CN1909403A - Pilot frequency creating method in multiple-input and multiple-output system - Google Patents

Abstract

According to the present invention, a pilot generation method in a MIMO system is proposed, the method includes: at the transmitting end, receiving the channel information fed back from the receiving end to obtain the channel fading matrix; using the channel estimation method to obtain the channel estimation Error; using the channel signal detection method, according to the output detection optimal mode, the correlation relationship between the channel estimation error and the channel matrix is obtained; according to the correlation relationship between the channel fading matrix and the pilot matrix determined by the channel signal detection method adopted, determine a pattern of pilot design; and generating pilots according to the pilot design pattern and the channel fading matrix.

Description

Pilot Generation Method in Multiple-Input Multiple-Output System

technical field

The present invention relates to a pilot generation method under a MIMO (Multiple Input Multiple Output) system, more specifically, the present invention relates to a pilot generation method under a related MIMO system, which is suitable for high-speed wireless networks under various cellular systems Communication systems and high-throughput wireless local area network systems, especially MIMO communication systems with strong correlation or slow time-varying characteristics.

Background technique

The MIMO (Multiple Input Multiple Output) multi-antenna system uses multiple antennas at the transmitting or receiving end to actively use the spatial orientation information of the user or the redundancy of the spatial channel, which can effectively increase the system capacity. As the MIMO system plays an important role in the wireless system, the pilot design in the MIMO environment has gradually become a research topic that has attracted more attention. In the MIMO environment, when using the designed pilot sequence for channel estimation, it is necessary not only to estimate the time domain information from each transmitting antenna to the receiving antenna, but also to estimate the difference of each antenna with the corresponding spatial characteristics. Therefore, when designing the pilot sequence of the MIMO system, not only the design of the pilot sequence of each transmitting antenna should be considered, but also the relationship between the pilot sequences of each antenna should be considered.

In the MIMO system, each antenna can transmit pilot signals at the same time, or some antennas can transmit pilot signals, and other antennas can transmit data signals. Fig. 1 shows several arrangements of pilot signals and data signals in each antenna and each symbol period. The gray circle in the figure represents the pilot frequency, the white circle represents the data signal, the horizontal axis is the time axis, and the values on the vertical axis correspond to different transmitting antennas. Among them, the most common transmission mode is the mode corresponding to Fig. 1(a). In several symbol periods, each antenna either transmits pilot signals or all transmits data signals. The present invention is developed based on this model.

When designing the pilot sequence, it is necessary to solve the channel estimation error matrix according to different channel estimation algorithms, and then design the pilot sequence according to the mode with the smallest mean square error of the matrix. The flow chart of traditional method pilot frequency design is as shown in Figure 3, in step 31, obtain channel estimate according to the selected channel estimation algorithm, in step 32, subtract channel true value with channel estimate, obtain channel estimation error, in step 33 , obtain the channel estimation error minimization expression, and in step 34, obtain the pilot sequence according to the expression.

However, the mean square error of channel estimation is only a measure of one aspect of estimation performance, and cannot fully reflect the quality of channel estimation. For example, if the zero-forcing method is used for channel estimation, the mean square error of the channel estimation is completely determined by the energy of the noise and the pilot sequence, and has nothing to do with the channel characteristics. Therefore, when generating the pilot sequence, it is impossible to combine the channel characteristics to further improve performance of channel estimation.

In fact, for the zero-forcing algorithm, different channel characteristics adapt to different pilot sequences. "MIMO channel estimation: optimal training and tradeoffs between estimation techniques" (ICC 2004-IEEE International Conference on Communications, vol.27, no.1, June 2004, pp.2658-2662) written by Mehrzad Biguesh, Alex B. Gershman et al. A scale zero-forcing algorithm is proposed, which multiplies a scalar scale factor with the channel estimate, uses channel information, and selects an appropriate scale factor, which can reduce the channel estimation error. Even with the same mean square error, different channel estimation errors have different effects on signal detection. The invention uses a zero-forcing signal detection algorithm to analyze the influence of channel estimation error on signal detection, and when the channel estimation error is orthogonal to the channel, the interference energy corresponding to the channel estimation error is the smallest. Therefore, it is somewhat one-sided to generate the pilot sequence only by using the channel estimation mean square error.

In the actual environment, there is a certain correlation between the channels between the transmitting and receiving antennas. In a correlated MIMO environment, the correlation characteristics of the channel change slowly with time, and this slow-varying component of the channel can be fed back to the transmitter for pilot design to generate pilots to improve the accuracy of subsequent channel estimation. In related contexts, re- A channel model is constructed, as shown in Figure 2, which expresses the channel as the product of the correlation of the transmitter, the independent channel part and the correlation of the receiver. When generating the pilot sequence, the pilot sequence is designed by combining the mutual information maximization mode of the received signal and channel fading and the minimum mean square error. However, this method needs to solve the problem of joint optimization of matrix trace and determinant, and the complexity is too high. The literature only gives approximate solutions under high and low SNR.

Contents of the invention

The present invention has been made in consideration of the above-mentioned problems. The invention relates to a method for generating pilot frequency in a correlation MIMO system, which is suitable for high-speed wireless communication systems and high-throughput wireless local area network systems under various cellular systems, especially MIMO with strong correlation or slow time-varying characteristics Communication Systems.

In order to achieve the above purpose, a pilot generation method in a MIMO system is proposed, the method includes: at the transmitting end, receiving the channel information fed back from the receiving end to obtain the channel fading matrix; using the channel estimation method to obtain the channel fading matrix Estimate error; Utilize the channel signal detection method, according to the optimal mode of output detection, obtain the correlation relationship between the channel estimation error and the channel matrix; According to the correlation relationship between the channel fading matrix and the pilot matrix determined by the channel signal detection method adopted, determining a pattern of pilot design; and generating pilots based on the pilot design pattern and the channel fading matrix.

Preferably, the channel signal detection method is a zero-forcing method.

Preferably, the optimal output detection mode is a mode that minimizes interference of channel estimation errors on detected output signals.

Preferably, one of the modes that minimize the interference of the channel estimation error on the detected output signal is the mode that minimizes the projection of the channel estimation error on the channel inverse matrix.

Preferably, the minimum mode projected by the channel estimation error on the channel inverse matrix is to minimize the energy of the product of the channel error and the pseudo-inverse of the channel matrix.

Preferably, the other mode that minimizes the channel estimation error's interference with the detected output signal is the most similar mode between the channel estimation error and the channel matrix.

Preferably, the most similar mode between the channel estimation error and the channel matrix is that the projection of the channel estimation error matrix on the channel matrix is a negative number with the largest modulus.

Preferably, based on the slow time-varying characteristics of the channel, the channel matrix is represented by a related channel matrix, and a previous channel matrix is used as a substitute for the currently required channel matrix.

Description of drawings

The above objects, advantages and features of the present invention will become apparent by referring to the following detailed description of preferred embodiments employed in conjunction with the accompanying drawings, wherein:

1(a), 1(b) and 1(c) are diagrams showing how pilot sequences and data transmission are transmitted at each antenna and each symbol period according to the prior art;

FIG. 2 is a diagram showing a virtual correlation channel model according to the prior art;

FIG. 3 is a flowchart showing a method for generating pilots based on channel estimation errors according to the prior art;

4 is a schematic diagram showing a system model for related MIMO system pilot generation according to an embodiment of the present invention;

FIG. 5 is a flowchart showing a method for generating a pilot based on signal detection according to an embodiment of the present invention;

FIG. 6 is a flow chart showing a specific implementation process of a method for generating a pilot based on signal detection according to an embodiment of the present invention;

FIG. 7 is a schematic diagram showing the principle of Channel Correlation Pilot Design Mode 1 according to an embodiment of the present invention;

FIG. 8 is a schematic diagram showing the principle of channel similarity pilot design mode 2 according to an embodiment of the present invention; and

9 and 10 are both graphs showing the results of Monte Carlo simulations.

Detailed ways

Preferred embodiments of the present invention will be described below with reference to the accompanying drawings.

In the pilot generation method of the present invention, a pilot design mode related to channel information is used, which is different from the previous pilot sequence optimization algorithm, which only starts from the channel estimation performance, and the present invention further solves the signal detection by the channel estimation value. Based on the results, the influence of channel estimation error matrix on signal detection performance is analyzed, and the channel similarity model and corresponding generation method for pilot sequence design are proposed. The present invention combines channel characteristics when generating pilot sequences, and can generate corresponding pilot sequences according to different channels.

Specifically, the present invention considers the characteristics of the channel when designing the pilot sequence, and generates corresponding pilot sequences for different channels, and its system model is shown in FIG. 4 . When generating the pilot, the transmitting end uses the channel information fed back by the receiving end through the feedback channel to adjust the phase and power of the pilot, insert it into the data, and send it out through each antenna after beamforming and radio frequency amplification.

Assume that the MIMO system has N transmitting antennas and M receiving antennas, where M≥N, there is a certain correlation between the transmitting ends of the MIMO system. After matched filtering and sampling, the received signal is X.

X＝HP+N _P (1)

Among them, each row of X corresponds to the channel received by each receiving antenna, and each column of it corresponds to the symbols received in the same symbol period. H is an M×N-dimensional channel matrix, and its elements [H] _ij correspond to the first The channel between the i transmit antenna and the j receive antenna is fading, and the wireless channel is a flat-fading Rayleigh channel. N _P is zero-mean additive Gaussian noise, spatial white noise, and obeys N(0, N ₀ I _N ) distribution. P is the pilot vector transmitted by each antenna, and the element p _ij in the ith row and the jth column corresponds to the pilot symbol sent by the i-th antenna in the j-th symbol period.

The influence of channel estimation error on signal detection will be described below.

When the traditional method designs and generates the pilot sequence, it only considers the influence of the pilot sequence on the mean square error of the channel estimation, and the channel estimation is for the signal detection service. The present invention further analyzes the influence of the channel estimation error on the signal detection, thereby A pilot sequence design pattern with the greatest similarity is obtained.

在步骤52，用信道估计

减去信道真值H，得到信道估计误差E，即信道估计误差矩阵The present invention proposes a pilot sequence design method based on signal detection, the flow chart of which is shown in Figure 5, in step 51, channel estimation is obtained according to the selected channel estimation algorithm

In step 52, use channel estimation

Subtract the channel true value H to get the channel estimation error E, which is the channel estimation error matrix

$\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}$

channel can be expressed as

$\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; =</span>\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

Here, the difference from the traditional pilot signal design method is that the present invention further considers the impact of channel estimation on signal detection when designing and generating pilot sequences, and its mode is to minimize the negative impact of channel estimation error matrix. In step 53, the output of channel detection is obtained according to different channel signal detection algorithms. If the transmitted signal is S, after signal fading, the signal received by the antenna is

Y=HS+N _s

Where N _s is the zero-mean, additive, Gaussian, and spatial white noise received by the receiving antenna. Using the zero-forcing method for signal detection, we get

$\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; =</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}}^{<span\; class="notranslate">\; +</span>}\mathrm{<span\; class="notranslate">\; Y</span>}<span\; class="notranslate">\; =</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}}^{<span\; class="notranslate">\; +</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; HS</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}<span\; class="notranslate">\; )</span>$

的伪逆， ${\hat{H}}^{+}={\hat{H}}^H{\left(\hat{H}{\hat{H}}^H\right)}^{-1}.$ 如果将(2)式代入上式，得到检测结果in channel estimation

the pseudo-inverse of ${\hat{h}}^{+}={\hat{h}}^h{\left(\hat{h}{\hat{h}}^h\right)}^{-1}.$ If (2) is substituted into the above formula, the test result is obtained

$\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; =</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}}^{<span\; class="notranslate">\; +</span>}\mathrm{<span\; class="notranslate">\; Y</span>}<span\; class="notranslate">\; =</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}}^{<span\; class="notranslate">\; +</span>}<span\; class="notranslate">\; (</span>\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; S</span>}<span\; class="notranslate">\; +</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}}^{<span\; class="notranslate">\; +</span>}{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; S</span>}<span\; class="notranslate">\; -</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}}^{<span\; class="notranslate">\; +</span>}\mathrm{<span\; class="notranslate">\; ES</span>}<span\; class="notranslate">\; +</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}}^{<span\; class="notranslate">\; +</span>}{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

造成。In step 54, based on the signal detection result, the influence of the channel estimation error on the signal detection is analyzed, and the pilot design mode is determined according to the mode with the best signal detection performance. (3) where is the estimated value of S, and the first item S on the right side is the transmitted symbol, which is the value we want to estimate. When designing the pilot sequence, we should try to increase the value of S in proportion of. The second item in the formula is related to the channel estimation error, and the third term consists of the noise

cause.

的值尽量小，由此，本发明提出如步骤63所示的模式一：信道估计误差在信道逆阵上投影最小模式，将在实施例一中介绍。若不将第二项看作干扰，当 接近于单位阵时，信道估计误差不仅不会降低检测的性能，反而会加强检测的性能，为此，本发明推导出如步骤64所示的模式二：信道估计误差与信道矩阵最相似模式，将在实施例二中介绍。最终，应用所选的模式一和模式二之一来生成导频。In step 55, the pilot is designed to generate the pilot, and the flow chart of the specific implementation method of the pilot design is shown in FIG. 6 . In step 61, an optimal mode of pilot design is determined, the main purpose of which is to reduce the negative impact of channel estimation error E on signal detection. In step 62, it is determined whether mode one or mode two is selected. If the second item As interference, in order to increase the proportion of S, it should make

The value of is as small as possible. Therefore, the present invention proposes the first mode as shown in step 63: the channel estimation error projects the minimum mode on the channel inverse matrix, which will be introduced in the first embodiment. If the second term is not considered as interference, when When it is close to the unit matrix, the channel estimation error will not only reduce the detection performance, but will strengthen the detection performance. Therefore, the present invention derives the second mode as shown in step 64: the channel estimation error is the most similar mode to the channel matrix, It will be introduced in the second embodiment. Finally, the selected one of mode one and mode two is applied to generate the pilot.

(Embodiment 1)

As mentioned above, when the second term in (3) is regarded as interference, in order to reduce the impact of this term, its energy should be minimized, and the first mode is obtained: the channel estimation error projects the minimum mode on the channel inverse matrix

min{‖EH ⁺ ‖ ² }

In mode 1, information of channel H is included, and when designing pilot sequences, corresponding pilot sequences are designed and generated according to different channel characteristics.

In the actual system, the information of the instantaneous channel matrix H cannot be obtained, and the correlation channel matrix is used instead of the instantaneous channel matrix during implementation. In a certain time interval, the channels have similar characteristics. In order to reduce the calculation amount of solving the correlation matrix, the last channel information can be used as the channel information required for optimizing the pilot sequence.

If the zero-forcing method is used for channel estimation, then the channel estimation error is

E＝N _P P ⁺

The minimum product energy of E and H ⁺ can be expressed as

$\mathrm{<span\; class="notranslate">\; min</span>}<span\; class="notranslate">\; \{</span>{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; P</span>}}{\mathrm{<span\; class="notranslate">\; P</span>}}^{<span\; class="notranslate">\; +</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}^{<span\; class="notranslate">\; +</span>}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; \}</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; min</span>}<span\; class="notranslate">\; \{</span>\mathrm{<span\; class="notranslate">\; tr</span>}<span\; class="notranslate">\; (</span>{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; P</span>}}^{<span\; class="notranslate">\; +</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}^{<span\; class="notranslate">\; +</span>}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; h</span>}}{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; P</span>}}}^{\mathrm{<span\; class="notranslate">\; h</span>}}{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; P</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; P</span>}}^{<span\; class="notranslate">\; +</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}^{<span\; class="notranslate">\; +</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; \}</span>$

$<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; min</span>}<span\; class="notranslate">\; \{</span>\mathrm{<span\; class="notranslate">\; tr</span>}<span\; class="notranslate">\; (</span>{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; P</span>}}^{<span\; class="notranslate">\; +</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}^{<span\; class="notranslate">\; +</span>}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; h</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; P</span>}}^{<span\; class="notranslate">\; +</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}^{<span\; class="notranslate">\; +</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; \}</span>$

$<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; min</span>}<span\; class="notranslate">\; \{</span>\mathrm{<span\; class="notranslate">\; tr</span>}<span\; class="notranslate">\; [</span>{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; HP</span>}<span\; class="notranslate">\; )</span>}^{{<span\; class="notranslate">\; +</span>}^{\mathrm{<span\; class="notranslate">\; h</span>}}}{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; HP</span>}<span\; class="notranslate">\; )</span>}^{<span\; class="notranslate">\; +</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \}</span>$

$<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; min</span>}<span\; class="notranslate">\; \{</span>\mathrm{<span\; class="notranslate">\; tr</span>}<span\; class="notranslate">\; [</span>{<span\; class="notranslate">\; \{</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; HP</span>}<span\; class="notranslate">\; )</span>{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; HP</span>}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; h</span>}}<span\; class="notranslate">\; \}</span>}^{<span\; class="notranslate">\; +</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \}</span>$

According to the characteristics of the trace, to minimize the product energy of E and H ⁺ , then

(HP)(HP) ^H =I.

Let P=H ⁺ , the above formula can be satisfied.

Figure 7 gives a physical explanation of the pilot design. The horizontal axis in FIG. 7 is the serial number of the antenna, and the vertical axis is the average power of each transmitting antenna. Fig. 7(a) is the average power of each antenna channel, and Fig. 7(b) is the average power of the corresponding pilot sequence of each antenna. The pilot is the pseudo-inverse of the channel matrix, that is to say, the higher the fading energy of the channel, the lower the average energy of the corresponding antenna pilot sequence. From this we can see an interesting phenomenon. When designing the pilot sequence, the channel with poorer performance will be assigned higher pilot energy, which is exactly the opposite of the requirement of the water filling algorithm.

(Example 2)

的信噪比提高。 接近于单位阵可以表示为From (3), it can be seen that when When it is close to the unit matrix, the channel estimation error will not only reduce the detection performance, but will strengthen the detection performance, so that

The signal-to-noise ratio is improved. Close to the identity matrix can be expressed as

$<span\; class="notranslate">\; -</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}}^{<span\; class="notranslate">\; +</span>}\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; cI</span>}<span\; class="notranslate">\; \&DoubleRightArrow;</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; c</span>}\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}<span\; class="notranslate">\; \&DoubleRightArrow;</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; c</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&DoubleRightArrow;</span>\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; =</span>\frac{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="c"\; filenames="A20051008939600142.png,A20051008939600161.png"\; state="\{\{state\}\}">c</figure-callout></span>}}{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; c</span>}}\mathrm{<span\; class="notranslate">\; h</span>}$

where c is a positive number. It can be seen from the above formula that when the channel estimation error and the channel matrix have similar characteristics, and the modulus of the projection of E on H is negative, the performance of signal detection can be improved. Therefore, the present invention proposes mode 2: the mode in which the channel estimation error is most similar to the channel matrix.

中，两者叠加后与S相乘，代表了检测结果的信号部分。(a)图中-e_h和i具有相同方向，两者相加后的向量具有更高的幅度，对应了(b)图中红线部分，增大了S在

中比例，提高输出信噪比。(c)图和(d)图则给出了相反的情况。Figure 8 shows the schematic diagram of the mode in this embodiment, in which vector i represents the ith row vector in the unit matrix, and -e _h represents The corresponding vector in the signal detection result

In , the two are superimposed and multiplied by S, which represents the signal part of the detection result. (a) -e _h and i have the same direction in the figure, and the vector after the addition of the two has a higher magnitude, which corresponds to the red line in (b) and increases the S in

Medium ratio, improve the output signal-to-noise ratio. Figures (c) and (d) show the opposite situation.

As mentioned above, the present invention proposes a method for generating pilots according to the pilot design pattern related to channel information, which is different from the previous pilot sequence optimization algorithm, which only starts from the channel estimation performance; the present invention further consists of channel The estimated value is used to solve the signal detection results, the influence of the channel estimation error matrix on the signal detection performance is analyzed, and the channel similarity model and corresponding design method for pilot sequence design are proposed. The present invention combines channel characteristics when designing and generating pilot sequences, and can design corresponding pilot sequences according to different channels.

It should be noted that although the zero-forcing method is used as an example to illustrate the present invention above, it is obvious that the idea of the present invention can be applied to other detection methods, as long as the channel estimation error matrix is considered in the pilot sequence design. performance impact.

Figure 9 and Figure 10 show the Monte Carlo simulation results of 10,000 times. The transmitting end and the receiving end have two antennas, the noise obeys the Gaussian distribution, the noise signals on each receiving antenna are independent of each other, the transmitted symbols adopt the BPSK modulation mode, and the transmitted symbols at each moment are independent of each other. The channel is a flat fading channel. In the simulation, the performance of signal detection after channel estimation obtained by the algorithm of the present invention and the algorithm of the prior art is compared. Both channel estimation and signal detection adopt zero-forcing algorithm, and the channel does not change for a short time. In the traditional algorithm, the pilot sequence is selected as the unit matrix.

Fig. 9 is a performance comparison between the inventive method and the traditional method under the condition that the signal-to-noise ratio is 10 dB during channel estimation and the signal-to-noise ratio is variable during signal detection. Fig. 10 is a performance comparison between the inventive method and the prior art method under the condition that the signal-to-noise ratio is 10 dB during signal detection and the signal-to-noise ratio is variable during channel estimation. When the signal-to-noise ratio of channel estimation is 10dB, the higher the signal-to-noise ratio of signal detection, the better the invention is compared with the traditional algorithm. When the signal-to-noise ratio of signal detection is 10dB, the lower the signal-to-noise ratio of channel estimation, the better the present invention is compared to traditional algorithms.

Although the present invention has been illustrated in conjunction with the preferred embodiments thereof, those skilled in the art will understand that various modifications, substitutions and alterations can be made to the present invention without departing from the spirit and scope of the invention. Accordingly, the invention should not be limited by the above-described embodiments, but by the appended claims and their equivalents.

Claims (8)

1, the pilot frequency creating method in a kind of multi-input multi-output system, described method comprises:

At transmitting terminal,

Reception obtains the channel fading matrix from the channel information of receiving terminal feedback;

Utilize channel estimation methods to obtain channel estimation errors;

Utilize the channel signal detection method, detect optimization model, obtain the incidence relation of channel estimation errors and channel matrix according to output;

According to incidence relation, determine the pattern of pilot design by determined channel fading matrix of the channel signal detection method that is adopted and pilot matrix; And

Generate pilot tone according to described pilot design pattern and described channel fading matrix.

2, method according to claim 1 is characterized in that described channel signal detection method is the ZF method.

3, method according to claim 2 is characterized in that it is to make channel estimation errors disturb minimum pattern to the signal that detects output that described output detects optimization model.

4, method according to claim 3 is characterized in that: the described channel estimation errors that makes is that channel estimation errors is at channel inverse matrix upslide shadow minimal mode to one of minimum pattern of the signal interference that detects output.

5, method according to claim 4 is characterized in that described channel estimation errors is at channel inverse matrix upslide shadow minimal mode: the long-pending energy minimum that makes channel errors and channel matrix pseudoinverse.

6, method according to claim 3 is characterized in that: described to make channel estimation errors disturb another minimum pattern to the signal that detects output be the parallel pattern of channel estimation errors and channel matrix.

7, method according to claim 6 is characterized in that the parallel pattern of described channel estimation errors and channel matrix is to make the negative that be projected as mould value maximum of channel estimation errors matrix on channel matrix.

8, according to any described method of claim 1 to 7, it is characterized in that the slow time-varying characteristics based on channel, described channel matrix adopts the correlated channels matrix to represent, and adopts previous channel matrix to be used as substituting of current required channel matrix.

Images