CN101359926A - Receiver, channel estimation method and device - Google Patents

Abstract

The present invention provides a receiver, channel estimation method and device, wherein the channel estimation method includes: performing multipath search on the pilot channel, and selecting a path whose energy in the channel window is greater than a preset condition; Estimate; remove the raised cosine interference from the channel estimate obtained by channel estimation to obtain a channel estimate without raised cosine interference; perform raised cosine convolution on the channel estimate without raised cosine interference to obtain all positions in the pilot channel diameter energy. When the present invention performs channel estimation, no new noise will be introduced, effectively ensuring the equalization performance of the LMMSE equalizer; in addition, due to the existence of RC influence, the interference between paths is avoided when performing RC convolution, thereby in channel estimation The deviation between the value and the actual channel value further ensures the equalization effect of the LMMSE equalizer.

Description

Receiver, channel estimation method and device

technical field

The present invention relates to communication technology, especially a receiver for eliminating interference between paths, a channel estimation method for channel estimation with the receiver and a channel estimation device for channel estimation in the receiver.

Background technique

In a Wideband Code Division Multiple Access (WCDMA) system, the multi-code signals at the transmitter are mutually orthogonal. However, when these signals propagate through multipath fading channels, the orthogonality weakens or disappears. In order to correctly identify the signals, a RAKE receiver is usually introduced at the receiving end to restore the orthogonality of the signals. Because of the wideband spread spectrum signal, WCDMA system can use RAKE receiver to distinguish different multipath components in the received signal, realize diversity reception, effectively overcome the influence of multipath fading, and improve the receiving performance of the system. However, in the downlink reception of the WCDMA system, intersymbol interference and multi-user interference caused by multipath limit the coherent reception performance of the RAKE receiver, especially in the case of high data transmission rates, such as: high-speed downlink packet access (High Speed Downlink Packet Access, hereinafter referred to as: HSDPA), due to the small spreading factor, not only the anti-interference ability of the RAKE receiver is reduced, but also the diversity gain introduced by multipath combination is also greatly reduced. At this time, in order to obtain a more ideal performance, the existing technology introduces the Linear Minimum Mean Square Error (Linear Minimum Mean Square Error, hereinafter referred to as: LMMSE) technology to realize the LMMSE receiver, and the LMMSE equalizer is used in the LMMSE receiver. The signal is equalized. The equalization principle of the LMMSE equalizer is that the mean square error between the equalized signal and the transmitted signal is the smallest. Since the intersymbol interference and multi-user interference caused by multipath are eliminated, the receiving performance of the LMMSE equalizer is better than that of the original RAKE receiver.

As shown in FIG. 1 , it is a schematic structural diagram of a WCDMA signal transceiving system in the prior art. At the transmitting end, the root raised cosine (Root Raised Cosine, hereinafter referred to as: RRC) filter at the transmitting end filters the signal x(k) and then transmits; the transmitted signal is transmitted through a multipath fading channel superimposed with Gaussian white noise, and then arrives at the receiving end The LMMSE receiver at the end obtains the signal y(k) after being filtered by the receiving end RRC filter in the LMMSE receiver, and sends the signal y(k) to the LMMSE equalizer and the channel estimation device respectively; the channel estimation device is based on the signal y( k) Perform channel estimation, estimate the energy of each path in the fading channel, and send the estimated channel estimation value h channel estimation matrix to the computing device. When there are multiple paths, the channel estimation value h is expressed as a channel estimation matrix H, at the same time, the noise function estimation device estimates the Gaussian white noise energy in the channel, and sends the estimated noise energy σ _n ² to the computing device; the computing device estimates the value h and the Gaussian white noise energy σ _n according to the channel estimation matrix ² Perform calculations to obtain the tap weight w _d of the LMMSE equalizer and send it to the LMMSE equalizer; the LMMSE equalizer equalizes y(k) and w _d according to the equalization principle to obtain the signal x′(k); descrambling The despreading device descrambles and despreads the signal x'(k) to obtain the transmitted signal x(k).

From the processing process of the received signal by the LMMSE receiver in Figure 1, it can be seen that the tap weight w _d of the LMMSE equalizer is determined by the channel estimated value h and the Gaussian white noise energy σ _n ² , the difference between the channel estimated value h and the actual channel value, It directly affects the performance of the LMMSE equalizer. The closer the channel estimate is to the actual channel situation, the better the equalization effect of the LMMSE equalizer will be. Therefore, how to perform channel estimation and obtain a channel estimation value close to the actual channel value is very critical.

The traditional channel estimation method directly estimates the channel fading of the pilot channel on the path position according to the multipath search indication. As shown in Figure 2, it is a schematic structural diagram of a traditional channel estimation device, and its flow process for channel estimation is as follows: when the multipath search module detects that there is a path in the pilot channel, it sends a common pilot channel (Common Pilot Channel, hereinafter referred to as: CPICH) descrambling and despreading module sends multipath search instruction information, and the multipath search instruction information includes the initial path position and channel length information of the pilot channel that needs channel estimation, because the initial path position is the beginning of the channel position, combined with the channel length, the channel window for which channel estimation needs to be performed can be known, thus, the end position of the channel window can also be known; the CPICH descrambling and despreading module performs descrambling, descrambling and spread to obtain pilot symbols; the division module divides the pilot symbols obtained by descrambling and despreading by the transmitted pilot symbols to obtain channel estimation values and send them to Finite Impulse Response (Finite Impulse Response, hereinafter referred to as: FIR) filtering The FIR filter smoothes the received channel estimate before and after; since the pilot signal is multiplied by a power weighting factor at the transmitting end, the current channel estimate is different from the actual channel estimate by a power weighting factor, and the power The weighting factor processing module divides the signal processed by the FIR filter by the power weighting factor to obtain the channel estimation value CE of each path position in the channel window.

When using the above-mentioned traditional channel estimation method for channel estimation, the path energy of all positions in the channel window is considered. Since the path energy of some positions is weak, and the signal-to-noise ratio of the position with weak energy is usually low, therefore, traditional channel estimation The method introduces a large noise when channel estimation, which affects the equalization performance of the LMMSE equalizer and makes its equalization performance decline.

In order to overcome the performance degradation of the LMMSE equalizer caused by the introduction of large noise in the traditional channel estimation method, the existing technology uses 1/4 chip as the precision to select the signal-to-noise ratio from the channel window detected by the multipath search module to reach the preset Threshold single path, and then use the traditional channel estimation method to perform channel estimation on this single path; since the equivalent channel to be estimated is the convolution of the actual channel, the receiving RRC filter, and the transmitting RRC filter, and the two RRC filters in series Together, they are equivalent to a raised cosine filter. The parameter matrix of the raised cosine filter is represented by RC, and the estimated channel estimation value is convoluted with RC to obtain the channel estimation of all phase points in the channel window. The accuracy of the phase point Also 1/4 chip. For complexity considerations, the number of taps in one chip of the equalizer is generally 2 or 1, and channel estimation at the chip level of 1 or 1/2 is required. Therefore, the 1/4 chip after convolution The precision channel estimation value is down-sampled to obtain a chip-level or 1/2 chip-level channel estimation value.

The inventor finds in the process of realizing the present invention: the above-mentioned method for channel estimation of a single path selected with 1/4 chip as the precision is compared with the traditional channel estimation method, although avoiding the LMMSE equalizer performance due to the introduction of noise However, since the estimated channel estimate has been mixed with the influence of RC, the estimated channel estimate is convoluted with RC at this time, which further introduces interference between paths, which will lead to There is a deviation between the estimated channel value of the channel and the actual channel value, so that the equalization effect of the LMMSE equalizer is reduced, especially when there are two paths with strong energy in the channel and the distance is less than 1 chip, due to RC The greater the interference between paths, the worse the channel estimation effect will be.

Contents of the invention

The technical problem to be solved by the embodiments of the present invention is to eliminate the interference between the channel estimation value and the path introduced by the convolution of the RC system due to the influence of RC, and then avoid the channel estimation after convolution due to the interference between the paths. There is a deviation between the value and the actual channel, and the equalization effect of the LMMSE equalizer is improved.

A channel estimation method provided according to an embodiment of the present invention includes:

Perform multipath search on the pilot channel, and select the path whose energy in the channel window is greater than the preset condition;

performing channel estimation on the path;

Eliminate the raised cosine interference from the channel estimation value obtained by channel estimation to obtain a channel estimation value without raised cosine interference;

Performing raised cosine convolution on the channel estimate without raised cosine interference to obtain the path energy at all positions in the pilot channel.

According to another aspect of the present invention, a channel estimation device is provided, including:

A multipath search module is used to perform multipath search on the pilot channel;

A selection module, configured to select a path with energy greater than a preset condition from the channel window obtained by multipath search;

A channel estimation module, configured to perform channel estimation on a path whose energy is greater than a preset condition;

A restoration module, used for eliminating raised cosine interference from the channel estimation value obtained by channel estimation;

The convolution module is configured to perform raised cosine convolution on the channel estimate without raised cosine interference obtained after removing the raised cosine interference.

A receiver provided according to yet another embodiment of the present invention includes a root-raised cosine filter, a channel estimation device, a noise power estimation device, a weight calculation device, an equalizer, and a descrambling and despreading device, and the channel estimation device include:

A multipath search module is used to perform multipath search on the pilot channel;

A selection module, configured to select a path with energy greater than a preset condition from the channel window obtained by multipath search;

A channel estimation module, configured to perform channel estimation on a path whose energy is greater than a preset condition;

A restoration module, used for eliminating raised cosine interference from the channel estimation value obtained by channel estimation;

The convolution module is configured to perform raised cosine convolution on the channel estimate without raised cosine interference obtained after removing the raised cosine interference.

When performing channel estimation in the embodiment of the present invention, channel estimation is only performed on the path whose energy in the selected channel window is greater than the preset condition, and no new noise will be introduced, effectively ensuring the equalization performance of the LMMSE equalizer; in addition, the traditional channel The channel estimation value estimated by the estimation method is convoluted after eliminating the influence of RC, which avoids the interference between the paths caused by the RC convolution after the RC influence, so that the channel estimation value is closer to the actual channel value, effectively The equalization effect of the LMMSE equalizer is guaranteed.

The technical solutions of the embodiments of the present invention will be described in further detail below with reference to the drawings and embodiments.

Description of drawings

FIG. 1 is a schematic structural diagram of a WCDMA signal transceiving system in the prior art.

Fig. 2 is a schematic structural diagram of a conventional channel estimation device in the prior art.

Fig. 3 is a flowchart of an embodiment of the channel estimation method of the present invention.

Fig. 4 is a flow chart of another embodiment of the channel estimation method of the present invention.

Fig. 5 is a schematic diagram of energy interception in the embodiment shown in Fig. 4 of the present invention.

FIG. 6 is a schematic structural diagram of an embodiment of a channel estimation apparatus according to the present invention.

Fig. 7 is a schematic structural diagram of a receiver embodiment of the present invention.

Detailed ways

In the embodiments of the present invention, channel estimation is performed on paths with energy greater than a preset condition in the channel window, and RC convolution is performed after eliminating RC interference on the channel estimation value to obtain path energies at all positions in the pilot channel.

As shown in Figure 3, it is a flowchart of an embodiment of the channel estimation method of the present invention, which includes the following steps:

Step 101, perform a multipath search on the pilot channel, and search for a path whose energy in the channel window satisfies a preset condition;

Step 102, performing channel estimation on paths whose energy satisfies a preset condition.

Channel estimation is only performed on the selected path whose energy meets the preset conditions, and no new noise is introduced, which effectively guarantees the equalization performance of the LMMSE equalizer;

Step 103, eliminating the RC influence on the channel estimation value obtained by channel estimation, and obtaining the channel estimation value of the corresponding path without RC interference;

Step 104, convolving the channel estimation value of the corresponding path without RC interference with the RC filter parameter matrix to obtain the path energy at all positions in the pilot channel.

After eliminating the influence of RC on the channel estimation value estimated by the channel, it is then convolved, which avoids the interference between paths introduced during RC convolution due to the existence of RC influence, so that there is a gap between the channel estimation value and the actual channel value A deviation is generated to make the estimated channel value closer to the actual channel value, further ensuring the equalization effect of the LMMSE equalizer.

As shown in FIG. 4, it is a flow chart of another embodiment of the channel estimation method of the present invention, which includes the following steps:

Step 201, perform multipath search on the pilot channel, measure the energy of each path in the searched channel window, select the path with the strongest energy in the channel window, and the path with the strongest energy is also called the main path, and delete the path There are two phase points on the left and right sides of the main diameter.

Step 202, except for the path with the strongest energy in the selected channel window, the path with the strongest energy among the remaining paths, that is, the path with stronger energy, the path with stronger energy is also called the secondary path, and delete the path with stronger energy There are three phase points on the left and right sides of the diameter. In the same way, multiple secondary paths with the strongest energy are sequentially selected from the remaining paths and three phase points on the left and right are deleted. The total number of selected secondary paths and main paths can be 8, or it can be other preset quantities.

Due to the impact of the RC system on multipath, the actual multipath is not the impulse response of a single path, and the path with stronger energy still has greater energy at a path position that is less than one chip away from the path. When performing multipath search, for the selected path position whose signal-to-noise ratio reaches the preset threshold, the left and right phase points at 3/4 chips are respectively deleted. Among the paths that meet the preset conditions found by multipath search, since the subsequent steps of this scheme can remove the interference of the RC system and restore the energy of the single path, it is beneficial to retain the channel estimates on both sides of the main path to more accurately simulate the channel . When removing the interference of the RC system, the inversion of the RC matrix must meet the condition of diagonal dominance. If you choose to remove the interference from the phase point diameters within 1/2 chips on the left and right of the main diameter, it is easy to cause the matrix to be ill-conditioned. inverse. For this reason, in this embodiment, choose to delete two phase points on the left and right sides of the main path, that is, delete the phase points in each 1/2 chip on the left and right sides of the main path, and delete three phase points on the left and right sides of the secondary path , one phase point is 1/4 chip, and there are four phase points in one chip.

In addition, three phase points on the left and right of the main diameter may be deleted, and two phase points on the left and right of the minor diameter may be deleted. However, since the main path has the strongest energy, compared to deleting three phase points on the left and right, deleting two phase points on the left and right can retain more energy on the main path, which is beneficial to obtain a more accurate channel estimation value.

Step 203 , from the selected path with the strongest energy and the path with stronger energy, select a path with energy greater than a preset condition.

Wherein, the condition that the energy is greater than a preset may specifically be: the energy or the signal-to-noise ratio is greater than a preset threshold, or the ratio of the energy on the path to the energy of the smallest path is greater than a preset value.

Step 204, perform channel estimation on each path whose energy satisfies a preset condition, and obtain a channel estimation value CE. Specifically, a traditional channel estimation device as shown in FIG. 2 is used for channel estimation.

Step 205, due to the RC interference between the paths, there will be interference from other paths to this path on the channel estimation value of each path obtained in step 204, since the RC parameters are known, the channel estimation is performed by the method of least squares The value CE eliminates the RC influence, and obtains the estimated CE' of the multipath without RC influence.

Assume that the path positions searched by multipath are: P1, P2, P3, P4, P5, P6, P7 and P8; the delay difference between the above paths is: ${\mathrm{\τ}}_{i,j}=\underset{i=1}{\overset{8}{\mathrm{\Σ}}}\underset{j=1;\mathrm{the\; ji}}{\overset{8}{\mathrm{\Σ}}}(P_i-P_j);$ The estimated channel values on each path obtained in step 204 are: CE ₁ , CE ₂ , CE ₃ , CE ₃ , CE ₄ , CE ₅ , CE ₆ , CE ₇ , and CE ₈ ; The estimated value of the channel without RC influence on is: CE' ₁ , CE' ₂ , CE' ₃ , CE' ₃ , CE' ₄ , CE' ₅ , CE' ₆ , CE' ₇ , CE'₈;

RC is the filter parameter matrix of RRC convolution RRC, then the RC influence can be eliminated by the following equation:

$\left[\begin{array}{cccccccc}\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; RC</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&tau;</span>}}_{\mathrm{<span\; class="notranslate">\; 1,2</span>}}<span\; class="notranslate">\; )</span>& \mathrm{<span\; class="notranslate">\; RC</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&tau;</span>}}_{\mathrm{<span\; class="notranslate">\; 1,3</span>}}<span\; class="notranslate">\; )</span>& \mathrm{<span\; class="notranslate">\; RC</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&tau;</span>}}_{\mathrm{<span\; class="notranslate">\; 1,4</span>}}<span\; class="notranslate">\; )</span>& \mathrm{<span\; class="notranslate">\; RC</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&tau;</span>}}_{\mathrm{<span\; class="notranslate">\; 1,5</span>}}<span\; class="notranslate">\; )</span>& \mathrm{<span\; class="notranslate">\; RC</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&tau;</span>}}_{\mathrm{<span\; class="notranslate">\; 1,6</span>}}<span\; class="notranslate">\; )</span>& \mathrm{<span\; class="notranslate">\; RC</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&tau;</span>}}_{\mathrm{<span\; class="notranslate">\; 1,7</span>}}<span\; class="notranslate">\; )</span>& \mathrm{<span\; class="notranslate">\; RC</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&tau;</span>}}_{\mathrm{<span\; class="notranslate">\; 1,8</span>}}<span\; class="notranslate">\; )</span>\\ \mathrm{<span\; class="notranslate">\; RC</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&tau;</span>}}_{\mathrm{<span\; class="notranslate">\; 2,1</span>}}<span\; class="notranslate">\; )</span>& \mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; RC</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&tau;</span>}}_{\mathrm{<span\; class="notranslate">\; 2,3</span>}}<span\; class="notranslate">\; )</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>\\ <span\; class="notranslate">\; .</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">\; 1</span>}& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>\\ <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</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">\; 1</span>}& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>\\ <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</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">\; 1</span>}& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>\\ <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</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">\; 1</span>}& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>\\ <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</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">\; 1</span>}& \mathrm{<span\; class="notranslate">\; RC</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&tau;</span>}}_{\mathrm{<span\; class="notranslate">\; 7,8</span>}}<span\; class="notranslate">\; )</span>\\ \mathrm{<span\; class="notranslate">\; RC</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&tau;</span>}}_{\mathrm{<span\; class="notranslate">\; 8,1</span>}}<span\; class="notranslate">\; )</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</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">\; RC</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&tau;</span>}}_{\mathrm{<span\; class="notranslate">\; 8,7</span>}}<span\; class="notranslate">\; )</span>& \mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right]\left[\begin{array}{c}{{\mathrm{<span\; class="notranslate">\; CE</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\\ {{\mathrm{<span\; class="notranslate">\; CE</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\\ {{\mathrm{<span\; class="notranslate">\; CE</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}\\ {{\mathrm{<span\; class="notranslate">\; CE</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}_{\mathrm{<span\; class="notranslate">\; 4</span>}}\\ {{\mathrm{<span\; class="notranslate">\; CE</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}_{\mathrm{<span\; class="notranslate">\; 5</span>}}\\ {{\mathrm{<span\; class="notranslate">\; CE</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}_{\mathrm{<span\; class="notranslate">\; 6</span>}}\\ {{\mathrm{<span\; class="notranslate">\; CE</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}_{\mathrm{<span\; class="notranslate">\; 7</span>}}\\ {{\mathrm{<span\; class="notranslate">\; CE</span>}}^{\text{<span class="notranslate"> \&prime;</span>}}}_{\mathrm{<span\; class="notranslate">\; 8</span>}}\end{array}\right]<span\; class="notranslate">\; +</span>\left[\begin{array}{c}{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\\ {\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\\ {\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}\\ {\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; 4</span>}}\\ {\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; 5</span>}}\\ {\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; 6</span>}}\\ {\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; 7</span>}}\\ {\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; 8</span>}}\end{array}\right]<span\; class="notranslate">\; =</span>\left[\begin{array}{c}{\mathrm{<span\; class="notranslate">\; CE</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\\ {\mathrm{<span\; class="notranslate">\; CE</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\\ {\mathrm{<span\; class="notranslate">\; CE</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}\\ {\mathrm{<span\; class="notranslate">\; CE</span>}}_{\mathrm{<span\; class="notranslate">\; 4</span>}}\\ {\mathrm{<span\; class="notranslate">\; CE</span>}}_{\mathrm{<span\; class="notranslate">\; 5</span>}}\\ {\mathrm{<span\; class="notranslate">\; CE</span>}}_{\mathrm{<span\; class="notranslate">\; 6</span>}}\\ {\mathrm{<span\; class="notranslate">\; CE</span>}}_{\mathrm{<span\; class="notranslate">\; 7</span>}}\\ {\mathrm{<span\; class="notranslate">\; CE</span>}}_{\mathrm{<span\; class="notranslate">\; 8</span>}}\end{array}\right]<span\; class="notranslate">\; ;</span>$

The above matrix equation can be abbreviated as: RC·CE'+n=CE, wherein, n is the noise coefficient of each above-mentioned path position.

最小。其中需求解的未知参数CE′的个数与方程的个数相等，通过最小二乘的方法可得到其解为：CE′＝RC^-1CE；When each component of the error vector has the same variance and each component is uncorrelated, the least square method can be used to obtain the smallest mean square error. For the above equation of eliminating inter-radius RC interference, the sum of the squares of the errors is minimized by the criterion of least squares, namely:

minimum. The number of unknown parameters CE' to be solved is equal to the number of equations, and the solution can be obtained by the method of least squares: CE'=RC ^-1 CE;

The RC matrix is further inverted to obtain the channel estimation value CE' of the path at each position before the RC influence. Since the RC matrix is a real symmetric matrix, the Gauss-Siedel (Gauss-Siedel) inversion method, the Gaussian elimination decomposition inversion method, or the Cholesky decomposition inversion method can be used to invert the RC matrix, so that Obtain the channel estimation value CE' of the corresponding position where the RC interference is eliminated.

Step 206, convolving the channel estimate CE' of the path without RC interference with the RC filter parameter matrix to obtain the path energy at all positions in the pilot channel.

信道估计值CE′与RC滤波器参数矩阵的卷积会造成能量的拖尾，按照能量由大到小的顺序，从卷积后得到能量序列

中，截取与卷积前CE′的在信道上的长度相同的部分能量。Step 207, since the pilot channel adopts 1/4 sampling rate, the channel estimation value CE' obtained after eliminating RC interference is convolved with the RC filter parameter matrix to obtain a 1/4 chip-level channel estimation value

The convolution of the channel estimation value CE′ and the RC filter parameter matrix will cause energy tailing, and the energy sequence is obtained from the convolution in the order of energy from large to small

In , intercept the same part of energy as the length of CE' on the channel before convolution.

As shown in FIG. 5 , it is a schematic diagram of energy interception in the present invention. Among them, Figure 5-1 shows the energy of CE' on each path position before convolution, assuming that the channel length is L; Figure 5-2 shows the energy of the RC filter on the corresponding path position, assuming that the channel length is M; Figure 5-3 shows the energy length of the convolution of the channel estimate CE′ and the RC filter parameter matrix. The channel length becomes L+M. In order to retain all the energy as much as possible, the larger energy part with the channel length L is intercepted. .

Step 208, the precision of the channel estimate that step 207 obtains is 1/4 chip, because the equalizer tap is generally one or two sampling points in one chip, so the channel estimate that needs also should be one or two sampling points in one chip. Two sampling points. For this purpose, the channel estimation values at all positions in the pilot channel are down-sampled to obtain the channel estimation value CE' of one chip or 1/2 chip.

In step 209, a channel estimation matrix is generated from the channel estimation value CE'.

In addition, the number of equations can also be increased so that the number of equations is greater than the number of unknowns, and the minimum mean square error (Minimum Mean Square Error, hereinafter referred to as: MMSE) or other least squares methods are used to eliminate the influence of RC on the channel estimation value CE , get the estimated CE' of multipath without RC influence.

Assume that through steps 201 and 202, the path positions searched by multipath are in order of energy: P1, P2, P3, P4, P5, P6, P7 and P8, and only path positions P1 and P2 meet the requirements in step 203. Preset conditions, select an additional path position for channel estimation from path positions P3, P4, P5, P6, P7, and P8, assuming that P3 is selected, and use the channel estimation values CE ₁ , CE ₂ of path positions P1, P2, and P3 and CE ₃ to restore the channel estimation values CE' ₁ and CE' ₂ of the out-path positions P1 and P2 without RC influence. The RC effect can be eliminated by the following equation:

$\left[\begin{array}{cc}\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; RC</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&tau;</span>}}_{\mathrm{<span\; class="notranslate">\; 1,2</span>}}<span\; class="notranslate">\; )</span>\\ \mathrm{<span\; class="notranslate">\; RC</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&tau;</span>}}_{\mathrm{<span\; class="notranslate">\; 2,1</span>}}<span\; class="notranslate">\; )</span>& \mathrm{<span\; class="notranslate">\; 1</span>}\\ \mathrm{<span\; class="notranslate">\; RC</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&tau;</span>}}_{\mathrm{<span\; class="notranslate">\; 3,1</span>}}<span\; class="notranslate">\; )</span>& \mathrm{<span\; class="notranslate">\; RC</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&tau;</span>}}_{\mathrm{<span\; class="notranslate">\; 2,3</span>}}<span\; class="notranslate">\; )</span>\end{array}\right]\left[\begin{array}{c}\mathrm{<span\; class="notranslate">\; C</span>}{{\mathrm{<span\; class="notranslate">\; E.</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\\ {{\mathrm{<span\; class="notranslate">\; CE</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\end{array}\right]<span\; class="notranslate">\; +</span>\left[\begin{array}{c}{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\\ {\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\\ {\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}\end{array}\right]<span\; class="notranslate">\; =</span>\left[\begin{array}{c}{\mathrm{<span\; class="notranslate">\; CE</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\\ {\mathrm{<span\; class="notranslate">\; CE</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\\ {\mathrm{<span\; class="notranslate">\; CE</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}\end{array}\right]<span\; class="notranslate">\; ;</span>$

的信道估计矩阵H′为：Using the method of least squares, it can be obtained

The channel estimation matrix H' of is:

的信道估计矩阵H′为：Using the MMSE method, it can be obtained

The channel estimation matrix H' of is:

Among them, σ _n ² is the noise energy, and I ₃ is the identity matrix.

As shown in FIG. 6 , it is a schematic structural diagram of an embodiment of the channel estimation device of the present invention. The channel estimation device includes a multipath search module 301 , a selection module 302 , a channel estimation module 303 , a restoration module 304 and a convolution module 305 connected in sequence. Wherein, the multipath search module 301 is used for carrying out multipath search to the pilot channel; The selection module 302 is used for selecting the path that energy is greater than the preset condition from the channel window that the multipath search goes out; The channel estimation module 303 is used for energy The path greater than the preset condition is used for channel estimation; the restoration module 304 is used to eliminate RC interference to the channel estimation value obtained by channel estimation; the convolution module 305 is used to perform RC convolution on the channel estimation value without RC interference obtained after eliminating the RC interference product.

In the channel estimation device shown in FIG. 6 , the selection module 302 may be specifically composed of a measurement unit 3021 and a selection unit 3022, wherein the measurement unit 3021 is connected to the multipath search module 301 for measuring the channel searched by the multipath search module 301 The energy or signal-to-noise ratio of the inner diameter of the window. Further, the measurement unit 3021 can also sequentially select the path with the strongest energy in the channel window and the path with stronger energy among the remaining paths; the selection unit 3022 is connected with the channel estimation module 303 for According to the measurement result of the measurement unit 3021 , a path whose energy in the channel window is greater than a preset condition is selected and sent to the channel estimation module 303 . The channel estimation apparatus in this embodiment can be used to implement the channel estimation method in the embodiment shown in FIG. 3 .

Further, the selection module may also include a deletion unit 3023, which is arranged between the measurement unit 3021 and the selection unit 3022, and is used to delete and select two phase points on the left and right of the main path with the strongest energy selected by the measurement unit 3021, and the energy is relatively low. There are three phase points on the left and right sides of the strong secondary diameter.

In addition, the channel estimation device shown in FIG. 6 may also include an information storage module 306, which is connected to the selection unit 3022 and is used to store condition information of a path with energy greater than a preset condition, for example: energy greater than a preset value A, or signal-to-noise The ratio is greater than the preset value B, or the ratio to the smallest path energy is greater than the preset value C; specifically, the selection unit 3022 is used to store the energy or signal-to-noise ratio and the information of the main path and the secondary path sent by the measurement unit 3021 The corresponding condition information in module 306 is compared, and the path whose energy in the channel window is greater than the preset condition is selected.

Further, the above-mentioned channel estimation device provided by the embodiment of the present invention may also include a parameter storage module 307, which is respectively connected to the restoration module 304 and the convolution module 305, and is used to store the RC filter parameter matrix; specifically, the restoration module 304 is used to Obtain the RC filter parameter matrix from the parameter storage module 307, and calculate the product of the channel estimate obtained by the inverse matrix of the RC filter parameter matrix and the channel estimate, to eliminate the RC interference for the channel estimate obtained by the channel estimate; convolution module 305 is used to obtain the RC filter parameter matrix from the parameter storage module 307, and perform raised cosine convolution on the estimated channel value without RC interference and the RC filter parameter matrix.

Referring again to FIG. 6 , the channel estimation apparatus according to the embodiment of the present invention may further include a downsampling module 308 for downsampling the path energy at all positions in the pilot channel obtained after RC convolution. Furthermore, the above channel estimation apparatus may further include a matrix generation module 309, configured to generate a channel estimation matrix according to the channel estimation value of 1/2 chip obtained by downsampling. The channel estimation apparatus in this embodiment can be used to implement the channel estimation method in the embodiment shown in FIG. 4 .

Referring to FIG. 7, it is a schematic structural diagram of an embodiment of a receiver of the present invention, the receiver of this embodiment includes an RRC filter, a channel estimation device, a weight calculation device, an equalizer, a descrambling and despreading device, and a weight The noise power estimation device connected to the value calculation device, the output end of the RRC filter is also connected to the input end of the equalizer. Wherein, the channel estimation device can adopt the channel estimation device of any embodiment as shown in Fig. 6, specifically, the output end of the RRC filter is connected with the multipath search module 301 in the channel estimation device, and the convolution in the channel estimation device The output terminal of the module 305 or the down-sampling module 308 is connected with the input terminal of the weight calculation device.

Specifically, the RRC filter is used to receive the transmitted signal transmitted through the multipath fading channel superimposed with Gaussian white noise, and filter the received signal and send it to the equalizer and the multipath search module 301 in the channel estimation device respectively;

The channel estimation device uses the method shown in Figure 3 or Figure 4 to perform channel estimation based on the filtered signal, and sends the estimated channel estimation value, that is, the path energy or channel estimation matrix at all positions in the pilot channel Giving a weight calculation device;

The noise power estimating device estimates the Gaussian white noise energy in the channel, and sends the estimated noise energy to the computing device;

The computing device calculates the channel estimation value sent by the channel estimating device and the noise energy sent by the noise power estimating device, obtains tap weights of the equalizer and sends them to the equalizer;

The equalizer equalizes the filtered signal sent by the RRC filter and the tap weights of the equalizer according to the equalization principle, and sends the equalized signal obtained after equalization to the descrambling and despreading device. The equalizer is an LMMSE equalizer;

The descrambling and despreading device descrambles and despreads the equalized signal to obtain the original transmission signal sent by the transmitter.

When performing channel estimation in the embodiment of the present invention, channel estimation is only performed on the path whose energy in the selected channel window is greater than the preset condition, and no new noise will be introduced, effectively ensuring the equalization performance of the LMMSE equalizer; in addition, the traditional channel After the channel estimation value estimated by the estimation method eliminates the influence of RC, it is then convoluted, which avoids the interference between paths due to the existence of RC influence during RC convolution, so that the difference between the channel estimation value and the actual channel value There is a deviation between them, which further ensures the equalization effect of the LMMSE equalizer.

Finally, it should be noted that: the above examples are only used to illustrate the technical solution of the present invention, rather than limiting the understanding of the present invention. Although the present invention has been described in detail with reference to the above-mentioned preferred embodiments, those skilled in the art should understand that: it can still modify or replace the technical solution of the present invention, and such modification or replacement does not depart from the technology of the present invention. The spirit and scope of the programme.

Claims (11)

1. A channel estimation method, characterized in that, comprising the following steps: Perform multipath search on the pilot channel, and select the path whose energy in the channel window is greater than the preset condition; performing channel estimation on the path; Eliminate the raised cosine interference from the channel estimation value obtained by channel estimation to obtain a channel estimation value without raised cosine interference; Performing raised cosine convolution on the channel estimate without raised cosine interference to obtain the path energy at all positions in the pilot channel. 2. The channel estimation method according to claim 1, wherein the energy is greater than a preset condition specifically includes: the energy is greater than a preset value, or the signal-to-noise ratio is greater than a preset value, or the ratio of the energy to the smallest path is greater than a preset value. set value. 3. The channel estimation method according to claim 1 or 2, characterized in that said performing multipath search on the pilot channel and selecting a path with energy greater than a preset condition in the channel window comprises: performing multipath search on the pilot channel Path search, select the main path with the strongest energy from the paths obtained by the multipath search operation and delete two or three phase points on the left and right of the main path; select the secondary path with the strongest energy from the positions of the remaining paths and delete the secondary path Two or three phase points on the left and right, select and delete two or three phase points on the left and right of multiple secondary paths in turn; select the path with energy greater than the preset condition from the selected main path and multiple secondary paths . 4. The channel estimation method according to claim 3, wherein the elimination of raised cosine interference from the channel estimation value obtained by channel estimation is specifically: using the least squares method to calculate the inverse matrix of the raised cosine filter parameter matrix and the channel The estimated channel estimation values are quadratured to eliminate the raised cosine interference of the channel estimation values. 5. The channel estimation method according to any one of claims 1 to 4, characterized in that, after obtaining the path energies at all positions in the pilot channel, further comprising: Carry out down-sampling to the path energy on all positions in the pilot channel, obtain the channel estimation value of 1/2 chip; generating a channel estimation matrix according to the channel estimation value of the 1/2 chip. 6. A channel estimation device, characterized in that it comprises: A multipath search module is used to perform multipath search on the pilot channel; A selection module, configured to select a path with energy greater than a preset condition from the channel window obtained by multipath search; A channel estimation module, configured to perform channel estimation on a path whose energy is greater than a preset condition; A restoration module, used for eliminating raised cosine interference from the channel estimation value obtained by channel estimation; The convolution module is configured to perform raised cosine convolution on the channel estimation value obtained after eliminating the raised cosine interference. 7. The channel estimation device according to claim 6, further comprising: The parameter storage module is used for storing the raised cosine filter parameter matrix; The restoring module is used for calculating the product of the inverse matrix of the raised cosine filter parameter matrix and the channel estimation value obtained by channel estimation, so as to eliminate the raised cosine interference for the channel estimation value obtained by channel estimation. The convolution module is configured to perform raised cosine convolution on the raised cosine interference-eliminated channel estimate and the raised cosine filter parameter matrix. 8. The channel estimation device according to claim 6 or 7, further comprising: A downsampling module is used to downsample the path energy at all positions in the pilot channel obtained after the raised cosine convolution; The matrix generating module is configured to generate a channel estimation matrix according to the channel estimation value of 1/2 chip obtained by downsampling. 9. A receiver, comprising a root-raised cosine filter, a channel estimation device, a noise power estimation device, a weight calculation device, an equalizer, and a descrambling and despreading device, wherein the channel estimation device includes: A multipath search module is used to perform multipath search on the pilot channel; A selection module, configured to select a path with energy greater than a preset condition from the channel window obtained by multipath search; A channel estimation module, configured to perform channel estimation on a path whose energy is greater than a preset condition; A restoration module, used for eliminating raised cosine interference from the channel estimation value obtained by channel estimation; The convolution module is configured to perform raised cosine convolution on the channel estimate without raised cosine interference obtained after removing the raised cosine interference. 10. The receiver according to claim 9, wherein the channel estimation device further comprises: The parameter storage module is used for storing the raised cosine filter parameter matrix; The restoring module is used for calculating the product of the inverse matrix of the raised cosine filter parameter matrix and the channel estimation value obtained by channel estimation, so as to eliminate the raised cosine interference for the channel estimation value obtained by channel estimation. The convolution module is configured to perform raised cosine convolution on the raised cosine interference-eliminated channel estimate and the raised cosine filter parameter matrix. 11. The receiver according to claim 9 or 10, wherein the channel estimation device further comprises: A downsampling module is used to downsample the path energy at all positions in the pilot channel obtained after the raised cosine convolution; The matrix generating module is configured to generate a channel estimation matrix according to the channel estimation value of 1/2 chip obtained by downsampling.

Images