CN106231621A

CN106231621A - A kind of many scene adaptives optimization method of propagation model in FDD LTE system

Info

Publication number: CN106231621A
Application number: CN201610614523.4A
Authority: CN
Inventors: 许贤泽; 郑成林; 程文强
Original assignee: Wuhan University WHU
Current assignee: Wuhan University WHU
Priority date: 2016-07-29
Filing date: 2016-07-29
Publication date: 2016-12-14

Abstract

The present invention relates to a kind of many scene adaptives propagation model optimization method based on FDD LTE wireless network, use the propagation model after optimizing correction to carry out field intensity prediction so that prediction signal can be closer to practical measurement signals.The present invention relates to the correction of COST231 Hata and SPM both propagation models, many displayings in view of communication environments, invention is one group with single base station data, often to organize data as object of study, has eventually formed a correction propagation model storehouse based on each base station.Propagation after one the optimum correction of selection of the scene covered for each base station, it is achieved thereby that the many scene adaptives of propagation model adjust.The communication environments of different scenes is also connected by propagation model storehouse by the present invention with propagation model, has the good reference value of the highest guidance, improving the range of application also improved while calibration accuracy for the follow-up network planning and the network optimization.

Description

A kind of many scene adaptives optimization method of propagation model in FDD-LTE system

Technical field

The present invention relates to field of wireless communications networks, particularly propose a kind of in FDD-LTE system propagation model many The method that scene adaptive optimizes.

Background technology

In recent years along with the extensive utilization of LTE wireless 4G network, information-intensive society enters network big data age, soon Speed development universal Intelligent mobile terminal applies have promoted increasing substantially of global mobile data traffic, in mobile big data Epoch, the opportunities and challenges of the development that mass data, the characteristic such as multiformity of data are brought to wireless network.Wireless FDD-LTE Being one of the whole world two big 4G standards, it is researched and developed earlier than TD-LTE, and technology is more ripe, and terminal is more rich, through being whole world model Enclosing interior most popular wireless 4G network, contrast with TD-LTE wireless network, FDD-LTE has fast, the applicable wide area of speed and covers The advantages such as lid.

Along with the day by day complexity of wireless network with wireless propagation environment, the correction of propagation model becomes Network Programe Design and network The aspect performances such as the key link of the work such as optimization, directly affects the final network coverage, network capacity amount, communication quality.Due to wireless Electric wave can be affected, by communication environments, multiple biographies such as can there is reflection, diffraction, scattering and direct projection during aloft propagating Broadcast mode and approach, this can cause receiving signal multipath is weak, signal time delay.Radio transmission model is primarily used to describe and sends out Machine of penetrating to receiving field intensity and the Changing Pattern thereof of signal at each point around the dissemination of electric wave signal between receiver and base station, and Conventional radio transmission model is on the basis of mass data is measured, for the electric wave of different frequency range at different communication environments Statistically empirical model obtained by lower.Therefore the propagation characteristic of propagation model is closely bound up with communication environments, for this propagating mode Type was particularly important with mating of communication environments, proposed a kind of propagation in FDD-LTE system for this patent of the present invention The method that many scene adaptives of model optimize.

Using at present more in propagation model revision method is CW (Continuous Wave, continuous wave) test, then passes through Analytical calculation CW test data carry out the correction optimization of propagation model, and then carry out the network plannings such as MPS process prediction again and set Meter.And CW test is to obtain test data by building non-directional transmitter before networking, it is wireless with bulk zone Environmental dissemination model is corrected to research unit, and this allows for the lacking of property of prediction of the wireless environment to concrete a certain community, Virtually add forecast error；Meanwhile bulk zone wireless environment has complexity, so can cause carrying out model Need the variable parameter considered to increase the when of correction, such as the distance of test point to transmitter, frequency, launch antenna the highest Degree, reception antenna effective depth, communication environments type etc.；Furthermore it is exactly the CW test mobile communications network not to different systems Treat with a certain discrimination.

Summary of the invention

It is contemplated that overcome defect that above-mentioned prior art exists and provide one to save time, laborsaving, economical and have higher Many scene adaptives optimization method of propagation model in FDD-LTE network of accuracy.

The concrete technic relization scheme of the present invention is as follows:

A kind of many scene adaptives optimization method of propagation model in FDD-LTE system, it is characterised in that the method comprises Following steps:

Step 1.1, by base station distribution situation on the electronic chart in region to be measured is analyzed, select test station and drive test Route；

Described drive test route selection is according to following condition:

Condition one: test route comprises base station on the way；

Condition two: test route is through the different electromagnetic propagation environment of at least two, and contains all roads；

Condition three: avoiding drive test route through high building shadow region, different distance, different directions are all contained in region to be measured；

The described following condition of website selection gist:

Condition one: selected website covers all types of ground objects in region to be measured；

Condition two: increase at least one overlapping region between each website；

Condition three: the base station antenna height of selected website is more than 20 meters, antenna is higher than nearest barrier more than 5 meters, website place Building should be higher than the average height of surrounding building；

The engineering parameter of the website to be measured selected in step 1.2, obtaining step 1.1 is primarily referred to as the basic engineering ginseng of base station Number, including: base station is numbered, latitude and longitude of base station, and the extension of antenna for base station is high, and azimuth, angle of declination, then described in download step 1.1 The drive test data that obtains of drive test route survey；

Step 1.3, the drive test data surveyed based on FDD-LTE network system is carried out pretreatment, the screening of data and filtration；

Step 1.4, to step 1.3 process after data carry out Data Discretization；Owing to GPS sampling rate is than the sampling of receiver Speed is slow, and this makes to be sequentially arranged multiple data on same longitude and latitude, need for this to carry out these data from Dissipate and launch, it is assumed that the speed in test process keeps consistent, and the time interval between each data is also equal, then for a long time In chronological order data can be carried out interpolation process, thus the data repeating to go up on one point are tiled in chronological order and comes；

The valid data that step 1.5, use processed through said method, extract the main plot and neighbour received at test point The base station signal strength in district, calculates the test point path loss in FDD-LTE community drive test point entirety, and utilizes data to carry out The adaptive optimization of propagation model；

Step 1.6, it is respectively completed the adaptively correcting optimization to particular propagation environment of COST231-Hata and the SPM propagation model, Select a correction propagation model closer to measured data according to error mean and standard deviation, and various scene will be applicable to Propagation model record formed be suitable for propagation model storehouse class.

In a kind of above-mentioned method that many scene adaptives of propagation model optimize in FDD-LTE system, described step 1.2 Middle drive test packet includes testing time, the longitude and latitude of test point, main serving base station cell number, adjacent area number, test frequency, main service Base station cell received signal strength and neighbor received signal intensity.

In a kind of above-mentioned method that many scene adaptives of propagation model optimize in FDD-LTE system, described step 1.3 Described pretreatment is to filter during drive test road measuring device to lose test point longitude and latitude, the latitude and longitude information of main plot, main little District's signal receiving strength；

The filtration of data mainly has filter type based on received signal level intensity and spacing based on test point Yu dominant base From filter type

Filtercondition one: filter type based on received signal level intensity, this filtercondition is to come based on received signal strength Carry out data filtering, during actual drive test received signal strength the strongest or the most weak to data process accuracy the most very Big impact, is unfavorable for that the optimization of propagation model processes；Level is generally set in actual mechanical process for this and filters thresholding；

Filtercondition two: data filtering mode based on test point Yu the spacing of dominant base, the survey of the position close to base station Amount point, is affected by black scope under base station towers, and nearly base station testing point does not often have direct path, and the nearly base station of LTE is covered Lid test point does not receives signal level when test or the signal level that receives is the most weak, and equipment fails to be identified, with Reason, during in distance base station farther out, base station signal makes through the reflection of communication environments, scattering, diffraction that signal intensity is the most weak to be caused Accepting device fails to receive effective information, avoids nearly base station and the test point away from base station when Path selection.

In a kind of above-mentioned method that many scene adaptives of propagation model optimize in FDD-LTE system, described step 1.5 In, to FDD-LTE community drive test point entirety path loss calculation method it is:

Step 4.1, when drive test data display test point receive only main serving cell signal, when being not received by neighboring area signal, The loss of test point Actual path just deducts the main serving cell signal intensity received, i.e. PL=P for base station transmitting power_BS- P_UE, wherein PL (PathLoss) is path loss, P_BSFor base station transmitting power, P_UEFor drive test point received signal strength；

Step 4.2, when drive test data show two or three sectors receive signal time, reject drive test point longitude and latitude for empty right The sector answered, only calculates remaining sector, if only remaining a sector, calculates by above-mentioned steps 4.1；

If step 4.3 residue sector is two sectors, wherein host sectors path loss is PL₀=P_BS0-P_UE0, another adjacent area Path loss is PL₁=P_BS1-P_UE1；Test accuracy is promoted, to main plot and neighbour in order to take into account main plot and information of adjacent cells The way of district's signal weighting calculates received signal strength；Through repeatedly test experiments, finally draw weight ratio be 7:3 time Time can realize the optimization of experimental result, so the path loss after last test point integration neighbor received signal is PL=0.7* PL₀+0.3*PL₁；

If step 4.4 remains three sectors, calculate host sectors path loss PL with reference to step 4.3₀, the second path, sector Loss PL₁With the 3rd sector path loss PL₂, second sector that wherein signal is stronger is calculated and presses 0.2*PL₁, for the 3rd Individual sector 0.1*PL₂, it is PL=0.7*PL that last test point integrates the path loss after neighbor received signal₀+0.2*PL₁+0.1* PL₂；

If step 4.5 remains the many excess-three in sector, retain the sector that host sectors is the strongest with two other signal, And calculate drive test point path loss according to above-mentioned steps 4.4, due to the complicated variety of communication environments, received signal strength is subject to Multifactor restriction, the data reliability that therefore signal intensity is higher is higher.

In a kind of above-mentioned method that many scene adaptives of propagation model optimize in FDD-LTE system, described step 1.5 Data after utilization process carry out the optimization of propagation model and specifically mainly comprise the steps that

Step 5.1, based on COST231-Hata propagation model, utilization process after data, take least-square fitting approach, and After checking propagation model optimizes, propagation model prediction data is less than 8dB with measured data standard deviation, and carries out next step；

Step 5.2, based on SPM propagation model, utilize SPM model coefficient initial value, carry out propagation model revision optimization, carry out data During matching, make the prediction data of SPM model minimum, such as less than with the standard deviation of measured data with the setting of each coefficient 8dB then counts model class libraries, otherwise terminates this step and carries out next step；

Step 5.3, according to step 5.1,5.2, compare above two propagation model optimize after standard deviation and error mean, select One correction result closest to prediction data is adaptive to this scene communication environments；Step 5.4, output adaptive are in this The communication environments of scape, and by communication environments and the Optimized model typing propagation model storehouse class being applicable to this communication environments, set up one The individual propagation model storehouse class with practicality, reference and assosting effect.

Present invention have the advantage that the correction of the propagation model in FDD-LTE wireless network directly can use this network Drive test data, it is to avoid have more specific aim while the CW test of loaded down with trivial details, inefficiency, it is possible to realize geographical passing different Broadcast the propagation model self-adaptative adjustment of environment, enrich propagation model storehouse class, for the network planning afterwards, optimize the reliable of offer Reference and auxiliary.

Accompanying drawing explanation

Fig. 1 is the flow chart of the present invention；

Fig. 2 is the propagation model revision flow chart of the present invention.

Detailed description of the invention

So that the object, technical solutions and advantages of the present invention are clearer, below in conjunction with the accompanying drawings with detailed description of the invention pair The present invention is described in detail.

The core concept of the present invention is: utilize the drive test data that existing actual mobile communications network obtains, with communication environments Many scenes are many scene adaptives correction that coupling object carries out propagation model.Advantages of the present invention is as described below:

On the one hand drive test data is to obtain based on real network, can not only reflect that the radio propagation channel of reality promotes accurately The accuracy of model correction, and drive test data has the highest exploitation value in the planning and designing and system optimization of wireless network Value, it is possible to the lifting for network communication quality, network capacity, the network coverage provides reliable data foundation and knows direction；

For CW tests, drive test need not build the non-directional transmitter for measuring specially, and whole implementation scheme saves time Laborsaving easy to implement, greatly reducing manpower and materials；

When carrying out propagation model revision, three propagation models are corrected simultaneously, finally can be according to prediction data and actual measurement number According to error mean and standard deviation select a propagation model revision result closest to actual measurement data to mate currently Measure scene, it is achieved that the adaptively correcting of many scenes, meanwhile also set up the propagating mode matched with concrete communication environments Type storehouse class.

Fig. 1 is the schematic flow sheet of the present invention, as it is shown in figure 1, this embodiment specifically includes that

Step 1, selects suitable website and drive test route；

Test station condition must be able to represent exemplary base station condition, and essential condition includes that antenna hangs height, the ground of surrounding enviroment The conditions such as thing landforms.Test route requires that the road of all directions all includes, during test, the position of various distances is all Should cover, the region near various atural objects will cover so that test data are evenly distributed as far as possible.

The principle of drive test route selection has:

1. test route comprises many base stations the most as far as possible；

2. test route is tried one's best through different electromagnetic propagation environment, contains more road as far as possible；

Avoid drive test route through high building shadow region, it is ensured that different distance, different directions are all contained in region to be measured the most as far as possible.

The principle that website selects has:

The most selected website should the most enough cover region to be measured in all types of ground objects；

Increase the overlapping region between each website the most as far as possible；

The base station antenna height of the most selected website should be greater than 20 meters, and antenna is higher than nearest barrier more than 5 meters, and website place is built Thing should be higher than the average height of surrounding building.

Step 2, obtains FDD-LTE network engineering parameter ginseng and drive test data；

FDD-LTE network work ginseng mainly include the longitude and latitude of test station, antenna hang high, launch power, test station cell id, Antenna feeder configures；Drive test data mainly includes testing time, the longitude and latitude of test point, main serving base station cell number, adjacent area number, test Frequency, main serving base station cell received signal strength and neighbor received signal intensity.

Step 3, road test data carries out pretreatment；

Aloft propagate due to radio wave and affected by multiple decline, and the receiving signal sensitivity of road measuring device is limited, Therefore have during drive test and some important information can be caused to lose.In addition radio wave many scenes in various communication environments are cut Change the major reason being also to cause information loss.The latitude and longitude information just like main plot relatively common in dropout is main little District's signal receiving strengths etc., should be filtered for the most invalid data.

Pretreated FDD-LTE drive test data information table is as indicated with 1.

Data message table after table 1 pretreatment:

Drive test information	Information explanation
		LogTime	Testing time
Lon	Test point longitude
		Lat	Test point latitude
ServerCellPCI	Main plot number (physical-layer cell identifier)
		ServerCellRSRP	Main plot received signal strength
NBCellPCI	Adjacent cell number (physical-layer cell identifier)
		NBCellRSRP	Adjacent cell received signal strength
DLFrequency	Test frequency

After invalid data in drive test data is filtered, remaining valid data are divided into different little according to the difference of base station Group.

Step 4, checks whether and filters complete by all invalid datas.

Step 5, in order to improve the accuracy of data, also needs road test data to be further processed, particular flow sheet such as Fig. 2 institute Show.

The screening of data and filtration:

The filtration of data should be set about in terms of two: filtration based on test point Yu dominant base distance；Based on the letter received The filtration of number intensity.

Data filtering based on level, it is contemplated that receive the shadow that signal accuracy that is too strong or that process data too much is the biggest Ring, level is set for this and filters thresholding (such as filtering less than-110dBm and the data point more than-50dBm).

Data filtering based on distance, the measurement point in the position close to base station, affected by black scope under base station towers, with And closely base station testing point does not often have direct path, the nearly base station coverage test point of LTE does not receives signal electricity when test Signal level that is flat or that receive is the most weak, and equipment fails to be identified, and in like manner, during in distance base station farther out, base station signal passes through The reflection of communication environments, scattering, diffraction make that signal intensity is the most weak causes accepting device to fail to receive effective information, therefore exist Avoid nearly base station and the test point away from base station, it is contemplated that base station range and communication environments etc. the when of Path selection as far as possible Factor measuring distance is typically advisable with 150m～3000m.

Data discrete, owing to GPS sampling rate is slower than the sampling rate of receiver, this makes on same longitude and latitude temporally Order arranges multiple data, needs for this these data are carried out discrete expansion, it is assumed that the speed in test process keeps one Cause, and the time interval between each data be also equal, then in chronological order data can be carried out interpolation process for a long time, Thus the data repeating to go up on one point are tiled in chronological order and comes.

Weighted calculation to cell received signal strength:

When drive test data display test point receives only main serving cell signal, and when being not received by neighboring area signal, test point is real Border path loss just deducts the main serving cell signal intensity received, i.e. PL=P for base station transmitting power_BS-P_UE, wherein PL (PathLoss) it is path loss, P_BSFor base station transmitting power, P_UEFor drive test point received signal strength；

When drive test data shows two or three sectors reception signal, rejecting drive test point longitude and latitude is the fan corresponding to sky District, only calculates remaining sector, if only remaining a sector, calculates by above-mentioned steps [0038]；

If residue sector is two sectors, wherein host sectors path loss is PL₀=P_BS0-P_UE0, another adjacent area path loss For PL₁=P_BS1-P_UE1.Promote test accuracy to take into account main plot and information of adjacent cells, main plot and neighboring area signal are added The way of power calculates received signal strength.Through repeatedly test experiments, finally draw and can realize the when that weight ratio being 7:3 The optimization of experimental result, so the path loss after last test point weighted calculation is PL=0.7*PL₀+0.3*PL₁；

If during three sectors of residue, calculating host sectors path loss PL with reference to step [0040]₀, the second sector path loss PL₁With the 3rd sector path loss PL₂, second sector that wherein signal is stronger is calculated and presses 0.2*PL₁, for the 3rd fan District 0.1*PL₂, the path loss after last test point weighted calculation is PL=0.7*PL₀+0.2*PL₁+0.1*PL₂；

If the when of the residue many excess-three in sector, retain the sector that host sectors is the strongest with two other signal, and according to upper State step [0041] and calculate drive test point path loss, due to the complicated variety of communication environments, received signal strength by many because of Element limits, and the data reliability that therefore signal intensity is higher is higher.

All propagation models are carried out adaptively correcting under various scenario by step 6.

This patent carries out parameter correction based on COST231-Hata and SPM propagation model.

COST231-Hata model formation is:

L_p(dB)=46.3+33.9lgf-13.82lgh_b-α(h_m)+(44.9-6.55lgh_b)lgd+C_M

Wherein,

The signal carrier frequency (MHz) that f: base station is used；

h_m: mobile portable antennas height (m)；

h_b: base station antenna height (m)；

D: the distance (km) between base station and mobile station.

SPM propagation model formula:

L_p(dB)=k₁+k₂lgd+k₃lgh_te+k₄lgh_re+k₅×Diff+k₆lgd×lgh_te+C_clutter

k₁For constant offset；

k₂For the range attenuation factor, default value is 44.9；

k₃For Base Transmitter antenna effective height correlation factor, default value is 5.83, h_teFor Base Transmitter antenna effective height；

k₄For mobile station reception antenna effective depth correlation factor, default value is 0, h_reFor mobile station reception antenna effective depth；

k₅For Diffraction Calculation correlation factor；

k₆For effective height of transmitting antenna and propagation distance correlation factor, default value is-6.55；

Diff is diffraction loss；

C_clutterFor the topographical correction factor；

For correcting after the data of propagation model parameter process according to preceding method, it is divided into different little according to each base station Group, observes above-mentioned propagation model formula, and for the convenient meter of correction, special is data acquisition system by Data Integration:

Data={ (x_ij,y_ij), i=1,2 ..., n；J=1,2 ... }；

Wherein, x_ij=log (d_ij) it is the logarithm of distance between Receiver And Transmitter, d_ijFor according to test point longitude and latitude and The actual range that dominant base longitude and latitude is tried to achieve after Coordinate Conversion, i is drive test data number, j base station number, the most yes Often group data constitute a little data acquisition system, facilitate the subsequent treatment of data.

Wherein y_ijFor in distance d_ijUnder the conditions of survey path loss, be cell received signal strength described in above-mentioned steps and add Path loss PL that power calculates.

Make b=46.3+33.9lgf-13.82lgh_b-α(h_m)+C_M, a=44.9-6.55lgh_bThen COST231-Hata model becomes For:

L_{p} (d B) = a l g d + b &DoubleRightArrow; L_{i j} = a_{j} x_{i j} + b_{j} .

For the correctness of subsequent authentication fitting result, before data matching, to retain the data of 30% for verifying that matching is tied The reasonability of fruit, is therefore only used for model correction matching by the 70% of overall data, owing to data are abundant, so being not concerned about number According to not, furthermore the checking being used for propagation model revision result with measurement data the most once is had more representativeness and cogency.

Use least square fitting that all data acquisition systems are fitted so that error sum of squares is minimum:

m i n Σ_{i = 1}^{n} {(y_{i j} - L_{p})}^{2} = m i n Σ_{i = 1}^{n} {(y_{i j} - a_{j} x_{i j} - b_{j})}^{2}

Judge whether model and measured data fitting degree after correction meet requirement, generally use error mean and standard deviation this Two indices is weighed.

e r r o r_m e a n = \frac{Σ_{i = 1}^{n} {error}_{i}}{n}; e r r o r_δ = \sqrt{\frac{Σ_{i = 1}^{n} {(e r r o r_m e a n - {error}_{i})}^{2}}{n - 1}}

Industry generally believes and illustrates when standard deviation is less than 8dB that calibration model meets actual propagation environment, i.e. this propagation at present The correction result of model is accurately, can serve as the foundation of the network planning and the network optimization, if being unsatisfactory for this standard, then Cast out this group propagation model revision result.

In like manner, it is exactly the process of the adjustment of a figure parameters for the correction of SPM model, k is set₁=-23.5, k₂=-44.9, k₃=5.83, k₅Taking 0.2 if urban district, suburb takes 0.4, and hills takes 0.5, k₅=0, k₆=-6.55, C_clutter=1, and utilize road Survey data and SPM propagation model is carried out field intensity prediction.

In view of correction efficiency and the impact of relevant parameter, only to k in the present invention₁、k₂、C_clutterCarry out parameter correction.

According to 70% drive test data imported, calculate whole during total error and mean square error, and adjust k₁、C_clutter The total error and the mean square error that make entirety are 0；

Ensure that above-mentioned parameter is constant, adjust k up and down₂Minimum to mean-square value.

After above-mentioned parameter adjusts, parameter is substituted into SPM propagation model and is predicted checking, in like manner, work as propagation model revision After the measured data predicted the outcome with another 30% standard deviation less than 8dB when think correction result be accurately, permissible As the network planning and the foundation of the network optimization, if being unsatisfactory for this standard, then cast out this group propagation model revision result.

Respectively COST231-Hata with SPM propagation model is corrected owing to employing drive test data, propagation model above Correction all carry out with each base station for correcting unit, it is now desired to select one more by error mean and standard deviation Close to the correction propagation model of measured data, thus for propagation model after covering many Scene realizations correction of different base station Self-adaptative adjustment.

In sum, the present invention not only achieves the propagation model self-adaptative adjustment being applicable to many scenes communication environments, but also The communication environments of different scenes is connected by propagation model storehouse class with propagation model, to the follow-up network planning and network There is for optimization the good reference value of the highest guidance, improving the range of application also improved while calibration accuracy.

Specific embodiment described herein is only to present invention spirit explanation for example.The technical field of the invention Described specific embodiment can be made various amendment or supplements or use similar mode to substitute by technical staff, but Without departing from the spirit of the present invention or surmount scope defined in appended claims.

Claims

1. many scene adaptives optimization method of propagation model in FDD-LTE system, it is characterised in that the method bag Containing following steps:

Described drive test route selection is according to following condition:

Condition one: test route comprises base station on the way；

The described following condition of website selection gist:

Condition two: increase at least one overlapping region between each website；

A kind of side that many scene adaptives of propagation model optimize in FDD-LTE system the most according to claim 1 Method, it is characterised in that in described step 1.2, drive test packet includes testing time, the longitude and latitude of test point, main serving base station cell Number, adjacent area number, test frequency, main serving base station cell received signal strength and neighbor received signal intensity.

A kind of side that many scene adaptives of propagation model optimize in FDD-LTE system the most according to claim 1 Method, it is characterised in that pretreatment described in described step 1.3 is to filter road measuring device during drive test to lose test point longitude and latitude Degree, the latitude and longitude information of main plot, main plot signal receiving strength；

The most according to claim 1 a kind of be applicable to FDD-LTE system propagation model many scene adaptives optimize Method, it is characterised in that in described step 1.5, to FDD-LTE community drive test point entirety path loss calculation method be:

If step 4.3 residue sector is two sectors, wherein host sectors path loss is PL₀=P_BS0-P_UE0, another adjacent area road Footpath loss is PL₁=P_BS1-P_UE1；Test accuracy is promoted, to main plot and adjacent area in order to take into account main plot and information of adjacent cells The way of signal weighting calculates received signal strength；Through repeatedly test experiments, finally draw the when that weight ratio being 7:3 The optimization of experimental result can be realized, so the path loss after last test point integrates neighbor received signal is

PL=0.7*PL₀+0.3*PL₁；

If step 4.4 remains three sectors, calculate host sectors path loss PL with reference to step 4.3₀, second path, sector damage Lose PL₁With the 3rd sector path loss PL₂, second sector that wherein signal is stronger is calculated and presses 0.2*PL₁, for the 3rd Sector 0.1*PL₂, it is PL=0.7*PL that last test point integrates the path loss after neighbor received signal₀+0.2*PL₁+0.1* PL₂；

A kind of side that many scene adaptives of propagation model optimize in FDD-LTE system the most according to claim 1 Method, it is characterised in that the data after described step 1.5 utilization process carry out the optimization of propagation model and mainly include following step Rapid:

Step 5.3, according to step 5.1,5.2, compare above two propagation model optimize after standard deviation and error mean, select One correction result closest to prediction data is adaptive to this scene communication environments；

Step 5.4, output adaptive are in the communication environments of this scene, and by communication environments and the optimization being applicable to this communication environments Model typing propagation model storehouse class, sets up a propagation model storehouse class with practicality, reference and assosting effect.