WO2007093111A1

WO2007093111A1 - Maximum likelihood estimation method and apparatus for a mimo system

Info

Publication number: WO2007093111A1
Application number: PCT/CN2007/000216
Authority: WO
Inventors: Bin Li
Original assignee: Huawei Technologies Co., Ltd.
Priority date: 2006-02-17
Filing date: 2007-01-19
Publication date: 2007-08-23
Also published as: CN101026428A

Abstract

A maximum likelihood estimation method for a MIMO system includes the steps of: estimating each received signal, and generating J maximum likelihood estimated values, wherein J is 1≤ J ≤S, S is the number of the received signals under all possible situations; generating the sequence of all possible received signals based on the J maximum likelihood estimated values of each received signal; calculating the Euclidian Distance corresponding to each possible received signal sequence, and finding out the minimum Euclidian distance; outputting the received signal sequence corresponding to the minimum Euclidian distance. A maximum likelihood estimation apparatus and receiver for a MIMO system are also provided. By applying this method, the apparatus and the MIMO system, the search space of the received signal sequence is furthest reduced and the calculating quantity is observably reduced, and the applicability is furthest enhanced, the signal detecting capability of the maximum estimation method of this invention approximates to the optimal one.

Description

Maximum likelihood estimation method and device for multiple input multiple output system

Technical field

The present invention belongs to the field of wireless communication technologies, and in particular, to a maximum likelihood estimation method and apparatus for a multiple input multiple output (MIMO) system. Background of the invention

MIMO technology can increase the capacity and spectrum utilization of communication systems without increasing bandwidth. The MIMO technology uses multiple antennas to transmit and receive signals at the transmitting end and the receiving end, respectively. Since the signals simultaneously transmitted by the respective transmitting antennas occupy the same frequency band, the communication bandwidth does not increase. There is a spatial channel between each transmitting antenna and each receiving antenna. If the channel impulse response of each spatial channel is unique, the MIMO system creates multiple connections between the transmitting end and the receiving end through multiple transmitting antennas and multiple receiving antennas. Parallel independent spatial channels. The information is transmitted independently through the spatial channel to improve the transmission data rate of the MIMO system. MIMO systems have received extensive attention and application in wireless communication systems due to their multiplying system capacity over single transmit antennas and single receive antenna systems.

In a MIMO system, on the receiver side, the signal received on the antenna will be a superposition and mutual interference of multiple transmit antenna signals. It is important to recover the original received signal from the received signal. Currently commonly used signal detection methods are the minimum mean square error (MMSE) method and the maximum likelihood estimation (ML) method. The MMSE method is easy to implement but the signal detection performance is poor. The performance of the ML method is good, but the complexity is extremely high and it is not easy to implement.

For example: In a MIMO system, there are M transmit antennas and N receive antennas. In the ^: symbol, the transmission sequence on the M antennas is ^ ^ , ^ ,···, ( ^Μ ))' , here ( ) ' indicates vector transpose. Assume that the received signal on one of the receiving antennas is =

, Then there are:

^ = ^H " ^d k + ^f formula (work) In equation (1), is a complex Μ Ν Ν channel matrix; ⁿ k = ( (1), ( ² ) '···, ' ( )' is _Ν a noise vector at the receiving antennas, each vector element is the variance ^σ "

The following describes the existing MMSE signal detection method and ML signal detection method separately:

In the existing MMSE signal detection method, the original signal recovered by the receiving antenna is

K = [H^H _k + σ YH r _k Formula ( 2 ) where " is the unit matrix; represents the transposed conjugate of the matrix, () ^_1 represents the inversion of the matrix. It can be seen from equation (2), MMSE signal detection method It is relatively easy to implement, but since the method does not eliminate interference between signals in signal detection, its detection performance is not good.

The ML signal detection method can significantly improve the signal detection performance. Figure 1 is a schematic flow chart of a conventional ML signal detection method. As shown in Figure 1, the following steps are included:

Step 101: generating a received signal sequence may be a total number of S ^M, assuming that the size of the constellation of the modulated signal S, M is the number of multi-input multi-output system of transmitting antennas;

Step 102: Calculate the Euclidean distance corresponding to each received signal sequence, and the Euclidean distance has a total of S ^M ;

Step 103: the number of all the Euclidean distance S ^M, find the smallest Euclidean distance; Step 104: Output the received signal sequence corresponding to the minimum Euclidean distance.

It can be seen that the constellation size of the modulated signal is S, and the set of constellation points is Ω = {Ω„Ω ₂ ,..., Ω ₅ } ₅ then the signal on each transmit antenna belongs to the set, ie ^d » ⁿ , l ≤ m ≤ M. So the signal sequence sent by the transmit antenna belongs to the set Ω is ^£ " The size of the collection is G = . In the ML method, using a maximum likelihood criterion, all of the signals in the search to find the best of a signal sequence, that is the minimum Euclidean distance, the mathematical ML signal detection method is expressed as: two \ r _k

Formula (3) Here HI ^χ ιΓ "Η^ ² +·'·+|1⁄2Γ ^{Χ =} { ^Χ Ι> ^χ 2'-' ^χ Ν))'

It can be seen from the flow and formula (3) that the ML signal detection method is high in complexity and is not easy to implement. For example, for an antenna system with NxM = ⁴ >< ⁴ , when using 16 Quadrature Amplitude Modulation (QAM) modulation, the search space size is: G = S ^M =16 ⁴ =65,536, which also means Repeat the calculation of Ik - ^Al ² value G = 6 ⁵ , ⁵³⁶ times, and then find the smallest one in it, the amount of calculation is obviously already large.

For example, when using the 64QAM modulation method, the search space size is: (7 = ^ = ⁶⁴⁴ =1 ⁶ , ⁷⁷⁷ , ² 1 ⁶ , which means that the Ik-Alf value must be calculated repeatedly.

^{G = 16} , ⁷⁷⁷ , ²¹⁶ times, and then find the smallest one among them, obviously, the amount of calculation is very large.

It can be seen that the existing ML valuation method is computationally intensive, which causes great defects in its applicability. Summary of the invention

In view of this, the main objective of the embodiments of the present invention is to provide a maximum likelihood estimation method and apparatus for a MIMO system, to overcome the high complexity and poor performance techniques in the prior art. Question.

Another object of embodiments of the present invention is to propose a receiver for a MIMO system to reduce the complexity of its maximum likelihood estimate.

To achieve the above objective, the technical solution of the embodiment of the present invention is implemented as follows: A maximum likelihood estimation method for a MIMO system, the method comprising the following steps: estimating each received signal to generate a maximum Likelihood estimate, where ■ is in the range of 1 ≤ ≤ S, where S is the number of all possible values of the received signal;

Generating all possible received signal sequences based on the J maximum likelihood estimates for each received signal;

Calculating the Euclidean distance corresponding to each possible received signal sequence and finding the smallest Euclidean distance from it;

A sequence of received signals corresponding to the minimum Euclidean distance is output.

The method for estimating the received signal is: performing a zero-break (ZF) equalization estimate or a minimum mean square error (MMSE) equalization estimate for each received signal.

\≤J≤-S

The value of the range is 3 .

l≤J≤-S

The value of ■ is in the range of 2.

Based on a maximum likelihood estimate for each received signal, all possible J ^M received signal sequences are generated, where M is the number of transmit antennas in the multiple input multiple output system.

A maximum likelihood estimation apparatus for a MIMO system, the apparatus comprising: a received signal estimate unit for estimating an estimated value of each received signal, generating a maximum likelihood estimate, wherein J is taken The value range is 1 ≤ J ≤ S, where S is the number of all possible values of the received signal;

a received signal sequence generating unit for estimating J maximum likelihoods for each received signal a value that produces all possible received signal sequences;

An Euclidean distance calculation unit, configured to calculate an Euclidean distance corresponding to each possible received signal sequence, and find a minimum Euclidean distance therefrom;

The received signal sequence output unit is configured to output a sequence of received signals corresponding to the minimum Euclidean distance.

The received signal estimated value unit is a ZF equalization estimated value unit or an MMSE balanced estimated value unit. The range of ■ is i ^{≤ J ≤} S.

l≤J < -5'

The value ranges from 2 to 2.

A receiver for a multiple input multiple output system, the receiver comprising a plurality of receive antennas, a demodulator and a maximum likelihood estimation device, wherein:

a plurality of receiving antennas for receiving multiple wireless signals;

a demodulator, configured to demodulate the received multi-channel wireless signal, and send the demodulated signal to a maximum likelihood estimation device;

The maximum likelihood evaluation device includes:

The received signal estimation value unit is configured to estimate an estimated value of each received signal sent by the demodulator, and generate J maximum likelihood estimates, wherein the value range of J is 1≤J≤S, S is The number of all possible values of the received signal;

a received signal sequence generating unit configured to generate all possible received signal sequences according to J maximum likelihood estimates of each received signal;

The received signal sequence output unit is configured to output a received signal sequence corresponding to the minimum Euclidean distance. It can be seen from the above technical solution that, compared with the existing ML signal detecting method, in the present invention, an estimated value is first obtained for each received signal, and a maximum likelihood estimate is generated, wherein the value range of J is obtained. is 1≤J≤S, S is the number of all possible values of the received signal, rather than the prior art simple possibility may generate ^M received signal sequence, the present invention significantly reduces all received symbols S The search space of the received signal sequence is reduced exponentially from the original to J ^M , which greatly reduces the search space of the received signal sequence, so the signal detection method of the present invention significantly reduces the amount of calculation, so that it is executed in practice. It is possible that the applicability is improved and the detection performance is close to the optimal ML signal detection method.

At the same time, in the present invention, it is possible to perform various types of estimation values for each received signal, and there are many alternative estimation methods, such as performing ZF equalization estimation values or MMSE equalization estimation values, etc., The invention is also very convenient to implement. BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic flow chart of an ML method in the prior art.

2 is a schematic flow chart of an ML method for a MIMO system according to an embodiment of the present invention.

3 is a comparative simulation diagram of an existing MMSE method and ML method according to an embodiment of the present invention.

4 is a schematic diagram showing an exemplary structure of an ML apparatus for a MIMO system according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION The present invention will be further described in detail with reference to the drawings and specific embodiments.

2 is a maximum likelihood estimate (ML) for a MIMO system in accordance with an embodiment of the present invention. A schematic flow chart of the method. As shown in FIG. 2, the method includes the following steps: Step 201: Perform an estimated value for each received signal, and generate a maximum likelihood estimate, wherein the value range of · is 1 ≤ J ≤ S, and S is a receiving The number of all possible values of the signal.

Here, there are various ways to estimate the value, and the estimated value of the received signal can be estimated in various ways, which is not limited in this embodiment. For example, the ZF equalization estimation method or the MMSE equalization estimation method may be used. .

Specifically, when the ZF equalization estimation method is adopted, the zero-breaking principle is used to obtain: k ={H" _k ) ^X H _k ^H r _k , where Λ is a column vector, Λ=(Λ(1),Λ( 2), "·, 3⁄4(Μ))'_; for each element, \<1<M , calculated with all

_{_{_{Ω = {Ω 15 Ω 2,}}} · "Ω 5 Euclidean distance between constellation points: ,,,.}" = | ( /) _ Ω ", |, i ≤m≤S ^ ^ where m is the first constellation Point. Find the minimum Euclidean distance from them. Their serial numbers are:

^"' ,..., ^ is the sequence number of a constellation point, and their corresponding constellation points are:

When using the MMSE equilibrium estimation method, specifically, the formula can be used.

To achieve an estimated value for each received signal, and generate J maximum likelihood estimates, where J has a value range of 1 ≤ J ≤ S, where S is the number of all possible values of the received signal.

Although the two estimation methods of the ZF equalization estimation value and the MMSE equalization estimation value are described in detail above, those skilled in the art can appreciate that the method for estimating the value of the present invention is not limited thereto, but can be applied. Any form of predictive value.

Step 202: Generate all possible received signal sequences based on a maximum likelihood estimate for each received signal. Here, a new signal subspace consisting of ^Ω TM ( l ≤ m ≤ M ) is formed.

Each element in ^(1)e , ( ² ) ^e , ...,

As can be seen from the above equation, the search and computation space has been reduced to ^(? '^ ⁼ " ⁷ "', where ≤ S. When J is much smaller than S, the new search and computation space will be much smaller than the original search and calculation. Space S ^M. Here, different J values can be selected according to different requirements for receiving performance. For example: When the receiving performance is high and the computational complexity is not strict, a larger value can be selected. When the degree is not high and the receiving performance is low, you can choose

l<J<-5

Small · value. Preferably, the range of values is 3; more preferably, the value of the range of values is x ≤ J < -5

Surrounded by 2 .

Step 203: Calculate the Euclidean distance corresponding to each possible received signal sequence, and find the smallest Euclidean distance from it.

Here, search for all the signals in ", find the best one of the signal sequences, that is, find the smallest Euclidean distance in the Euclidean distance corresponding to each possible received signal sequence, then receive the signal sequence. The mathematics is expressed as: , , , ,

Formula (4) Step 204: Output a received signal sequence corresponding to the minimum Euclidean distance.

Heretofore, the basic method of the embodiment of the present invention has been described in detail, and a comparison of the embodiment of the present invention with the existing MMSE method and ML method will be described below.

Fig. 3 is a comparative simulation diagram of a conventional MMSE method and an ML method according to an embodiment of the present invention.

In FIG. 3, assuming a MIMO system having two transmit antennas and two receive antennas, And the modulation method is 64QAM. FIG. 3 depicts a situation in which the embodiments of the present invention correspond to J=5, 7, and 10, respectively. It can be seen from the simulation comparison of Fig. 3 that the performance of the embodiment of the present invention is far superior to the MMSE method and is close to the optimal ML method.

As can be seen from Fig. 3, the embodiment of the present invention has a gain of about 5.0 dB - 7.0 dB in the case of a symbol error rate of 0.01 as compared with the conventional MMSE method, and therefore the performance of the embodiment of the present invention is good.

As can be seen from FIG. 3, when the J is =5, 7, and 10, respectively, the complexity of the embodiment of the present invention is 80 times, 50 times, and 30 times lower than the complexity of the ML, respectively, as compared with the conventional ML method. Embodiments of the present invention greatly reduce complexity.

At the same time, compared with the traditional ML method, the gain variation of the embodiment of the present invention is not large, so the performance is very close to the existing ML method.

An ML device for a MIMO system can also be proposed based on an embodiment of the present invention. 4 is a schematic diagram showing an exemplary structure of an ML apparatus for a MIMO system according to an embodiment of the present invention. As shown in FIG. 4, the ML device 400 includes:

The received signal estimation value unit 401 is configured to perform an estimated value for each received signal to generate a maximum likelihood estimate, where the value range of J is 1 ≤ ^ ≤ S, where S is all possible values of the received signal. Number

Received signal sequence generating unit 402 is configured to generate all possible received signal sequences according to J maximum likelihood estimates of each received signal;

The Euclidean distance calculation unit 403 is configured to calculate a Euclidean distance corresponding to each possible received signal sequence and find a minimum Euclidean distance therefrom;

The received signal sequence output unit 404 is configured to output a received signal sequence corresponding to the minimum Euclidean distance.

The received signal estimation value unit 401 may be a ZF equalization estimation unit that performs a ZF equalization estimation algorithm, or an MMSE equalization algorithm that performs an MMSE equalization estimation algorithm. Estimated value unit.

Similarly, when the reception performance is high and the computational complexity is not critical, a larger value can be selected; when the computational complexity is required to be low and the reception performance is low, a smaller value can be selected. Preferably, the value ranges from ^ ^ . l≤J < -5

More preferably, the value may be in the range of ² .

The ML apparatus of the embodiment of the present invention can be applied to a receiver of a MIMO system. Obviously, this will greatly improve the performance of the MIMO receiver.

For example, in one embodiment of the present invention, a receiver for a multiple input multiple output system is provided, the receiver including a plurality of receive antennas, a demodulator, and a maximum likelihood estimation device, wherein: a receiving antenna for receiving a plurality of wireless signals; a demodulator for demodulating the received multiple wireless signals, and transmitting the demodulated signals to a maximum likelihood estimating device; Value devices include:

The above is only the preferred embodiment of the present invention and is not intended to limit the scope of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and scope of the present invention are intended to be included within the scope of the present invention.

Claims

Claim

A maximum likelihood estimation method for a multiple input multiple output system, characterized in that: estimating a value for each received signal, and generating J maximum likelihood estimates, wherein the range of J is 1 ≤ J ≤ S , S is the number of all possible values of the received signal;

Calculate the Euclidean distance corresponding to each possible received signal sequence, and find the smallest Euclidean distance from it;

The method according to claim 1, wherein the method for estimating the value is: performing a zero-break ZF equalization estimation value or a minimum mean square error MMSE equalization estimation value for each received signal.

The method according to claim 1 or 2, wherein the value of J is in the range of i ≤ J ≤ > s.

3

The method according to claim 3, wherein the value of J is in the range of i ≤ J ≤ s.

2

5. The method of claim 1 or claim 2, characterized in that the number of all possible received signal sequence is a J ^M, where M is the number of multi-input multi-output system of transmitting antennas.

6. A maximum likelihood estimating apparatus for a multiple input multiple output system, the apparatus comprising:

The received signal estimation value unit is configured to perform an estimated value for each received signal, and generate J maximum likelihood estimates, where J has a value range of 1 ≤ J ≤ S, and S is all possible values of the received signal. Number a received signal sequence generating unit, configured to generate all possible received signal sequences according to J maximum likelihood estimates of each received signal;

The maximum likelihood estimating apparatus according to claim 6, wherein the received signal estimated value unit is a ZF equalized estimated value unit that uses a ZF equalization method to perform an estimated value, or uses an MMSE equalization pre-stage. Valuation method The MMSE equilibrium estimate unit for the estimated value.

The maximum likelihood estimating apparatus according to claim 6 or 7, wherein the value of J is in a range of 1 ≤ J ≤ ^S.

3

9. The maximum likelihood estimating apparatus according to claim 8, wherein the value of J is in a range of l ≤ J ≤ i > S.

2

10. A receiver for a multiple input multiple output system, comprising: a plurality of receiving antennas, a demodulator, and a plurality of receiving antennas for receiving multiple wireless signals;

Characterizing in that the receiver further comprises a maximum likelihood estimation device;

The maximum likelihood evaluation device includes:

a received signal sequence generating unit for estimating J maximum likelihoods for each received signal a value that produces all possible sequences of received words;