CN-122028126-A - 5G uplink coverage enhancement method

Abstract

The invention relates to a 5G uplink coverage enhancement method which comprises the steps of collecting data of a 5G network, comprising RSRP, SINR, txPower, service types, uplink loads, switching success times, switching failure times, transmission rate and block error rate, dividing the data into a training set and a testing set, constructing a multi-dimensional threshold SUL trigger model based on a deep neural network, training the multi-dimensional threshold SUL trigger model by using the training set, constructing a three-layer framework comprising a global load sensing layer, a regional resource scheduling layer and a terminal dynamic adaptation layer, improving a dynamic time slot matching algorithm, inputting the testing set into the trained multi-dimensional threshold SUL trigger model, and outputting uplink parameter configuration by combining the dynamic time slot matching algorithm to enhance the 5G uplink coverage. Compared with the prior art, the method has the advantages of reducing misjudgment of SUL switching, realizing uplink coverage enhancement and the like.

Inventors

• LIU CHUANZHI
• JIANG XIANG
• JIN YAN
• HUA FENG
• XI CHENG
• YU HONGFAN
• SHAO YI
• REN ZIYU
• GUO YANYAN
• WU DONGYAN
• CAI TIANCHENG
• XU LIQUN
• PENG YAN
• LIU QING
• YANG JUN
• JIANG HAO
• CHEN ZIYU
• TANG JIA
• TIAN YUAN
• WANG JIANYONG
• WU JIAHUI
• WANG BIN
• DING LAIGEN
• CHEN LINXING
• LV ZHONG
• ZHU YILUN
• Gu Zhuoyue
• Lang Yunjing
• ZHANG WEI
• YUE JIAMIN
• XU BING
• Xi Shunjin
• MA XIAOHONG
• QIN YI
• ZHU YE
• GU ZHENNING
• XU XIAOMING
• HUANG YIJIE
• XU WEI
• LU WEIWEN
• FAN JINGCHUN
• WEI MIN
• QIN LING
• ZHU XIAOLIANG
• SONG WENBIN
• MA XIAOYAN
• LU WEIMIN
• XU WEIDONG
• LI GU
• GU MIAO
• SONG XUECHENG
• LIANG YING
• WU MINJUAN
• ZHU JUN
• JIN GANG
• SUN DANQING
• JIANG QIANG
• XU MIN
• ZHANG ZHOU
• NIU XIAOJUN
• LIU FANGHUI
• LI YUANZHI
• XU FENG
• GUO LIPING
• Shen Daina
• XU BIN

Assignees

• 上海共联通信信息发展有限公司

Dates

Publication Date

20260512

Application Date

20251231

Claims (10)

1. 1. The 5G uplink coverage enhancement method is characterized by comprising the following steps: Collecting 5G network data, including RSRP, SINR, txPower, service types, uplink load, switching success times, switching failure times, transmission rate and block error rate, and dividing the data into a training set and a testing set; Constructing a multi-dimensional threshold SUL trigger model based on a deep neural network, wherein the deep neural network comprises an input layer, a 3-layer hidden layer and an output layer, the dimension of the input layer corresponds to RSRP, SINR, path loss and uplink load in acquired 5G network data, and parameter weights of all layers are optimized through a back propagation algorithm, and the multi-dimensional threshold comprises an RSRP threshold Pth, a PL threshold Lth and an SINR compensation threshold Sth; Constructing a three-layer architecture comprising a global load sensing layer, a regional resource scheduling layer and a terminal dynamic adaptation layer, and improving a dynamic time slot matching algorithm; Inputting the test set into a trained multi-dimensional threshold SUL trigger model, and outputting the parameter configuration of the uplink by combining an improved dynamic time slot matching algorithm, wherein the parameter configuration comprises an RSRP threshold Pth, an SINR compensation threshold Sth, whether to trigger SUL access and time slot configuration, and the 5G uplink coverage is enhanced by the parameter configuration.
2. 2. The 5G uplink coverage enhancement method according to claim 1, wherein the multi-dimensional threshold SUL trigger model dynamically adjusts the multi-dimensional threshold by establishing a threshold adjustment mapping table based on uplink load and service type related data in a training set, wherein the mapping table comprises RSRP threshold Pth and SINR compensation threshold Sth corresponding to different uplink load intervals and different service types; and the output result of the multi-dimensional threshold SUL trigger model is a binary decision value, 1 indicates that SUL access is triggered, and 0 indicates that SUL carrier is exited or the original access mode is maintained.
3. 3. The method for enhancing 5G uplink coverage according to claim 1, wherein when the UE uses the mid-high band, the SUL carrier access is triggered if any one of the following conditions is satisfied and the duration is greater than the TTT: the first condition is that RSRP is less than or equal to RSRP threshold Pth and PL is more than or equal to PL threshold Lth; and the second condition is that SINR is less than or equal to SINR compensation threshold Sth.
4. 4. The method for enhancing 5G uplink coverage according to claim 1, wherein when the SUL carrier is used by the UE, the SUL carrier is exited and the medium-high band access is returned if the following two conditions are satisfied at the same time: A third condition is RSRP > RSRP threshold Pth or PL < PL threshold Lth; fourth condition, SINR > SINR compensation threshold Sth.
5. 5. The 5G uplink coverage enhancement method according to claim 1, wherein the multidimensional threshold SUL trigger model classifies uplink loads, and implements differentiated RSRP thresholds Pth for uplink loads of different classes, specifically: If the uplink load is smaller than or equal to the first uplink load UL1, maintaining a default RSRP threshold Pth; the uplink load is located between the first uplink load UL1 and the second uplink load UL2, and the RSRP threshold Pth is increased by 3dB; And if the uplink load is greater than or equal to the second uplink load UL2, the SUL is forcedly triggered, and the limit of the SINR compensation threshold Sth is ignored.
6. 6. The method for enhancing 5G uplink coverage according to claim 1, wherein the SINR compensation threshold Sth is used to implement differentiated triggering SUL in combination with a service type, specifically: uRLLC traffic corresponds to a first SINR compensation threshold; eMBB traffic corresponds to a second SINR compensation threshold; mMTC traffic corresponds to a third SINR compensation threshold; Wherein the first SINR compensation threshold > the second SINR compensation threshold > the third SINR compensation threshold.
7. 7. The method for enhancing 5G uplink coverage according to claim 1, wherein the global load sensing layer periodically monitors a downlink traffic ratio α on the whole network, and starts an uplink coverage enhancement mode when α is greater than a traffic ratio threshold, and adjusts a timeslot ratio to 2dl:2ul; the regional resource scheduling layer is arranged in the SUL carrier coverage area of the hot spot region, and additionally increases 1 UL time slot; The terminal dynamic adaptation layer is bound with the service type output by the multi-dimensional threshold SUL trigger model, and the granularity of time slot configuration is dynamically adjusted.
8. 8. The method for enhancing 5G uplink coverage according to claim 7, wherein the dynamically adjusting granularity of the slot configuration comprises: to eMBB business, 2 UL time slots and 1 flexible time slot are allocated continuously; reserving a fixed UL time slot and an emergency time slot request channel for uRLLC services; For mMTC traffic, 1 UL slot is allocated using time division multiplexing.
9. 9. The method of claim 1, wherein the method further comprises dynamically adjusting the number of available UL slots according to the real-time SINR, specifically: , , Wherein SL actual is the actual time slot number; SL basic is the basic time slot number, which is preset by the terminal dynamic adapting layer according to the service type, SL borrow is the dynamic borrowing time slot, the actual available number is decided by the global load sensing layer; the SUL carrier SINR value reported in real time for the UE; the threshold Sth is compensated for the SINR corresponding to the current traffic type, An ideal SINR upper limit for the SUL carrier; The link quality correction coefficient is formed, the value range is [ -1,1], beta is the time slot enhancement factor, which is dynamically mapped by the uplink load level, gamma is the borrowed time slot effectiveness coefficient, which is related to the region type, delta is the BLER attenuation coefficient, And forming a link reliability correction term, wherein BLER is the block error rate reported by the UE in real time.
10. 10. The method for 5G uplink coverage enhancement as set forth in claim 1, wherein, the method further comprises improving uplink quality closed loop control, including: the UE periodically reports SINR and BLER of the SUL carrier; The base station dynamically adjusts time slot allocation according to BLER, namely increasing 1 UL time slot and reducing 1-2 level MCS level if BLER is larger than a first BLER threshold value, and reducing 1 UL time slot and increasing 1-3 level MCS level if BLER is smaller than a second BLER threshold value, wherein the first BLER threshold value is larger than the second BLER threshold value; and predicting resource requirements in a future period by a linear regression algorithm according to the historical record at the core network side, and reserving UL time slot resources of 2 SUL carriers in advance if no additional downlink resource requirements exist in the future period, so as to compensate the transmission loss of an uplink coverage edge region.

Description

5G uplink coverage enhancement method Technical Field The invention relates to the technical field of 5G mobile communication, in particular to a 5G uplink coverage enhancement method. Background The 5G mobile communication system faces the challenge of unbalanced uplink and downlink coverage while promoting the development of communication technology, and especially has a defect in uplink coverage. With the development of service diversification, such as ultra-high definition video communication, big data acquisition, intelligent monitoring and the like, higher requirements on uplink capacity and coverage are provided. The main stream commercial frequency band (such as 3.5 GHz) of 5G adopts a Time Division Duplex (TDD) mode, and the downlink greatly improves the capacity and coverage capacity by introducing MassiveMIMO and other technologies and by means of accurate beam forming and scanning. However, due to the frequency band characteristics and uneven uplink and downlink time slot ratio in the TDD system, the uplink and downlink coverage of the C-BandTDD system is unbalanced, and the uplink coverage capability and capacity of the 3.5GHz 5G network need to be improved. Meanwhile, with the development of services such as mobile internet and internet of things, the requirement for uploading mass data is rapidly increased, and a severe test is provided for the 5G uplink performance. The 5G mobile network aggravates the problem of unbalanced uplink and downlink due to technical characteristics of high frequency bands (such as Sub-6GHz and millimeter waves), flexible time slot proportion, large-scale MIMO and the like, and is specifically expressed as follows: 1. The transmission loss of the high frequency band is large, and the millimeter wave path loss is about 18dB (1 km distance) higher than the Sub-6 GHz; 2. The proportion of the time slots is unbalanced, namely the proportion of the downlink time slots is high (for example, 8:2), and the uplink transmission opportunity of the terminal is reduced; The difference of the massive MIMO gains is that the base station multi-antenna beam forming improves the downlink gain, but the number of the terminal antennas is limited; 4. The uplink service demand is increased, and the application such as video live broadcast and the like needs continuous high-quality uplink coverage. The existing scheme does not quantitatively analyze uplink and downlink coverage differences of different frequency bands, and lacks a technical means for systematically enhancing uplink coverage. Through searching, chinese patent application publication No. CN116056224A discloses an uplink coverage enhancement method in a 5G mobile communication system, which comprises the steps of configuring a working frequency band which can be used by a terminal, carrying out link budget to obtain the maximum allowed path loss, sending a detection signal in a high frequency band, detecting to obtain signal-to-noise ratio and receiving power information, calculating the proportion and path loss of tap energy of a direct path accounting for the total energy of all paths, selecting an uplink data transmission frequency band according to an initial detection result and timely feeding back the selection result to a user, tracking the uplink coverage condition under the high frequency band communication, sending the detection signal again when an initial detection time interval reaches a period, and carrying out frequency band switching by the user according to the detection condition by the base station. The prior patent application has the problems that only path loss and SINR are used for triggering SUL, SINR is not considered in combination with service types, and frequency band switching misjudgment and service types are not considered in complex scenes. How to realize the enhancement of 5G uplink coverage and compensate the imbalance of uplink and downlink coverage becomes a technical problem to be solved. Disclosure of Invention The invention aims to overcome the defects of the prior art and provide a 5G uplink coverage enhancement method. The aim of the invention can be achieved by the following technical scheme: according to one aspect of the present invention, there is provided a 5G uplink coverage enhancement method, including: Collecting 5G network data, including RSRP, SINR, txPower, service types, uplink load, switching success times, switching failure times, transmission rate and block error rate, and dividing the data into a training set and a testing set; Constructing a multi-dimensional threshold SUL trigger model based on a deep neural network, wherein the deep neural network comprises an input layer, a 3-layer hidden layer and an output layer, the dimension of the input layer corresponds to RSRP, SINR, path loss and uplink load in acquired 5G network data, and parameter weights of all layers are optimized through a back propagation algorithm, and the multi-dimensional threshold comprises an