Network resource self-optimization method and system for automatically adjusting network slice configuration

Abstract

The application relates to the technical field of 5G network optimization, and discloses a network resource self-optimization method and system for automatically adjusting network slice configuration, which is suitable for dynamic optimization of 5G wireless sub-slices.The method comprises the following steps: a UE measures CSI information and maps physical layer measurement results into a CQI index, and sends the CQI index to a gNB at regular time intervals or according to event triggering; the gNB queries an MCS mapping table according to the CQI index, obtains a preliminary MCS index, and offsets according to slice service requirements; according to the MCS index and the number of allocated RBs, a transport block TBS is obtained, and a physical layer scheduling instruction is generated; the gNB sends data to the UE according to the scheduling instruction; after receiving the data, the UE demodulates and decodes the data, and feeds back data receiving status to the gNB through HARQ; the gNB records the HARQ feedback result, counts the number of continuous successes / failures, and dynamically adjusts the MCS under a lag constraint mechanism according to real-time transmission results.The application realizes differentiated link adaptation through LA and MCS optimization, improves the dynamic adaptation speed, and can realize slice priority conflict processing.

Description

Technical Field

[0001] This invention relates to the field of wireless communication technology, and in particular to dynamic configuration optimization of 5G communication network slices. Specifically, it relates to a network resource self-optimization method and system that automatically adjusts the network slice configuration. Background Technology

[0002] 5G network slicing technology is a core capability for achieving "one network, multiple uses" in 5G. It constructs multiple logically isolated virtual networks on a unified physical infrastructure to meet the differentiated needs of various industry scenarios. A 5G network slice (end-to-end network slice) consists of radio sub-slices, transport sub-slices, and core network sub-slices. The core network sub-slice is the control center of the end-to-end slice, aiming at the virtualization, instantiation, and logical isolation of network functions. The transport sub-slice is responsible for ensuring the path connectivity and quality of service of the end-to-end slice, enabling programmable control of network topology and traffic paths. The radio sub-slice is the access-side entry point for the end-to-end slice, and its core management function is the isolation and dynamic allocation of radio resources.

[0003] Wireless subslices are a key component of dynamic configuration. Through multi-dimensional resource slicing in the time domain (time slots / symbols), frequency domain (RB resource blocks), and spatial domain (beams / antenna ports), dedicated resource pools can be allocated to different services. For example, uRLLC slicing can reserve fixed time-frequency resource blocks (hard isolation), while eMBB slicing employs dynamic resource scheduling (soft isolation). Based on the slice ID (S-NSSAI), data streams are identified at the MAC layer (Media Access Control) and PHY layer (Physical Layer), and logical isolation is achieved through Radio Bearer (DRB) mapping, thus realizing dynamic allocation of resource partitioning and isolation.

[0004] In existing technologies, the Radio Access Network (RAN) controller (such as the CU in a CU / DU architecture) receives resource allocation policies through the Slice Management Function (SMF) and combines this with Radio Resource Management (RRM) algorithms to achieve real-time configuration and scheduling of slices. Existing technologies typically use reinforcement learning (RL)-based methods for dynamic decision-making, modeling the optimization problem of radio subslices as a Markov Decision Process (MDP), and learning the optimal resource scheduling strategy through RL algorithms (such as DQN and PPO). For example, in multi-slice coexistence scenarios, the RL agent can autonomously decide how to allocate RB resources to maximize overall system efficiency. Alternatively, deep learning models such as LSTM and CNN can be used to predict slice traffic trends (such as periodic data transmission in industrial slices) and allocate resources in advance to avoid sudden congestion. For example, predicting that factory equipment will send control commands in 10 seconds can reserve time-frequency resources for uRLLC slices in advance. However, traditional RL dynamic decision-making methods rely on large amounts of training data and iterative optimization, requiring long-term training to converge to the optimal strategy. They lack real-time performance and struggle to cope with the rapidly changing characteristics of wireless channels, especially in multi-slice priority conflict scenarios. The complexity of joint optimization in multi-slice scenarios increases exponentially with the number of slices, resulting in complex decision logic and an inability to directly correlate with physical layer parameters (such as MCS). Deep learning methods, on the other hand, rely on historical data to predict future data. Sudden interference in the wireless environment (such as interference abrupt changes caused by neighbor cell handover) can lead to prediction failure, resulting in improper resource allocation and a lack of dynamic adjustment capabilities in the face of sudden channel changes. Summary of the Invention

[0005] The purpose of this invention is to provide a network resource self-optimization method that automatically adjusts network slice configuration, realizes dynamic resource scheduling optimization of wireless sub-slices based on adaptive adjustment of wireless parameters, and achieves dynamic adaptation to different QoS requirements through link adaptation (LA) and modulation and coding (MCS), maximizing spectrum utilization efficiency while ensuring reliability, and providing physical layer support for end-to-end network slicing.

[0006] According to a first aspect of the present invention, a network resource self-optimization method for automatically adjusting network slice configuration is proposed, comprising the following processes performed at the UE and gNB ends:

[0007] The UE measures signal quality through a reference signal, obtains CSI information, and combines it with a CQI mapping table to convert physical layer measurement results into a CQI index;

[0008] Based on the preset feedback cycle or event trigger, send the CQI index to the gNB;

[0009] gNB queries the MCS mapping table based on the received CQI index to obtain a preliminary MCS index. Each MCS index in the MCS mapping table corresponds to a combination of modulation scheme and coding rate.

[0010] The initial MCS index is offset and adjusted according to the service requirements of the wireless sub-slice;

[0011] Based on the MCS index and the number of allocated resource blocks, the TBS table is queried to obtain the transport block size TBS, and a physical layer scheduling instruction is generated to indicate the modulation and coding scheme of the current transport block. The modulation and coding scheme includes the MCS index, the transport block location, and the transport block size TBS.

[0012] Data is sent to the UE according to the modulation and coding scheme indicated by the scheduling instruction;

[0013] After receiving data, the UE demodulates and decodes it, and then feeds back the data reception status to the gNB via HARQ.

[0014] gNB records HARQ feedback results and counts the number of consecutive successes / failures; and

[0015] Based on the real-time feedback results from HARQ and BLER, the MCS is dynamically adjusted and switched under the hysteresis constraint mechanism.

[0016] According to a second aspect of the present invention, a network resource self-optimization system for automatically adjusting network slice configuration is proposed, comprising a UE and a gNB, wherein:

[0017] At the UE side, the signal quality is first measured by the reference signal to obtain CSI information and combine it with the CQI mapping table to convert the physical layer measurement results into CQI index;

[0018] Then, based on the preset feedback cycle or event trigger, the CQI index is sent to the gNB;

[0019] At the gNB end, the MCS mapping table is first queried based on the received CQI index to obtain the preliminary MCS index. Each MCS index in the MCS mapping table corresponds to a combination of modulation scheme and coding rate.

[0020] Then, the initial MCS index is offset and adjusted according to the service requirements of the wireless sub-slice;

[0021] Subsequently, based on the MCS index and the number of allocated resource blocks, the TBS table is queried to obtain the transport block size (TBS), and a physical layer scheduling instruction is generated, indicating the modulation and coding scheme for the current transport block. This modulation and coding scheme includes the MCS index, transport block location, and transport block size (TBS).

[0022] Data is sent to the UE according to the modulation and coding scheme indicated by the scheduling instruction;

[0023] At the UE end, after receiving data, it demodulates and decodes it, and then feeds back the data reception status to the gNB via HARQ;

[0024] On the gNB side, it records HARQ feedback results and counts the number of consecutive successes / failures; and

[0025] Based on the real-time feedback results from HARQ and BLER, the MCS is dynamically adjusted and switched under the hysteresis constraint mechanism.

[0026] The network resource self-optimization method for automatically adjusting network slice configuration according to the above embodiments of the present invention performs closed-loop optimization based on physical layer parameters (CQI, MCS), and dynamically adjusts the link adaptation (LA) and modulation and coding scheme to achieve rapid adaptation and priority isolation of wireless sub-slice resources, thus making up for the inherent shortcomings of the upper layer algorithm in physical layer optimization.

[0027] Compared with existing technologies, the significant advantages of the network resource self-optimization method for automatically adjusting network slice configuration proposed in this invention are:

[0028] This invention is based on Link Adaptation (LA) to dynamically adjust transmission parameters (such as modulation scheme and coding rate) according to the wireless channel state. By matching channel quality and modulation and coding scheme (MCS), it maximizes spectral efficiency while ensuring transmission reliability. Differentiated link adaptation is achieved through LA and MCS optimization, realizing dynamic adaptation speed improvement and cross-layer scheduling optimization. Furthermore, when multiple slices share resources, low-order MCS is prioritized for URLLC slices to ensure reliability, while high-order MCS is used for eMBB slices on the remaining resources to maximize rate. This achieves slice priority conflict handling and meets the differentiated requirements of each slice for rate, delay, and reliability. Attached Figure Description

[0029] Figure 1 This is a timing flowchart of a network resource self-optimization method for automatically adjusting network slice configuration according to an embodiment of the present invention. Detailed Implementation

[0030] To better understand the technical content of the present invention, specific embodiments are described below in conjunction with the accompanying drawings.

[0031] Various aspects of the invention are described in this disclosure with reference to the accompanying drawings, in which numerous illustrative embodiments are shown. The embodiments of this disclosure are not necessarily intended to encompass all aspects of the invention. It should be understood that the various concepts and embodiments described above, as well as those described in more detail below, can be implemented in any of many ways, because the concepts and embodiments disclosed herein are not limited to any particular implementation. Furthermore, some aspects of the invention disclosed may be used alone or in any suitable combination with other aspects of the invention disclosed.

[0032] {Example 1}

[0033] Combination Figure 1 As shown, the network resource self-optimization method for automatically adjusting network slice configuration according to an embodiment of the present invention includes the following processes performed at the UE end and the gNB end:

[0034] The UE measures signal quality through a reference signal, obtains CSI information, and combines it with a CQI mapping table to convert physical layer measurement results into a CQI index;

[0035] Based on the preset feedback cycle or event trigger, send the CQI index to the gNB;

[0036] gNB queries the MCS mapping table based on the received CQI index to obtain a preliminary MCS index. Each MCS index in the MCS mapping table corresponds to a combination of modulation scheme and coding rate.

[0037] The initial MCS index is offset and adjusted according to the service requirements of the wireless sub-slice;

[0038] Based on the MCS index and the number of allocated resource blocks, the TBS table is queried to obtain the transport block size TBS, and a physical layer scheduling instruction is generated to indicate the modulation and coding scheme of the current transport block. The modulation and coding scheme includes the MCS index, the transport block location, and the transport block size TBS.

[0039] Data is sent to the UE according to the modulation and coding scheme indicated by the scheduling instruction;

[0040] After receiving data, the UE demodulates and decodes it, and then feeds back the data reception status to the gNB via HARQ.

[0041] gNB records HARQ feedback results and counts the number of consecutive successes / failures; and

[0042] Based on the real-time feedback results from HARQ and BLER, the MCS is dynamically adjusted and switched under the hysteresis constraint mechanism.

[0043] As an optional embodiment, the UE measures signal quality through a reference signal to obtain CSI information, including:

[0044] The UE performs power spectrum analysis based on the downlink reference signal to calculate SNR / SINR;

[0045] The CQI index is obtained by querying the mapping table defined by 3GPP based on SNR / SINR.

[0046] In embodiments of the present invention, the Channel Quality Indicator (CQI) refers to the channel quality quantization value measured by the terminal UE and fed back to the base station gNB, which reflects the current transmission conditions of the wireless channel and is directly associated with the highest supported modulation and coding level.

[0047] For example, a terminal UE calculates the signal-to-noise ratio (SNR) or signal-to-interference-plus-noise ratio (SINR) by measuring downlink reference signals (such as CSI-RS, DM-RS).

[0048] Based on the mapping table defined by 3GPP (such as TS 38.214), physical layer measurements are converted into CQI indices (0-15 or higher, depending on the 5G configuration).

[0049] The UE further feeds back the CQI to the base station gNB through the MAC layer (PUCCH / PUSCH) according to periodic reporting (e.g., every 20ms) or event triggering (e.g., channel mutation, SNR change > 3dB), thereby converting the physical layer channel quality into a quantitative indicator that can be used for gNB decision-making and providing a basis for MCS selection.

[0050] In embodiments of the present invention, the signal-to-noise ratio (SNR) / signal-to-interference-plus-noise ratio (SINR) refers to the ratio of signal power to noise / interference power, and is a fundamental physical layer parameter for evaluating channel quality. The UE processes the received signal at the physical layer, separating the useful signal, thermal noise, and neighboring cell interference. The SNR and SINR can then be calculated through power spectral density analysis.

[0051] For example: SNR = signal power / noise power; SINR = signal power / (noise power + interference power).

[0052] In another embodiment, the gNB can also estimate the uplink channel's SNR and SINR using the uplink sounding reference signal (SRS), for example, in a closed-loop LA scenario.

[0053] Based on the obtained CQI, the mapping table defined by 3GPP can be further queried according to SNR / SINR to obtain the CQI index. For example, SNR=15dB can be mapped to CQI=12.

[0054] At the gNB end, the initial selection of MCS is performed based on the received CQI.

[0055] As an optional embodiment, after receiving the CQI feedback from the UE, the gNB can perform filtering processing, such as moving average (moving window filtering), to eliminate jitter caused by short-term fast fading.

[0056] Then, based on the filtered CQI index, the MCS mapping table is queried to obtain the suggested MCS index. For example, based on the 5GNR-based MCS table (as defined in 3GPP TS 38.214), the CQI is mapped to the corresponding MCS index, with each index corresponding to the modulation scheme (QPSK, 16QAM, 256QAM, etc.) and coding rate (e.g., 1 / 3, 5 / 6).

[0057] The MCS indexes (0-28) correspond to a fixed mapping table defined by 3GPP, for example:

[0058] MCS20: 64QAM+ coding rate 0.656, corresponding to a theoretical spectral efficiency of approximately 4.5 bits / s / Hz;

[0059] MCS28: 256QAM+ coding rate 0.877, spectral efficiency approximately 6.0 bit / s / Hz.

[0060] Each MCS index corresponds to a fixed modulation order (e.g., 2). m m is the number of bits per symbol and the coding rate (code rate = information bits / total bits).

[0061] Based on the initial MCS index, the initial MCS index can be further offset and adjusted according to the service requirements of the wireless sub-slice, including:

[0062] Based on the service requirements of wireless sub-slices in URLLC, eMBB, and mMTC scenarios, MCS offset processing is performed according to the slice type priority criterion, where:

[0063] eMBB: Employs a positive offset, ΔMCS = +1 to +2, to increase MCS when channel quality permits;

[0064] URLLC: Uses negative offset, ΔMCS = -2 to -1, to force a reduction in MCS;

[0065] mMTC: Employs a deep negative offset, ΔMCS = -3 to -2, triggering a retransmission strategy;

[0066] Where ΔMCS represents the adjustment value of the initial MCS index based on the slice type, with positive numbers indicating an increase in MCS and negative numbers indicating a decrease in MCS;

[0067] Furthermore, global constraints are applied based on channel quality:

[0068] If the CQI index is less than 5, forward offset of all types of wireless sub-slices is prohibited, and ΔMCS is forced to be 0 to avoid a sudden increase in bit error rate.

[0069] If the CQI index is greater than 10, then for wireless sub-slices in eMBB and URLLC scenarios, the positive and negative offset ranges can be dynamically updated. For example, for eMBB slices, a larger positive offset (ΔMCS = +3) can be allowed, and for URLLC slices, the negative offset can be relaxed (ΔMCS = -1).

[0070] In embodiments of the present invention, for further parameter MCS offset processing, the service characteristics and requirements of different types of slices are described as follows:

[0071] The core requirement of eMBB (enhanced mobile broadband) is high speed, which requires improving spectrum efficiency through high MCS level, while allowing a certain bit error rate (target BLER ≤ 10%).

[0072] The core requirements of URLLC (Ultra-Reliable Low-Latency Communication) are high reliability and low latency, which requires increasing coding redundancy through a low MCS level (target BLER ≤ 10). -5 This reduces the number of retransmissions.

[0073] The core requirements of mMTC (massive machine-type communication) are low bit error rate and low power consumption, which necessitates reducing transmission complexity through a conservative MCS level, with a target BLER ≤ 10. -6 This also reduces power consumption.

[0074] Therefore, in the offset criteria of this invention, a positive offset is used for eMBB slices to increase the MCS when channel quality allows, maximizing the peak rate. A negative offset is used for URLLC slices to forcibly reduce the MCS to increase redundancy, while combining HARQ-ACK feedback for rapid adjustment. For mMTC slices, a deep negative offset combined with repetitive transmission technology is used to ensure low bit error rate and low power consumption.

[0075] As mentioned earlier, offset configuration, by pre-setting offsets to match core service requirements (e.g., using a high MCS for data rate in eMBB, and a low MCS for reliability in URLLC), avoids insufficient or wasted spectrum resource utilization caused by uniformly setting MCS configurations, achieving performance isolation and resource optimization between slices. Simultaneously, combined with dynamic channel quality constraints, it prevents the bit error rate from exploding due to blindly increasing MCS when the channel deteriorates, or from being overly conservative and wasting resources when the channel is good. A balance is struck between service requirements and channel conditions; for example, URLLC can appropriately relax reliability constraints under good channel conditions to improve resource utilization. Conversely, URLLC can sacrifice some reliability to prioritize transmission timeliness when latency is critical.

[0076] Therefore, by combining channel quality and service characteristics, a modulation and coding scheme suitable for the current slice is initially determined, balancing rate and reliability.

[0077] In a preferred embodiment, the MCS offset is also automatically adjusted based on the real-time BLER and the target values ​​of BLER statistics for different slice types:

[0078] eMBB: If real-time BLER > 10% and ΔMCS > 1, ΔMCS will automatically decrease by 1;

[0079] URLLC: If real-time BLER > 10 -5 ΔMCS is automatically decremented by 1, and a retransmission policy is triggered if there are 3 consecutive failures.

[0080] Therefore, the offset is further dynamically fine-tuned based on the actual transmission results, forming a closed-loop control of dynamic adjustment and verification (BLER statistics) to adapt to the time-varying characteristics of the channel. For example, in fast fading channels, eMBB can quickly converge to the optimal MCS through feedback to avoid continuous bit errors.

[0081] As an example, the offset processing for different slice types is as follows:

[0082] eMBB scenario:

[0083] When CQI = 12 (corresponding to MCS = 22), the channel quality is excellent. Applying ΔMCS = +2, the adjusted MCS = 24, and the spectral efficiency increases from 6.75bps / Hz to 7.5bps / Hz, meeting the high-speed requirements.

[0084] URLLC scenario:

[0085] When CQI=10 (corresponding to MCS=19), apply ΔMCS=-2, and after adjustment, MCS=17. The coding redundancy increases, and BLER decreases from 10% to 10^-5. At the same time, fast retransmission is achieved through HARQ feedback to ensure latency <1ms.

[0086] mMTC scenarios:

[0087] When CQI = 8 (corresponding to MCS = 15), applying ΔMCS = -3, the adjusted MCS = 12, combined with three repeated transmissions, reduces BLER to 10. -6 The following measures will simultaneously reduce the transmit power by 3dB, extending the UE's battery life.

[0088] Therefore, based on the offset mechanism driven by slice demand and combined with the dynamic constraints of channel quality, and combined with BLER feedback closed-loop optimization, the MCS configuration and service requirements can be accurately matched.

[0089] As an optional embodiment, the step of querying the TBS table based on the MCS index and the number of allocated resource blocks to obtain the transport block size TBS, and generating a physical layer scheduling instruction to indicate the modulation and coding scheme of the current transport block, includes:

[0090] The gNB determines the number of transport blocks allocated based on the resource quota of different slices. For example, the eMBB slice is allocated 100 RBs.

[0091] Based on the MCS index and the number of allocated transport blocks, query the 3GPP TBS table to obtain the transport block size (TBS), ensuring that the database size matches the coding rate and avoiding transmission efficiency waste; for example, if MCS = 16 and RB = 100, the corresponding TBS = 81888 bits.

[0092] Based on the MCS index, transport block location, and transport block size (TBS), a physical layer scheduling instruction is generated, indicating the modulation and coding scheme of the current transport block.

[0093] It should be understood that Transport Block Size (TBS) represents the amount of data (bits) transmitted at the physical layer in a single transmission, and is determined by the number of MCS and Resource Blocks (RBs). The TBS value can be obtained by looking up the TBS value in the table in 3GPP TS 38.214, by inputting the MCS index and the number of allocated RBs.

[0094] The calculation logic for TBS is as follows (using the MCS index and the number of allocated RBs as input):

[0095] First, look up the modulation order (mod_order) and coding rate (code_rate) based on the MCS index; for example: MCS = 16 → mod_order = 6 (64QAM, 2 6 =64), code_rate=0.656;

[0096] Then, calculate the number of symbols for each RB (number of symbols in a slot, such as 14 symbols / slot);

[0097] Then calculate the number of effective bits = number of RBs × mod_order × number of symbols × code_rate;

[0098] Finally, the number of valid bits is rounded and mapped to the nearest value in the TBS table (the TBS table defined by 3GPP already contains rounding logic).

[0099] Therefore, the MCS is transformed into the actual amount of data that can be transmitted, ensuring that resource utilization efficiency matches transmission reliability.

[0100] Furthermore, the base station gNB sends data according to the scheduling instructions. After receiving the data, the terminal UE demodulates and decodes it, and determines whether the decoding is successful. It then feeds back to the base station gNB via HARQ-ACK / NACK, where ACK indicates success and NACK indicates failure.

[0101] The base station records the HARQ feedback results and counts the number of consecutive successes / failures (e.g., counters cnt_success and cnt_failure).

[0102] This enables the use of a link-layer retransmission mechanism to compensate for instantaneous channel degradation, while providing real-time feedback for dynamic MCS adjustment and enabling dynamic closed-loop optimization.

[0103] As an optional embodiment, the step of dynamically adjusting and switching the MCS based on the real-time feedback results of HARQ and BLER under a hysteresis constraint mechanism includes:

[0104] Based on the data reception status reported by the UE, the gNB uses ACK to indicate successful reception and NACK to indicate reception failure, and combines this with channel quality to determine whether the current MCS is compatible.

[0105] If HARQ-ACK is successful N1 times consecutively and the current MCS has not reached its maximum value, the MCS index is incremented by 1 to improve spectrum efficiency.

[0106] If the difference between the current CQI and the previous one does not exceed ΔCQI, and the BLER statistic is lower than the BLER statistic target value of the corresponding type slice, then MCS boosting is triggered, and the MCS index is incremented by 1.

[0107] If HARQ-NACK fails N2 times consecutively, or if the current CQI decreases by more than ΔCQI compared to the previous one, the MCS index is reduced by 1 to reduce the modulation order or coding rate.

[0108] The hysteresis constraint mechanism is set as follows:

[0109] Configure the threshold range for MCS switching: When the SINR exceeds the uplink threshold of the target MCS index, do not switch immediately, but wait for a duration of ΔT time slots (especially 5-10 time slots) before performing cross-region switching.

[0110] As an optional embodiment, the threshold value N1 for the consecutive successes of HARQ-ACK is configured to be 5. The threshold value N2 for the consecutive failures of HARQ-NACK is configured to be 3.

[0111] As an optional embodiment, boundary verification processing is also included after dynamic adjustment of the MCS:

[0112] If the adjusted MCS exceeds the legal range [0, 28], then the boundary value is taken.

[0113] This allows for the correction of MCS selection based on actual transmission results, avoiding increased bit error rate due to CQI feedback delays or channel abrupt changes, thus achieving closed-loop optimization. Simultaneously, a hysteresis mechanism is incorporated to prevent frequent switching, such as MCS oscillations caused by small channel fluctuations.

[0114] {Example 2}

[0115] Combination Figure 1 Based on the method execution shown in the above embodiments, an embodiment of the present invention also proposes a network resource self-optimization system for automatically adjusting network slice configuration, including UE and gNB.

[0116] At the UE end, the signal quality is first measured by the reference signal to obtain CSI information and combine it with the CQI mapping table to convert the physical layer measurement results into CQI index;

[0117] Then, based on the preset feedback cycle or event trigger, the CQI index is sent to the gNB.

[0118] At the gNB end, the MCS mapping table is first queried based on the received CQI index to obtain the preliminary MCS index. Each MCS index in the MCS mapping table corresponds to a combination of modulation scheme and coding rate.

[0119] Then, the initial MCS index is offset and adjusted according to the service requirements of the wireless sub-slice;

[0120] Subsequently, based on the MCS index and the number of allocated resource blocks, the TBS table is queried to obtain the transport block size (TBS), and a physical layer scheduling instruction is generated, indicating the modulation and coding scheme for the current transport block. This modulation and coding scheme includes the MCS index, transport block location, and transport block size (TBS).

[0121] Data is sent to the UE according to the modulation and coding scheme indicated by the scheduling instructions.

[0122] Furthermore, at the UE end, after receiving the data sent by the gNB, it demodulates and decodes it, and then feeds back the data reception status to the gNB via HARQ.

[0123] Furthermore, at the gNB end, it records the HARQ feedback results, counts the number of consecutive successes / failures, and dynamically adjusts and switches the MCS under the hysteresis constraint mechanism based on the real-time feedback results of HARQ and BLER.

[0124] It should be understood that the above-mentioned dynamic adjustment and switching of MCS based on the real-time feedback results of HARQ and BLER, under the hysteresis constraint mechanism, includes:

[0125] Based on the data reception status reported by the UE, the gNB uses ACK to indicate successful reception and NACK to indicate reception failure, and combines this with channel quality to determine whether the current MCS is compatible.

[0126] If HARQ-ACK is successful N1 times consecutively and the current MCS has not reached its maximum value, the MCS index is incremented by 1 to improve spectrum efficiency.

[0127] If the difference between the current CQI and the previous one does not exceed ΔCQI, and the BLER statistic is lower than the BLER statistic target value of the corresponding type slice, then MCS boosting is triggered, and the MCS index is incremented by 1.

[0128] If HARQ-NACK fails N2 times consecutively, or if the current CQI decreases by more than ΔCQI compared to the previous one, the MCS index is reduced by 1 to reduce the modulation order or coding rate.

[0129] The hysteresis constraint mechanism is set as follows:

[0130] Configure the threshold range for MCS switching: When the SINR exceeds the uplink threshold of the target MCS index, do not switch immediately, but wait for a duration of ΔT time slots (especially 5-10 time slots) before performing cross-region switching.

[0131] Furthermore, after dynamic adjustment of the MCS, boundary verification processing is also included:

[0132] If the adjusted MCS exceeds the legal range [0, 28], then the boundary value is taken.

[0133] This allows for the correction of MCS selection based on actual transmission results, avoiding increased bit error rate due to CQI feedback delays or channel abrupt changes, thus achieving closed-loop optimization. Simultaneously, a hysteresis mechanism is incorporated to prevent frequent switching, such as MCS oscillations caused by small channel fluctuations.

[0134] While the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the invention. Those skilled in the art can make various modifications and refinements without departing from the spirit and scope of the invention. Therefore, the scope of protection of the present invention shall be determined by the claims.

Claims

1. A method for automatically adjusting network slice configurations to optimize network resources, characterized in that, This includes the following procedures performed at both the UE and gNB ends: The UE measures signal quality through a reference signal, obtains CSI information, and combines it with a CQI mapping table to convert physical layer measurement results into a CQI index; Based on the preset feedback cycle or event trigger, send the CQI index to the gNB; gNB queries the MCS mapping table based on the received CQI index to obtain a preliminary MCS index. Each MCS index in the MCS mapping table corresponds to a combination of modulation scheme and coding rate. The initial MCS index is offset and adjusted according to the service requirements of the wireless sub-slice; Based on the MCS index and the number of allocated resource blocks, the TBS table is queried to obtain the transport block size TBS, and a physical layer scheduling instruction is generated to indicate the modulation and coding scheme of the current transport block. The modulation and coding scheme includes the MCS index, the transport block location, and the transport block size TBS. Data is sent to the UE according to the modulation and coding scheme indicated by the scheduling instruction; After receiving data, the UE demodulates and decodes it, and then feeds back the data reception status to the gNB via HARQ. gNB records HARQ feedback results and counts the number of consecutive successes / failures. as well as Based on the real-time feedback results from HARQ and BLER, the MCS is dynamically adjusted and switched under the hysteresis constraint mechanism. The step of adjusting the initial MCS index based on the service requirements of the wireless sub-slice includes: Based on the service requirements of wireless sub-slices in URLLC, eMBB, and mMTC scenarios, MCS offset processing is performed according to the slice type priority criterion, where: eMBB: Employs a positive offset, ΔMCS = +1 ~ +2, to increase MCS when channel quality permits; URLLC: Uses negative offset, ΔMCS = -2 ~ -1, to force a reduction in MCS; mMTC: Employs a deep negative offset, ΔMCS = -3 ~ -2, triggering a retransmission strategy; The ΔMCS represents the adjustment value of the initial MCS index based on the slice type; a positive number indicates an increase in MCS, and a negative number indicates a decrease in MCS. Furthermore, global constraints are applied based on channel quality: If the CQI index is less than 5, then all types of wireless sub-slices are prohibited from forward offset, and ΔMCS is forced to be set to 0. If the CQI index is greater than 10, then for wireless sub-slices in eMBB and URLLC scenarios, dynamic updates of the positive and negative offset ranges are allowed.

2. The network resource self-optimization method for automatically adjusting network slice configuration according to claim 1, characterized in that, The UE measures signal quality using a reference signal and obtains CSI information, including: The UE performs power spectrum analysis based on the downlink reference signal to calculate SNR / SINR; The CQI index is obtained by querying the mapping table defined by 3GPP based on SNR / SINR.

3. The network resource self-optimization method for automatically adjusting network slice configuration according to claim 1, characterized in that, The offset adjustment of the initial MCS index based on the service requirements of the wireless sub-slice also includes: The MCS offset is automatically adjusted based on real-time BLER and BLER statistical target values ​​for different slice types: eMBB: If real-time BLER > 10% and ΔMCS > 1, ΔMCS will automatically decrease by 1; URLLC: If real-time BLER > 10 -5 ΔMCS is automatically decremented by 1, and a retransmission policy is triggered if there are 3 consecutive failures.

4. The network resource self-optimization method for automatically adjusting network slice configuration according to claim 1, characterized in that, The step of querying the TBS table based on the MCS index and the number of allocated resource blocks to obtain the transport block size (TBS), and generating a physical layer scheduling instruction to indicate the modulation and coding scheme of the current transport block, includes: The gNB determines the number of transport blocks allocated based on the resource quotas of different slices; Based on the MCS index and the allocated number of transport blocks, query the 3GPP TBS table to obtain the transport block size (TBS), ensuring that the database size matches the coding rate; and Based on the MCS index, transport block location, and transport block size (TBS), a physical layer scheduling instruction is generated, indicating the modulation and coding scheme of the current transport block.

5. The network resource self-optimization method for automatically adjusting network slice configuration according to claim 1, characterized in that, The dynamic adjustment and switching of MCS based on the real-time feedback results of HARQ and BLER under the hysteresis constraint mechanism includes: Based on the data reception status reported by the UE, the gNB uses ACK to indicate successful reception and NACK to indicate reception failure, and combines this with channel quality to determine whether the current MCS is compatible. If HARQ-ACK is successful N1 times consecutively and the current MCS has not reached its maximum value, the MCS index is incremented by 1 to improve spectrum efficiency. If the difference between the current CQI and the previous one does not exceed ΔCQI, and the BLER statistic is lower than the BLER statistic target value of the corresponding type slice, then MCS boosting is triggered, and the MCS index is incremented by 1. If HARQ-NACK fails N2 times consecutively, or if the current CQI decreases by more than ΔCQI compared to the previous one, the MCS index is reduced by 1 to reduce the modulation order or coding rate. The hysteresis constraint mechanism is set as follows: Configure the threshold range for MCS switching: When the SINR exceeds the uplink threshold of the target MCS index, the switching will not be performed immediately, but will wait for a duration of ΔT time slots before the cross-region switching is performed.

6. The network resource self-optimization method for automatically adjusting network slice configuration according to claim 5, characterized in that, The threshold value N1 for the consecutive successes of HARQ-ACK is configured to be 5.

7. The network resource self-optimization method for automatically adjusting network slice configuration according to claim 5, characterized in that, The threshold value N2 for the consecutive number of HARQ-NACK failures is configured to be 3.

8. The network resource self-optimization method for automatically adjusting network slice configuration according to claim 5, characterized in that, After dynamic adjustment of the MCS, boundary verification processing is also included: If the adjusted MCS exceeds the legal range [0, 28], then the boundary value is taken.

9. A network resource self-optimization system that automatically adjusts network slice configuration, characterized in that, Including UE and gNB, where: At the UE end, the signal quality is first measured by the reference signal to obtain CSI information and combine it with the CQI mapping table to convert the physical layer measurement results into CQI index; Then, based on the preset feedback cycle or event trigger, the CQI index is sent to the gNB; At the gNB end, the MCS mapping table is first queried based on the received CQI index to obtain the preliminary MCS index. Each MCS index in the MCS mapping table corresponds to a combination of modulation scheme and coding rate. Then, the initial MCS index is offset and adjusted according to the service requirements of the wireless sub-slice; Subsequently, based on the MCS index and the number of allocated resource blocks, the TBS table is queried to obtain the transport block size (TBS), and a physical layer scheduling instruction is generated, indicating the modulation and coding scheme for the current transport block. This modulation and coding scheme includes the MCS index, transport block location, and transport block size (TBS). Data is sent to the UE according to the modulation and coding scheme indicated by the scheduling instruction; At the UE end, after receiving data, it demodulates and decodes it, and then feeds back the data reception status to the gNB via HARQ; On the gNB side, it records HARQ feedback results and counts the number of consecutive successes / failures; and Based on the real-time feedback results from HARQ and BLER, the MCS is dynamically adjusted and switched under the hysteresis constraint mechanism. The step of adjusting the initial MCS index based on the service requirements of the wireless sub-slice includes: Based on the service requirements of wireless sub-slices in URLLC, eMBB, and mMTC scenarios, MCS offset processing is performed according to the slice type priority criterion, where: eMBB: Employs a positive offset, ΔMCS = +1 ~ +2, to increase MCS when channel quality permits; URLLC: Uses negative offset, ΔMCS = -2 ~ -1, to force a reduction in MCS; mMTC: Employs a deep negative offset, ΔMCS = -3 ~ -2, triggering a retransmission strategy; The ΔMCS represents the adjustment value of the initial MCS index based on the slice type; a positive number indicates an increase in MCS, and a negative number indicates a decrease in MCS. Furthermore, global constraints are applied based on channel quality: If the CQI index is less than 5, then all types of wireless sub-slices are prohibited from forward offset, and ΔMCS is forced to be set to 0. If the CQI index is greater than 10, then for wireless sub-slices in eMBB and URLLC scenarios, dynamic updates of the positive and negative offset ranges are allowed.

Images