CN104486772B - A kind of normalized LTE network evaluation optimal sysytem of end-to-end various dimensions - Google Patents

Abstract

The invention discloses a kind of end-to-end normalized LTE network evaluation optimal sysytems of various dimensions, including transacter, collected data type includes that soft data, north orientation MR data, the LTE of adopting of LTE adopts data, frequency sweep data, drive test and ATU data, OMC performance statistic and network element device basic data firmly；Data integrity assesses device, and the integrality for the data to collection is assessed；Model evaluation device, the model evaluation device includes the data evaluation module that collected data are carried out with classification assessment, and the corresponding data evaluation module of different types of data, corresponding data evaluation module respectively assess to obtain assessed value corresponding data type；Normalization assessment device, the normalization assessment device are that the data evaluation module of corresponding data type is provided with corresponding weight, and can obtain normalized value by weight and assessed value.Assessment result of the invention has regression nature, continuity and trackability.

Description

A kind of normalized LTE network evaluation optimal sysytem of end-to-end various dimensions

Technical field

The present invention relates to wireless communication field, especially a kind of normalized LTE network assessment optimization of end-to-end various dimensions System.

Background technique

LTE is since China's Mainland deployment is commercial so far without a kind of intelligent automatic normalized appraisal procedure.Net Network operation management person is to LTE network covering level, annoyance level, resource efficiency, network performance, network load, terminal perception etc. Lack unification, comprehensively understand and grasp, so that there are the movements such as resource investment is shot at random, the network optimization is palliative The phenomenon that common carrier is unwilling to see.

LTE network planning leans on experience by emulation, optimization；Customer complaint is leaned in base station investment construction direction；Analyze data mining Drive test is depended merely on, north orientation MR, adopts firmly, soft one of the data sources such as adopt.The emulation of LTE planning is from true also relatively large distance.It is excellent The experience of change lacks uniformity and cannot be widely applied.Drive test data can acquisition range it is limited, and labor intensity is high, efficiency It is low；The interference of north orientation MR data deficiency downlink and the relevant data sample of scheduling of resource, cannot reflect Network status comprehensively；Number is adopted firmly According to only focusing on core network service face；Soft data of adopting stress wireless side, pay close attention to control plane.The above technology cannot be end-to-end more Dimension uniformly carries out the assessment optimization of LTE network comprehensively, moreover, different time difference districts and cities different technologies human assessment As a result all different, optimization means will appear repeatedly with inconsistent situation, not have a regression nature, continuity and traceable Property.

Summary of the invention

To solve the above problems, the purpose of the present invention is to provide a kind of end-to-end normalized LTE networks of various dimensions to comment Estimate optimization system, assessment optimization is carried out to LTE network by the way of unified, and makes assessment result that there is regression nature, continuity And trackability.

Technical solution used by the present invention solves the problems, such as it is:

A kind of normalized LTE network evaluation optimal sysytem of end-to-end various dimensions, comprising:

Transacter, collected data type include LTE it is soft adopt data, north orientation MR data, LTE adopt firmly data, Frequency sweep data, drive test and ATU data, OMC performance statistic and network element device basic data；

Data integrity assesses device, and the integrality for the data to collection is assessed；

Model evaluation device, the model evaluation device include that the data for carrying out classification assessment to collected data are commented Estimate module, and the corresponding data evaluation module of different types of data, corresponding data evaluation module is respectively to corresponding data Type is assessed to obtain assessed value；

Normalization assessment device, the normalization assessment device are that the data evaluation module of corresponding data type is provided with phase The weight answered, and normalized value can be obtained by weight and assessed value.

Further, further includes:

Output module is assessed, for exporting assessment result and corresponding Optimizing Suggestions.

Further, the transacter establishes data source using multidimensional normalizing weighted model and collects corresponding number According to.

Further, data integrity assessment device base station corresponding to the data of some data type of acquisition is small When the ratio of the area base station cell Zhan Zong is less than setting ratio, it will be prompted to transacter and continue to acquire the data of the type, or Person carries out single station assessment to acquired base station cell by model evaluation device.

Further, the model evaluation device includes corresponding network structure reasonable evaluation module, covering proficiency assessment Module, annoyance level evaluation module, network load level evaluation module, energy efficiency evaluation module, System Performance Analysis module Evaluation module is perceived with terminal user.

The beneficial effects of the present invention are:

The present invention uses a kind of end-to-end normalized LTE network evaluation optimal sysytem of various dimensions, weights mould using normalizing Type establishes data source and collects corresponding data, and when meeting condition, assesses data, and evaluation process uses unified mould Type assesses device, and model evaluation device includes data evaluation module corresponding with categorical data, and assessment result can be made to have and returned One property, continuity and trackability；Scope of assessment covers the various aspects of LTE network, comprehensive can comment LTE network Estimate, comprehensive reaction network situation.

Detailed description of the invention

The invention will be further described with example with reference to the accompanying drawing.

Fig. 1 is the structural schematic diagram of system of the present invention.

Specific embodiment

Embodiment 1:

Shown in referring to Fig.1, a kind of end-to-end normalized LTE network evaluation optimal sysytem of various dimensions of the invention, comprising:

Transacter, collected data type include LTE it is soft adopt data, north orientation MR data, LTE adopt firmly data, Frequency sweep data, drive test and ATU data, OMC performance statistic and network element device basic data；

Data integrity assesses device, and the integrality for the data to collection is assessed；

Model evaluation device, the model evaluation device include that the data for carrying out classification assessment to collected data are commented Estimate module, and the corresponding data evaluation module of different types of data, corresponding data evaluation module is respectively to corresponding data Type is assessed to obtain assessed value；

Normalization assessment device, the normalization assessment device are that the data evaluation module of corresponding data type is provided with phase The weight answered, and normalized value can be obtained by weight and assessed value；

Output module is assessed, for exporting assessment result and corresponding Optimizing Suggestions.

Wherein, the transacter establishes data source using multidimensional normalizing weighted model and collects corresponding data, Data source include eat dishes without rice or wine, it is soft adopt, adopt firmly, network management, test, the whole network terminal survey report etc..LTE is soft adopt data, north orientation MR data, LTE adopts the data such as data, frequency sweep data, drive test and ATU data, OMC performance statistic and network element device basic data firmly Acquisition can be acquired according to corresponding acquisition principle, can be in assessment of the invention through distinct methods data collected It is applied, but the determination of different application form decides assessment result.

When assessing data, different data volumes must satisfy certain requirement, and data are not complete then cannot be to LTE The truth of network carries out comprehensive total evaluation.Data integrity assesses device can be according to the requirement limitation pair of corresponding setting Corresponding data carries out integrity assessment, and when corresponding data do not meet integrity demands, continues to adopt by transacter Collection.But some data are under different conditions, assessment can also be normalized by the assessment to single base station.Usual situation Under, the integrality of data is the base station cell base station cell Zhan Zong according to corresponding to the data of some data type collected What ratio was determined, when it is less than setting ratio, then illustrate imperfect.For example, for north orientation MR data, it is desirable that it is acquired When being lower than 65% to the base station cell accountings of data, it will be prompted to operator and continue to acquire data, provide and need the base station cell that acquires Inventory.

In order to fully assess to above data, model evaluation device is provided with different according to the evaluation index of setting Evaluation module, the evaluation index of usual LTE include network structure reasonability, covering level, annoyance level, network load level, Energy efficiency, system performance and terminal user's perception etc., model evaluation device is respectively provided with corresponding evaluation module and is commented Estimate, including corresponding network structure reasonable evaluation module, covering proficiency assessment module, annoyance level evaluation module, network are negative It carries proficiency assessment module, energy efficiency evaluation module, System Performance Analysis module and terminal user and perceives evaluation module etc., with reality Regression nature, continuity and the trackability of existing evaluation structure；The various aspects of the above scope of assessment covering LTE network, can be entirely square LTE network is assessed in position, comprehensive reaction network situation.

Model evaluation device can obtain a corresponding assessed value, normalization assessment device according to different evaluation modules The data evaluation module of corresponding data type is provided with corresponding weight, and normalization can be obtained by weight and assessed value Value.Assessment output module can export assessment result, and assessment result includes each LTE net obtained to model evaluation device The assessed value of network index, according to corresponding assessed value, corresponding Optimizing Suggestions can be exported by assessing output module.

Embodiment 2:

Specifically to be explained to the present invention, following text based on north orientation MR data, network management data, drive test data model Algorithm carries out detailed paraphrase to the appraisal procedure of corresponding evaluation index.Specific data in following text are the present invention in reality Application data in border, the value etc. of setting, specific number including coefficient are stated for sake of convenience with assuming, setting etc..

The integrality of north orientation MR data is assessed first, when collecting the base station cell accountings of data lower than 65%, then Continue to acquire, the present invention meets the condition herein；

Later, covering level normalization assessment is carried out, comprising:

1. the sum of calculating serving cell or adjacent area RSRP > -100dBm MR account for the ratio of all north orientation MR, which is A kind of universal method of covering is assessed, it can be horizontal from a dimension assessment network coverage；

2. calculating the cell number of weak coating ratio > 50%: removing the website that MR total sample number is less than 500, cover difference TOPN cell number accounts for the ratio of whole cells, the foundation as network coverage level normalization assessment deduction；

3. calculating the cell number of weak coating ratio < 25%: removing the website that MR total sample number is less than 500, covering is preferable TOPN cell number accounts for the ratio of whole cells, the foundation as network coverage level normalization assessment bonus point；

4. if be collected into the base station 66 of MR data, sector 162 in region, and a certain date collects from 18:00 to 22:00 It is more than cell 135 of 500 to MR sum, and weak station of covering (< -100dbm) ratio lower than 25% has 49 in this 135 stations It is a, accounting 36.3%；Cell higher than 50% has 41, accounting 30.4%.The good coverage rate of base station has reacted the resource effect of base station Rate and quality have a direct impact the covering level of network, and the ultimate aim of network construction is to realize seamless coverage, and coverage rate exists Greater weight should be accounted in score-system, setting weighting coefficient takes 0.8；The ratio that base station cell covers embodies previous work Effect, in line with the principle for improving first-line staff's job enthusiasm and enthusiasm, bonus point weight when normalization assessment covering is horizontal Coefficient takes 0.6, comprehensive bonus point coefficient=0.2*0.6=0.12；The ratio of base station cell covering difference has reacted base station selection, prospecting and The deficiency of design work is direction and the emphasis of next stage work, deduction weight system when normalization assessment covering is horizontal Number takes 0.4, comprehensive deduction coefficient=0.2*0.4=0.08.

5. in conclusion network normalizes covering level=63.7*0.8+0.2*(30.4*0.12-36.3*0.08) * 100= 65.57 points.

Later, the normalization assessment of network interferences degree is carried out, comprising:

3 annoyance level of 1.PCI mould.Based on north orientation MR count ratio that the whole network cell interfered by PCI mould 3 (adjacent area PCI with Main serving cell PCI 3 results of mould are identical, while RSRP and main serving cell RSRP absolute value of the difference < 3dB MR and the cell are complete The ratio of portion MR), it is assumed that the cell accounting more than 40% is 16.25% to get 26/160=16.25 point, and normalization weighted value is gone 0.3(is daily work by reasonably planning and effectively optimization can drop to very low because the interference of PCI mould 3 is internal interference The emphasis of work) afterwards score 4.86 divide.

6 annoyance level of 2.PCI mould.Based on north orientation MR count ratio that the whole network cell interfered by PCI mould 6 (adjacent area PCI with Main serving cell PCI 6 results of mould are identical, while RSRP and main serving cell RSRP absolute value of the difference < 3dB MR and the cell are complete The ratio of portion MR), it is assumed that cell accounting more than 25% is 28/160=17.5% to get 17.5 points, normalization weighted value go 0.2(because For PCI mould 6 interference be internal interference, by reasonably plan and effectively optimization can drop to it is very low, be routine work time Emphasis) afterwards score 3.5 divide.

3. network overlapped level of coverage.Ratio (adjacent area RSRP and the main service of the small area overlapping of the whole network are counted based on north orientation MR The number of PCI of the cell RSRP absolute value of the difference less than 3dB >=3 MR and cell whole MR ratio is attached most importance to folded coverage coefficient PCI number of 1,3dB > absolute value of the difference >=6dB >=3 ratio is attached most importance to folded coverage coefficient 2,6dB > absolute value of the difference >=9dB PCI number >=3 ratio is attached most importance to folded coverage coefficient 3), if overlapping coverage coefficient 1 be more than 35% cell accounting for 28/160= 17.5%.Up to 17.5 points, normalization weighted value removes 0.3*0.7.The cell accounting that overlapping coverage coefficient 2 is more than 35% is 35/160 =21.86%.Up to 21.86 points, normalization weighted value removes 0.3*0.2.The cell accounting that overlapping coverage coefficient 3 is more than 35% is 44/ 160=27.5%.Up to 27.5 points, normalization weighted value removes 0.3*0.1.(because of overlapping coverage coefficient, by reasonably planning and Effectively optimization can drop to very low, be the emphasis of routine work, so weighting total value takes 0.3, and because overlapping coverage coefficient 1 It is the emphasis subitem weighting coefficient 0.7 of Topological expansion, overlapping coverage coefficient 2 is the secondary emphasis subitem of Topological expansion Weighting coefficient 0.2, overlapping coverage coefficient 3 are the secondary emphasis subitem weighting coefficients 0.1 of Topological expansion), comprehensive 3 overlappings are covered The weight score of lid coefficient, 17.5*0.3*0.7+21.86*0.3*0.2+27.5*0.3*0.1=3.96.

4. level of coverage is crossed in base station.Based on north orientation MR count the whole network cell get over area cross covering ratio (cross first lap and Second circle adjacent area, the ratio of the MR and cell whole MR of RSRP and other cells RSRP absolute value of the difference less than 3dB were to cover Lid coefficient crosses covering identification and implements relative complex, and algorithm is not detailed), there are 8 to there is covering in 160 cells, accounts for Than being 8/160=5%.Up to 5 points, normalization weighted value goes 0.2, and level of coverage is last excessively obtains 1 point.

5. uplink interference degree.It is counted according to the jamming power RIP value that ENB in north orientation MR is received, in practice it has proved that such as There are uplink interferences for fruit, then continuous longer period of time has.RIP > 27 (27-126=- 99dB) in one day are counted by base station cell Sum > 3000 cell accounting, if in 160 cells, uplink interference level >=- 99dBm belongs to the cell 41 of serious interference It is a, accounting 25.63%, weighting coefficient take 0.5(and downlink interfere constitute 100% local score-system) * 0.7(is tight according to interference The weighting of weight degree), score 8.97 is divided.- 99dBm > uplink interference level >=- 105dBm, there are cell 62 of interference, accounting 38.75%.Weighting coefficient takes 0.5*0.2, and score 3.86 is divided.- 105dBm > uplink interference level >=- 110dBm cell 85, Accounting 53.13%.Weighting coefficient takes 0.5*0.1, and score 2.66 is divided.Uplink interference level is in addition to inside and outside annoyance level Have outside the Pass, it is also closely related with network load level, ICIC efficiency etc..Network load level more high interference is bigger, ICIC difficulty Bigger, judging network load level most straightforward approach is uplink PRB utilization rate.The 19% of enchashment net calculates, score (8.97+ 3.86+2.66) * (1-19%)=12.55.

6. uplink signal-to-noise ratio and the received jamming power of ENB.North orientation MR uplink signal-to-noise ratio and the received jamming power two of ENB Uplink interference degree can be assessed from two dimensions simultaneously by tieing up statistical data, it is assumed that RIP>-100dBm and upstream sinr<0dB Accounting 12.46%, score 12.46 are divided.Uplink interference assessment can be than more fully reacting net made by described above 5th and the 6th Network situation, can be with objective after arithmetic average, therefore the uplink interference degree final score (12.55+ after comprehensive 5,6 / 2=12.5 point 12.46).

7. downlink annoyance level.Corresponding pass downlink RSRQ unfixed with SINR or throughput in north orientation MR data System, so the basic index without reacting downlink annoyance level.So needing to count downlink annoyance level according to drive test data. If the CDF statistics of network level SINR≤0 is 9.11%, score 6.38 is divided after weighting 0.7.Corresponding SINR < 0 cell-level CDF=5% Cell accounting is 37/160=23.13%, and score 6.94 is divided after weighting 0.3.Downlink annoyance level final score is (6.38+ 6.94) * 0.5=6.66 points.

8. network RSRQ is assessed.The whole network RSRQ is big, and portion concentrates between -11.5 ~ -5.5dB, the corresponding RSRQ of CDF5% about - 15dB.The experience obtained is analyzed from mass data, RSRQ is lower than -15dB, and SINR is generally poor, but RSRQ high, SINR are not One sets.Downlink RSRQ is horizontal with external disturbance, is overlapped coverage, 3 mould of PCI mould, 6 annoyance level, PRB load level correlation.Institute It can reflect the degree of network signal quality difference with RSRQ < -14.5dB ratio 3.13%, RSRQ assessment score 3.13 is divided, weighted The sum of 0.7(network level score and cell-level score are weighted statistics by 100%) afterwards score 2.19 divide.There are RSRQ data and remittance General Logistics Department's MR sum is more than cell 153 in total of 50, and poor (RSRQ < -14.5dB) ratio of matter is more than cell 9 of 10%, accounting 5.88%.Score 5.88 is divided, and score 1.76 is divided after weighting 0.3.Network RSRQ assesses final score 2.19+1.76=3.95 point.

7 score values of comprehensive 1-8 above, are enumerated as follows:

3 annoyance level score 4.86 of 1.PCI mould is divided

6 annoyance level score 3.5 of 2.PCI mould is divided

3. network overlapped level of coverage score 3.96 is divided

Divide 4. level of coverage score 1 is crossed in base station

5. uplink interference degree final score 12.5 divides

6. downlink annoyance level final score 6.66 divides

7. network RSRQ assessment final score 3.95 divides

Obtain network interferences degree normalization assessment score 4.86+3.5+3.96+1+12.5+6.66+3.95=36.43 point.

Later, statistical correlation data.Assuming that statistical result are as follows: network location entry number 10,000,000 passes through network management Or it is soft adopt, adopt firmly signaling platform statistics network registration number of users (1,000,000), network active users (660,000), network base station number (6000), cell number (15000), uplink PRB utilization rate (19%), downlink PRB utilization rate (30%).

Later, normalized output assessment result and Optimizing Suggestions:

1. network active users (660,000)/cell number (15K)=4.4；

2. network registry number of users (1,000,000)/network base station number (6K)=16.67；

3. uplink PRB utilization rate (19%)=19(theory full marks 100)；

4. downlink PRB utilization rate (30%)=33(theory full marks 100)；

5. network interferences degree=36.43(theory full marks 100)；

6. network coverage level=65.57(theory full marks 100)；

Then network normalized value=0.4* [uplink PRB utilization rate (19%)+downlink PRB utilization rate (30%)]/2+0.6* [(100- network interferences degree 36.43)+network coverage level 65.57]/2+4.4+16.67

=[(19+30)*0.4+(63.57+65.57)*0.6]/2+4.4+16.67

=48.54+4.4+16.67

=69.61。

The present invention does normalization assessment to Network resource allocation, covering efficiency, load level, annoyance level etc., finally Divide 69.61 points, network room for promotion is larger.Optimizing Suggestions are more, and particular content is pressed following table such as and exported:

1. exporting weak coverage cell list；

The 3 interfered cell list of 2.PCI mould；

The 6 interfered cell list of 3.PCI mould；

4. network overlapped coverage cell list；

5. coverage cell list is crossed in base station；

6. the serious cell list of uplink interference；

7. downlink serious interference cell list；

8. the serious cell list of external disturbance (opens table pack according to front 7 to screen, existing uplink interference has downlink interference again But interfere lower or overlapping covering little without the interference of PCI mould or PCI mould).

It is whole to be indicated with figure, including following application mode:

1. normalized value is ordinate, and drawn trend chart can objectively respond each using the time as abscissa coordinate The effect of period network construction optimization.

2. normalized value is ordinate using geographical location as coordinate, drawn trend chart can objectively respond prefectures and cities The difference of LTE network construction can be used to carry out task management.

3. subsidiary assessment report and Optimizing Suggestions book can be expressly recited the acquired achievement of network construction optimization and not Foot, can be used for achievement.

Be explained above based on north orientation MR data, network management data, drive test data model algorithm, LTE is soft to be adopted data, adopts firmly The algorithm of data, frequency sweep data and ATU data in assessment network coverage level and annoyance level is essentially the same, while these are counted According to further including a large amount of contents, can be perceived from user, quality of service, the dimensions such as migration efficiency further improve assessment content, this It is not described here in detail for invention.

The above, only presently preferred embodiments of the present invention, the invention is not limited to above embodiment, as long as It reaches technical effect of the invention with identical means, all should belong to protection scope of the present invention.

Claims (3)

1. a kind of normalized LTE network evaluation optimal sysytem of end-to-end various dimensions characterized by comprising

Transacter, collected data type include that soft data, north orientation MR data, the LTE of adopting of LTE adopts data, frequency sweep firmly Data, drive test and ATU data, OMC performance statistic and network element device basic data；

Data integrity assesses device, and the integrality for the data to collection is assessed, the data integrity assessment dress When setting the ratio of the base station cell base station cell Zhan Zong corresponding to the data of some data type of acquisition less than setting ratio, It will be prompted to transacter to continue to acquire the data of the type, or by model evaluation device to acquired base station cell Carry out single station assessment；

Model evaluation device, the model evaluation device are commented comprising corresponding network structure reasonable evaluation module, covering level Estimate module, annoyance level evaluation module, network load level evaluation module, energy efficiency evaluation module, System Performance Analysis mould Block and terminal user perceive evaluation module, and corresponding data evaluation module respectively assesses to obtain corresponding data type Assessed value；

Normalization assessment device, the normalization assessment device are that the data evaluation module of corresponding data type is provided with accordingly Weight, and normalized value can be obtained by weight and assessed value.

2. LTE network evaluation optimal sysytem according to claim 1, which is characterized in that further include:

Output module is assessed, for exporting assessment result and corresponding Optimizing Suggestions.

3. LTE network evaluation optimal sysytem according to claim 1, which is characterized in that the transacter uses Multidimensional normalizing weighted model establishes data source and collects corresponding data.