not have downlink transmission capability.

[0027] During the uplink transmission phase, K UEs send uplink signals to the network. L TRPs and M RPs simultaneously activate their receive links to receive uplink signals from the UEs and upload the received baseband signals to the CPU via the fronthaul network. The CPU aggregates the L+M received signals and, combined with the uplink channel state information estimated by each AP, uses joint detection algorithms such as Maximum Ratio Combining (MRC), Zero Forcing (ZF), or Minimum Mean Square Error (MMSE) to demodulate and decode the uplink data, thereby improving uplink reception performance.

[0028] During the downlink transmission phase, M RPs do not participate in signal transmission and can shut down their RF front-ends or enter a low-power state; only L TRPs perform downlink cooperative transmission based on the precoded data sent by the CPU. The CPU obtains the downlink channel state information from the L TRPs, which is obtained through uplink pilots using channel reciprocity in TDD systems and through UE feedback in FDD systems. It then uses algorithms such as Maximum Ratio Transmission (MRT), ZF, or regularized ZF to calculate the downlink precoding matrix and distributes the precoded data stream to each TRP for synchronous transmission.

[0029] In a TDD system, the TRP and RP synchronize time through a common clock source, ensuring coordinated reception in the uplink time slot and only TRP transmission in the downlink time slot. In an FDD system, the TRP supports both uplink frequency band reception and downlink frequency band transmission, while the RP is only configured with uplink frequency band reception hardware and cannot physically operate in the downlink frequency band.

[0030] The CPU can also dynamically select the optimal uplink receiving set (from L+M APs) and downlink service set (from L TRPs) for each UE based on the UE's location, service type, channel quality, and system load, thus achieving user-centric service. Furthermore, the system can introduce edge distributed units (EDUs) to combine local AP signals, or adopt a hybrid fronthaul architecture (such as TRP fiber connections and partial RP wireless backhaul) to balance performance and deployment costs.

[0031] Figure 2 The figure shows the Monte Carlo simulation results of the distributed MMSE receiver used in this invention, comparing the cumulative distribution of uplink spectral efficiency with different numbers of RPs. Figure 3 The image shows Monte Carlo simulation results of the centralized MMSE receiver used in this invention, comparing the cumulative distribution of uplink spectral efficiency with different numbers of RPs. The results indicate that by employing asymmetric deployment and introducing a dedicated RP for uplink reception, the uplink UE spectral efficiency is significantly improved regardless of whether a distributed or centralized receiver is used, effectively alleviating the uplink performance bottleneck. Through this approach, this invention effectively enhances uplink reception capabilities without increasing downlink complexity, achieving efficient, scalable, and low-cost cellular network deployment.

Claims

1. An asymmetric access point deployment system for non-cellular massive MIMO systems, characterized in that, include: There are L full-function access points (TRPs), M dedicated receiving access points (RPs), K user equipment (UEs), and a central processing unit (CPU). The TRPs and RPs are connected to the CPU via a fronthaul network. During the uplink transmission phase, all TRPs and RPs jointly receive uplink signals from the UEs, forming L+M receiving signals. The CPU performs joint processing on the uplink received signals from the L+M access points (APs). During the downlink transmission phase, only the L TRPs perform downlink data transmission, and the CPU only distributes downlink transmission data to the L TRPs.

2. The asymmetric access point deployment system for non-cellular massive MIMO systems as described in claim 1, characterized in that, It also includes edge distributed units (EDU) and user-centric distributed units (UCDU). EDUs are deployed near APs to perform local joint signal processing.

3. The asymmetric access point deployment system for non-cellular massive MIMO systems as described in claim 2, characterized in that, UCDU performs data stream merging and distribution, as well as resource scheduling and allocation, further supporting system scalability.

4. The asymmetric access point deployment system for non-cellular massive MIMO systems as described in claim 1, characterized in that, TRP and RP synchronize their time using a common clock source.

5. The asymmetric access point deployment system for non-cellular massive MIMO systems as described in claim 1, characterized in that, The CPU performs joint processing on the uplink received signals from L+M access points (APs). It receives uplink baseband signals from L+M APs and performs joint detection based on the uplink channel state information of L+M APs using maximum ratio combining, zero-forcing reception, or minimum mean square error reception algorithms. The CPU then demodulates and decodes the signals to obtain the UE's uplink data. Alternatively, it performs local signal detection on the L+M APs and sends the detection results to the edge distributed unit (EDU) for joint detection. After detection, the results are sent to the upper layer for merging.

6. The asymmetric access point deployment system for non-cellular massive MIMO systems as described in claim 5, characterized in that, The UE's uplink data is an uplink pilot signal. L TRPs and M RPs receive the pilot signal, and L+M APs estimate the uplink channel state information based on the received pilot signal and upload it to the CPU for joint processing.

7. The asymmetric access point deployment system for non-cellular massive MIMO systems as described in claim 1, characterized in that, During the downlink transmission phase, only L TRPs transmit downlink data, and the downlink precoding matrix is ​​calculated using zero-forcing precoding, regularized zero-forcing precoding, or maximum ratio transmission algorithm.

8. The asymmetric access point deployment system for non-cellular massive MIMO systems as described in claim 1, characterized in that, In a time-division duplex system, downlink channel state information is obtained by estimating uplink pilot signals using channel reciprocity; in a frequency-division duplex system, the UE feeds back downlink channel state information to L TRPs.

9. The asymmetric access point deployment system for non-cellular massive MIMO systems as described in claim 1, characterized in that, RPs are deployed in coverage blind spots, edge areas, or areas with high uplink traffic demand to achieve efficient uplink coverage enhancement.

10. The asymmetric access point deployment system for non-cellular massive MIMO systems as described in claim 1, characterized in that, The fronthaul network uses fiber optic fronthaul and supports eCPRI or CPRI protocols; the wireless fronthaul uses millimeter wave or terahertz frequency bands; the hybrid fronthaul uses fiber optic connections for TRPs and wireless connections for some RPs.

Images