CN105791181A - Channel estimation and balancing method for rail transit high-speed moving scene - Google Patents

Abstract

The invention discloses a channel estimation and balancing method for a rail transit high-speed moving scene. The method comprises following steps of S1, obtaining a current train position; S2, setting up a channel model at a prediction train position; S3, determining basic impact response of the channel at the prediction train position based on the current position of the train; S4, tracking the variable quantity of the channel response at the prediction train position; rapidly estimating the actual channel impact response at the prediction position and S5, balancing the channel according to the actual channel impact response at the prediction position determined in the step S4. According to the technical scheme provided by the invention, channel estimation is carried out on the known channel model at the train moving position through combination of the rail transit scene and the channel modeling; the channel estimation time delay is very short; the estimation precision is high; and the method is applicable to precise channel estimation and balancing in high-speed moving and complex scenes.

Description

Channel estimation and equalization method for rail transit high-speed mobile scene

Technical Field

The invention relates to the technical field of wireless mobile communication, in particular to a channel estimation and equalization method for a high-speed moving scene of rail transit.

Background

With the rapid development of rail transit systems such as railways, subways, intercity railways, particularly high-speed railways and the like, reliable, real-time and efficient broadband wireless network services are provided for train passengers, and the broadband wireless network services become hot spots of domestic and foreign broadband mobile communication research. However, the rapid time-varying of the channel due to high speed movement and complex scenarios of rail traffic (viaducts, cutting, tunnels, marshalling stations, railroad junctions, etc.) can distort the transmitted train control signals. If the channel distortion is not equalized or compensated, the signal recovery of a receiving end is seriously influenced, and the driving safety is influenced. Therefore, the channel estimation technology is significant for a rail transit system for transmitting train control information and ensuring safe operation of trains.

The channel estimation method is mainly divided into three categories: non-blind channel estimation, blind channel estimation and semi-blind channel estimation. Non-blind channel estimation methods are further classified into training sequence-based and pilot frequency-based methods. The pilots may be divided into block pilots, comb pilots, trellis pilots, etc. according to the pilot insertion manner. The channel estimation criteria are roughly 3: least Squares (LS) algorithm, Minimum Mean Square Error (MMSE) algorithm, and Maximum Likelihood (ML) algorithm. The blind channel estimation is based on the characteristics and statistical characteristics of the transmitted information symbols, and has potential in improving the capacity and reliability of the communication system, but has a slow convergence rate. Semi-blind channel estimation is a compromise between data transmission efficiency and convergence rate, that is, less training sequences are used to obtain channel information.

Based on the method, the terminal moving speed in the traditional mobile communication scene is low, and enough time is provided for acquiring and tracking channel synchronization so as to realize channel estimation. But under high-speed moving conditions, it is required to achieve fast synchronization acquisition and accurate channel estimation in a very short time. The conventional channel estimation technology suitable for medium and low speed mobile scenarios is not suitable for high speed mobile scenarios. The existing channel estimation and equalization technology suitable for a high-speed moving scene is based on accurate acquisition of channel state information or estimation and interpolation in a frequency domain based on pilot frequency, and the channel state information is outdated quickly in a high-speed moving state; a large number of inserted pilots will increase the estimation delay significantly. The methods still have the problems of higher complexity of channel estimation, long delay time of channel estimation, incapability of timely tracking channel change in a high-speed moving state, difficulty in obtaining accurate channel estimation and equalization in a very short time in a complex scene and the like. These methods often require a compromise in estimation accuracy and estimation time.

Therefore, it is desirable to provide a channel estimation method that is suitable for accurate channel estimation and equalization in a complex scenario with high-speed movement.

Disclosure of Invention

The technical problem to be solved by the invention is to provide a channel estimation and equalization method for a rail transit high-speed moving scene, which combines a rail transit scene and channel modeling, reduces the time delay of channel estimation and improves the estimation precision by carrying out channel estimation on a known channel model at the moving position of a train, and is suitable for accurate channel estimation and equalization under high-speed moving and complex scenes.

In order to solve the technical problems, the invention adopts the following technical scheme:

a channel estimation and equalization method for a rail transit high-speed moving scene comprises the following steps:

s1, acquiring the current position of the train;

s2, constructing a channel model of the train prediction position;

s3, determining the basic impact response of the train prediction position channel based on the current position of the train;

s4, tracking the variation of the channel response of the predicted position, and quickly estimating the actual channel impact response of the predicted position;

and S5, performing channel equalization according to the actual impact response of the predicted position determined in the step S4.

Preferably, the step S1 includes:

s11, the transponder arranged on the ground base station sends the ID of the transponder to the train;

s12, measuring the relative position of the train and the ground base station;

s13, determining the current position P of the train₁：P₁Transponder ID + relative position of train to said ground base station.

Preferably, the relative position of the train and the ground base station is measured by a wheel track distance meter arranged on the train.

Preferably, the step S3 includes:

storing basic parameters corresponding to each operation position of the train into a channel database in advance;

and reading corresponding basic parameters from a channel database based on the channel model of the predicted position of the train and the current position of the train, and estimating the actual channel impact response of the predicted position.

Preferably, the basic parameters include: the number of the multi-paths of the channel corresponding to each running position of the train and the time delay and average energy data of each path.

Preferably, the step S2 includes:

s21, determining all propagation paths of the electromagnetic wave signals between the current position of the train obtained in the step S1 and the ground base station;

s22, calculating the transmission loss of each propagation path by using the path loss formula of electromagnetic wave transmission, PL (dB) - △₁+74.52+26.16log₁₀(f)-13.82log₁₀(h_b)-3.2log₁₀(11.75h_m)²+[44.9-6.55log₁₀(h_b)+△₂]log₁₀(D)

Wherein f represents the operating frequency band, h_bAnd h_mRespectively showing the effective height of the base station antenna and the effective height of the train antenna, and D showing the effective height of the trainDistance of transmission path between front position and ground base station, △₁And △₂Is a constant related to the transmission environment;

s23, calculating the transmission delay of each propagation path;

and S24, constructing a channel model of the current position based on the energy loss and the transmission delay of each transmission path.

Preferably, the step S4 includes:

s41, based on the receiving sequence of the pilot signal transmitted in the channel, estimating the channel response at the frequency point by using the minimum mean square error criterion;

s42, obtaining responses of all points in the channel by using an interpolation algorithm, namely predicting the variation of the channel response of the position;

s43, estimating the actual channel impact response of the predicted position: and the actual channel impulse response of the predicted position is equal to the basic impulse response of the predicted position channel and the variation of the channel response of the predicted position.

Preferably, the step S5 includes

In the running process of the train, all the sub-channels are recorded as: y is_i＝H_ix_i+w_i；

Equalising the channel using the minimum mean square error criterion, i.e.

minE{(g_iy_i-x_i)^H(g_iy_i-x_i)}，

Then there is a change in the number of,

$\left\{\begin{array}{c}{g}_i={{\hat{H}}_i}^{*}/({{\hat{H}}_i}^2+{{\mathrm{\σ}}_n}^2/{{\mathrm{\σ}}_s}^2)\\ {\hat{x}}_i=y_i{{\hat{H}}_i}^{*}/({{\hat{H}}_i}^2+{{\mathrm{\σ}}_n}^2/{{\mathrm{\σ}}_s}^2)\end{array}\right.$

wherein, g_iTo equalize coefficients, σ_n ²In order to be the variance of the noise,to equalize the output.

The invention has the following beneficial effects:

the technical scheme of the invention combines a rail transit scene and channel modeling, and carries out channel estimation on the known channel model at the moving position of the train, so that the time delay of the channel estimation is very short, the estimation precision is high, and the method is suitable for accurate channel estimation and equalization under high-speed moving and complex scenes.

Drawings

The following describes embodiments of the present invention in further detail with reference to the accompanying drawings;

fig. 1 shows a schematic diagram of a chain-like coverage of a rail transit private mobile communication network;

FIG. 2 is a diagram illustrating pilot insertion according to the present invention;

fig. 3 shows a flow chart of channel estimation and equalization according to the present invention.

Detailed Description

In order to more clearly illustrate the invention, the invention is further described below with reference to preferred embodiments and the accompanying drawings. Similar parts in the figures are denoted by the same reference numerals. It is to be understood by persons skilled in the art that the following detailed description is illustrative and not restrictive, and is not to be taken as limiting the scope of the invention.

The existing channel estimation techniques mainly have two types: one is for medium and low speed moving scenes and one is for high speed moving scenes. The channel estimation suitable for the medium and low speed mobile scene is not suitable for the high speed mobile scene due to the factors of estimation complexity, large delay and the like. The method facing the high-speed mobile scene is either based on the acquisition of channel state information to perform channel estimation, or based on the channel estimation of pilot frequency in the frequency domain and then interpolation; the method carries out channel estimation by utilizing the known channel model at a certain position, has very short time delay of channel estimation and high estimation precision, and is suitable for high-speed moving complex scenes.

As shown in fig. 1, the basic idea of the present invention is: positioning of the train position by means of a wheel-track-based distance meter and transponder (assumed: P)₁) By using at the position P₁The position P is calculated according to the scene information and the channel model₁And the corresponding channel impulse response is accurately acquired under the channel condition, so that the rapid and accurate channel estimation can be realized under the complex high-speed moving scene. The scene information and the basic channel model can be obtained through a large amount of experimental data statistics and measurement.

The invention particularly discloses a channel estimation and equalization method for a high-speed moving scene of rail transit, which comprises the following steps:

step S1, obtaining the current position of the train

S2, constructing a channel model of the train predicted position;

step S3, determining the basic impact response of the train prediction position channel based on the current position of the train;

step S4, tracking the variation of the channel response of the predicted position, and quickly estimating the actual channel impact response of the predicted position;

step S5, channel equalization is performed based on the actual impulse response of the predicted position determined in step S4.

For the present solution, the step S1 specifically includes: the transponder arranged on the ground base station sends the ID of the transponder to the train, and meanwhile, the relative position of the train and the ground base station is measured by using a wheel-track distance meter arranged on the train. Determining the current position P of the train according to the ID of the transponder and the relative position₁：P₁Transponder ID + relative position of train to said ground base station. As shown in fig. 1, the distance D is a fixed value, because the track of the rail transit train is fixed, the base station distance D of the communication network is also fixed.

For the present solution, the step S3 specifically includes: the method comprises the steps that the number of multi-paths of a channel corresponding to each running position of a train and time delay and average energy data of each path are stored in a channel database in advance; and reading corresponding basic parameters from a channel database based on the channel model of the predicted position of the train and the current position of the train, and estimating the actual channel impact response of the predicted position.

For the present solution, the step S2 specifically includes: determining all possible propagation paths of the electromagnetic wave signals between the current position of the train obtained in the step S1 and the ground base station, including a direct path, a reflected path and the like. And calculating the transmission loss of each propagation path by using a path loss formula of electromagnetic wave transmission, and calculating the transmission delay of each propagation path, wherein the energy loss and the transmission delay of each propagation path obtained by the calculation can be used for constructing a channel model of the current position.

For the present solution, the step S4 specifically includes:

s41, based on the receiving sequence of the pilot signal transmitted in the channel, estimating the channel response at the frequency point by using the minimum mean square error criterion;

s42, obtaining responses of all points in the channel by using an interpolation algorithm, namely predicting the variation of the channel response of the position;

s43, estimating the actual channel impact response of the predicted position: and the actual channel impulse response of the predicted position is equal to the basic impulse response of the predicted position channel and the variation of the channel response of the predicted position.

The step S5 specifically includes: in the running process of the train, all the sub-channels are recorded as: y is_i＝H_ix_i+w_i(ii) a Equalising the channel using the minimum mean square error criterion, i.e.

minE{(g_iy_i-x_i)^H(g_iy_i-x_i)}，

Then there is a change in the number of,

$\left\{\begin{array}{c}{g}_i={{\hat{H}}_i}^{*}/({{\hat{H}}_i}^2+{{\mathrm{\σ}}_n}^2/{{\mathrm{\σ}}_s}^2)\\ {\hat{x}}_i=y_i{{\hat{H}}_i}^{*}/({{\hat{H}}_i}^2+{{\mathrm{\σ}}_n}^2/{{\mathrm{\σ}}_s}^2)\end{array}\right.$

wherein, g_iTo equalize coefficients, σ_n ²In order to be the variance of the noise,to equalize the output.

The invention is further illustrated by the following set of examples:

the core idea of the scheme is as follows: the fast and accurate channel estimation and equalization method based on a channel model at a certain position has low complexity, and avoids the reduction of transmission efficiency and large estimation delay caused by the utilization of feedback channel state information or the insertion of a large number of pilot frequencies/training sequences.

In this example, as shown in fig. 3, a channel estimation and equalization method for a high-speed moving scene in rail transit specifically includes:

the first step is as follows: as shown in FIG. 1, the transponder provided in the ground base station transmits its ID to the train, the position P of the train₁The transponder ID + the relative position of the ground base station and the train obtained by the train wheel-track distance meter through the rotation of the wheels. Because the track traffic train running track is fixed, the base station distance D of the communication network is also fixed, and D1 is the distance from the train running position to the ground base station.

The second step is that: the large-scale fading channel model corresponding to the predicted position stored by the train-mounted receiving station is as follows:

PL(dB)＝△₁+74.52+26.16log₁₀(f)-13.82log₁₀(h_b)-3.2log₁₀(11.75h_m)²+[44.9-6.55log₁₀(h_b)+△₂]log₁₀(D)

wherein f represents the operating frequency band, h_bAnd h_mRespectively representing the effective height of the base station antenna and the effective height of the train antenna, D representing the distance of the transmission path between the current position of the train and the ground base station, △₁And △₂Is a constant related to the transmission environment.

Table 1 channel model in different scenes of rail transit

Table 1 shows viaducts and roads in rail transitGraben, station, tunnel, urban area, suburb, country and river under the scene large-scale channel model. Where the parameters a, B represent two fixed values. f denotes the operating frequency band, h_bAnd h_mRespectively representing the effective heights of the base station antenna and the train antenna, D₂And indicating the distance from the predicted position of the positioned train to the base station of the next cell to be switched.

The third step: and acquiring basic channel impulse response related to the train operation environment.

The estimation of the channel impulse response can be divided into a basic channel response part and a channel detail variation part. In the second step we have obtained a large scale fading prediction of the channel. As the track traffic train running track is fixed, for each running position, channel basic response of the position related to a channel fixed background, such as parameters of the number of multi-paths of a channel, time delay of each path, average energy and the like, can be obtained in advance through sufficient data measurement, the parameters related to the position can be stored in a channel database, the basic parameters of the channel can be rapidly predicted and read according to the specific position of the train obtained in the first step, the impulse response of the part is basically determined by the specific position of a predicted point and is recorded as h₁。

The fourth step: the detail variation of the channel is quickly estimated and tracked.

Because the channel of the train can be changed in detail in the moving process, the channel impulse response can not be completely the same even if the train passes through the same position at different moments, and the variation can be superposed on the channel response basic value obtained in the third step and recorded as h₂。h₂Can be obtained by conventional pilot/training sequence or like assistance data, but statistically h₂Relative to h₁Generally smaller, which is simply the amount of change in channel response relative to the base value, we can insert fewer pilot/training sequences for fast tracking relative to the full channel response estimate of a conventional transmission system. In the traditional channel estimation algorithm under the fast time-varying multipath channel, a large number of pilot frequencies are often required to be inserted/trainedThe training sequence can obtain a better channel estimation effect, and when the doppler spread is larger, the insertion amount of the pilot frequency/training sequence even reaches 50% of the transmission data amount, which seriously reduces the transmission efficiency of effective data. The method of the invention divides the channel estimation into a basic response part and a detail change part, the pilot frequency/training sequence is inserted only for tracking the detail change of the channel, particularly, when the channel response can be basically determined by the channel environment corresponding to the current position of the train, namely h₂Relative to h₁The amount of pilot/training sequence insertion can be neglected to be zero. Therefore, the method improves the complexity of the traditional channel estimation method, and greatly reduces the reduction of transmission efficiency and larger estimation delay caused by the insertion of the pilot frequency/training sequence. The final channel estimation result is recorded as h ═ h₁+h₂。h₂The specific tracking process of (2) is as follows:

taking pilot insertion as an example, fig. 2 shows a scattered pilot insertion pattern in an OFDM system. Generally, we need to first estimate the channel response of the pilot frequency point, and then obtain the complete channel response result through interpolation.

Let X be diag (X) for the transmitted pilot signal₁,X₂,…,X_P) The received sequence after passing through the channel is equal to

Y＝XH+N

Where H is the P × 1 channel matrix and N is the P × 1 dimensional noise vector. We estimate the channel response at the pilot points using the minimum mean square error criterion:

${\hat{H}}_{MMSE}=R_{HY}{R_{YY}}^{-1}Y$

wherein,

R_HY＝E(HY^H)

R_YY＝E(YY^H)＝XR_HHX^H+σ_n ²I_P

e () denotes the mean value, σ_n ²Representing the noise power.

In practical applications, to reduce the complexity of the calculation, we can choose the amplitude of the pilot frequency such that XX is^H＝σ_s ²I_PWherein σ is_s ²Is the signal power. In this case, the channel estimation result can be simplified as follows:

${\hat{H}}_{MMSE}=R_{HH}{(R_{HH}+{{\mathrm{\σ}}_n}^2/{{\mathrm{\σ}}_s}^2{I}_P)}^{-1}{\hat{H}}_{LS}$

after the channel responses at the pilot points are obtained, the channel responses at the other points can be obtained by interpolation, the simplest first-order linear interpolation is

$\hat{H}\left(k\right)=\hat{H}(mL+l)=(1-\frac{l}{L}){\hat{H}}_p\left(m\right)+\frac{l}{L}{\hat{H}}_p(m+1)$

Wherein,andthe channel responses are for two pilot points.

The fifth step: and carrying out channel equalization.

After the fast and accurate channel estimation result is obtained, the data can be equalized by adopting a time domain or frequency domain method to obtain demodulated data. Similarly, taking the OFDM system as an example, after obtaining the channel frequency domain impulse response, the transmission model on the ith sub-channel is recorded as

y_i＝H_ix_i+w_i

Wherein x is_iFor transmission symbols on the ith subchannel, y_iFor received symbols, w_iIs gaussian noise. Still adopting the minimum mean square error criterion to carry out equalization, and setting the equalization coefficient as g_iIt satisfies:

minE{(g_iy_i-x_i)^H(g_iy_i-x_i)}

then there is a change in the number of,

$\left\{\begin{array}{c}{g}_i={{\hat{H}}_i}^{*}/({{\hat{H}}_i}^2+{{\mathrm{\σ}}_n}^2/{{\mathrm{\σ}}_s}^2)\\ {\hat{x}}_i=y_i{{\hat{H}}_i}^{*}/({{\hat{H}}_i}^2+{{\mathrm{\σ}}_n}^2/{{\mathrm{\σ}}_s}^2)\end{array}\right.$

in the formula, σ_n ²Is the variance of the noise, and is,i.e. the equalized output.

It should be understood that the above-mentioned embodiments of the present invention are only examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention, and it will be obvious to those skilled in the art that other variations or modifications may be made on the basis of the above description, and all embodiments may not be exhaustive, and all obvious variations or modifications may be included within the scope of the present invention.

Claims (8)

1. A channel estimation and equalization method for a rail transit high-speed moving scene is characterized by comprising the following steps:

s1, acquiring the current position of the train;

s2, constructing a channel model of the train prediction position;

s3, determining the basic impact response of the train prediction position channel based on the current position of the train;

s4, tracking the variation of the channel response of the predicted position, and quickly estimating the actual channel impact response of the predicted position;

and S5, performing channel equalization according to the actual impact response of the predicted position determined in the step S4.

2. The channel estimation and equalization method according to claim 1, wherein the step S1 comprises:

s11, the transponder arranged on the ground base station sends the ID of the transponder to the train;

s12, measuring the relative position of the train and the ground base station;

s13, determining the current position P of the train₁：P₁Transponder ID + relative position of train to said ground base station.

3. The channel estimation and equalization method of claim 2 wherein the relative position of the train and the ground base station is measured using a wheel track distance meter disposed on the train.

4. The channel estimation and equalization method according to claim 1, wherein the step S3 comprises:

storing basic parameters corresponding to each operation position of the train into a channel database in advance;

and reading corresponding basic parameters from a channel database based on the channel model of the predicted position of the train and the current position of the train, and estimating the actual channel impact response of the predicted position.

5. The channel estimation and equalization method of claim 4 wherein said base parameters comprise: the number of the multi-paths of the channel corresponding to each running position of the train and the time delay and average energy data of each path.

6. The channel estimation and equalization method according to claim 1, wherein the step S2 comprises:

s21, determining all propagation paths of the electromagnetic wave signals between the current position of the train obtained in the step S1 and the ground base station;

s22, calculating the transmission loss of each propagation path by using the path loss formula of electromagnetic wave transmission:

PL(dB)＝△₁+74.52+26.16log₁₀(f)-13.82log₁₀(h_b)-3.2log₁₀(11.75h_m)²+[44.9-6.55log₁₀(h_b)+△₂]log₁₀(D)

wherein f represents the operating frequency band, h_bAnd h_mRespectively representing the effective height of the base station antenna and the effective height of the train antenna, D representing the distance of the transmission path between the current position of the train and the ground base station, △₁And △₂Is a constant related to the transmission environment;

s23, calculating the transmission delay of each propagation path;

and S24, constructing a channel model of the current position based on the energy loss and the transmission delay of each transmission path.

7. The channel estimation and equalization method according to claim 1, wherein the step S4 comprises:

s41, based on the receiving sequence of the pilot signal transmitted in the channel, estimating the channel response at the frequency point by using the minimum mean square error criterion;

s42, obtaining responses of all points in the channel by using an interpolation algorithm, namely predicting the variation of the channel response of the position;

s43, estimating the actual channel impact response of the predicted position: and the actual channel impulse response of the predicted position is equal to the basic impulse response of the predicted position channel and the variation of the channel response of the predicted position.

8. The channel estimation and equalization method according to claim 1, wherein said step S5 includes

In the running process of the train, all the sub-channels are recorded as: y is_i＝H_ix_i+w_i；

Equalising the channel using the minimum mean square error criterion, i.e.

minE{(g_iy_i-x_i)^H(g_iy_i-x_i)}，

Then there is a change in the number of,

$\left\{\begin{array}{c}{g}_i={{\hat{H}}_i}^{*}/({{\hat{H}}_i}^2+{{\mathrm{\σ}}_n}^2/{{\mathrm{\σ}}_s}^2)\\ {\hat{x}}_i=y_i{{\hat{H}}_i}^{*}/({{\hat{H}}_i}^2+{{\mathrm{\σ}}_n}^2/{{\mathrm{\σ}}_s}^2)\end{array}\right.$

wherein, g_iTo equalize coefficients, σ_n ²In order to be the variance of the noise,to equalize the output.