CN103647624A - Shift-based low-complexity MIMO detection method - Google Patents

Abstract

The invention is a shift-based low-complexity MIMO detection method which is mainly used for signal detection of a multiple-input multiple-output system. The method comprises the following steps: firstly, calculating an unconstrained solution of a transmitting signal according to the current channel state when a detector receives new data; then, arbitrarily selecting a point from a star map as a reference signal point and calculating a difference vector between the unconstrained solution and the reference signal point; next, rounding up and rounding down each element in the difference vector to an integer and combining the rounded elements to obtain a plurality of candidate star points; and finally, carrying out minimum distance decision on the obtained candidate star points and outputting a decided signal vector. The invention gives full consideration to the problem that complexity and performance are difficult to balance in MIMO detection and proposes a signal detection method based on shift and rounding. The receiving performance of an ML can be achieved by carrying out minimum distance decision on a small number of candidate constellation points, and the complexity is reduced significantly.

Description

A kind of low-complexity MIMO detection method based on displacement

Technical field

The invention belongs to wireless communication technology field, relate to the input design in mimo system, mainly for the input problem of MIMO down link.

Background technology

MIMO technology refers to the technology of using a plurality of transmissions, reception antenna.MIMO technology can improve exponentially capacity and the availability of frequency spectrum of communication system in the situation that not increasing bandwidth, therefore MIMO technology is acknowledged as the key technology of Next-Generation Wireless Communication Systems, the reliability that it can effectively utilize random fading and space multi-way footpath to propagate to improve transmission rate and improve transmission.Meanwhile, MIMO has also brought array gain and has disturbed the benefits such as inhibition.In LTE system, MIMO combines with OFDM, and both are had complementary advantages, and improves the availability of frequency spectrum of system by MIMO technology, simultaneously by the frequency selective fading effect in OFDM modulation system antagonism channel.

The reception technique of mimo system is conventionally paid close attention to improving performance and is reduced detection algorithm complexity.Improving performance means and makes receiving algorithm can retain as far as possible or fully excavate the various gains that mimo system self can provide, and systematic function can be evaluated by the error sign ratio (or bit error rate) under different signal to noise ratio conditions.With regard to reducing complexity, should consider in actual applications, due to the restriction of hardware handles ability, to reduce as much as possible the complexity of algorithm, can be by hardware supports.Yet performance and complexity are difficult to take into account conventionally, when optimizing a side, tend to lose the opposing party.In general, the correlative study major part of current detection technology is to pay close attention to independently wherein some aspects, only has small part research can consider performance and complexity simultaneously.

In mimo system, current existing ZF detection algorithm, least mean-square error detection algorithm all cannot reach optimum detection performance (performance is often poor), and the optimal maximum likelihood algorithm of performance has too high complexity, along with the increase of number of transmit antennas, the complexity of detection can increase by exponentially.Ball decoding algorithm is when guaranteeing optimum (or approaching optimum) performance, can in average meaning, effectively reduce the search volume of detecting, but its complexity is unfixing, and the search volume under worst case equals the search volume that ML detects, therefore ball decoding technique there is no method and reaches hard-wired requirement at present.For complexity, do not fix, do not have the algorithm in the clear and definite instantaneous complexity upper bound, if this class algorithm is used in complexity constrained environment, may in detecting, some there is finishing in advance the situation (be subject to complexity restriction and cause) of operation, this situation is by the detection performance of loss system, and this loss conventionally cannot be quantitative.

Summary of the invention

The present invention is directed to the deficiencies in the prior art, a kind of low-complexity MIMO detection method based on displacement is provided, make to reduce the complexity of detection algorithm under the precondition of loss system optimal performance not.The technical solution used in the present invention is a kind of low-complexity MIMO signal detecting method based on displacement, and the multiple mimo system that actual 16QAM modulates, dual-mode antenna is 2 * 2 of take is example, and the method comprises the following steps:

1) this complex system is expressed as to real system of equal value.Make y _jRwith y _jIrepresent respectively to receive on j reception antenna real part and the imaginary part of signal, i.e. y _jR=Re{y _j, y _jI=Im{y _j.Similarly, the input signal on transmitting terminal i transmit antennas is expressed as: x _iR=Re{x _i, x _iI=Im{x _i.For 2 * 2MIMO channel, receive real part and imaginary part that signal can be expressed as it according to the method described above, expression is as follows:

$\left[\begin{array}{c}{y}_{1R}+{\mathrm{jy}}_{1I}\\ y_{2R}+{\mathrm{jy}}_{2I}\end{array}\right]=\left[\begin{array}{cc}{h}_{11R}+{\mathrm{jh}}_{11I}& h_{12R}+{\mathrm{jh}}_{12I}\\ h_{21R}+{\mathrm{jh}}_{21I}& h_{22R}+{\mathrm{jh}}_{22I}\end{array}\right]\left[\begin{array}{c}{x}_{1R}+{\mathrm{jx}}_{1I}\\ x_{2R}+{\mathrm{jx}}_{2I}\end{array}\right]+\left[\begin{array}{c}{z}_{1R}+{\mathrm{jz}}_{1I}\\ z_{2R}+{\mathrm{jz}}_{2I}\end{array}\right]$

H wherein _ijR=Re{h _ij, h _ijI=Im{h _ij, z _iR=Re{z _i, z _iI=Im{z _i.The real part of above formula and imaginary part can be expressed as:

$\left[\begin{array}{c}{y}_{1R}\\ y_{2R}\end{array}\right]=\left[\begin{array}{cc}{h}_{11R}& h_{12R}\\ h_{21R}& h_{22R}\end{array}\right]\left[\begin{array}{c}{x}_{1R}\\ x_{2R}\end{array}\right]-\left[\begin{array}{cc}{h}_{11I}& h_{12I}\\ h_{21I}& h_{22I}\end{array}\right]\left[\begin{array}{c}{x}_{1I}\\ x_{2I}\end{array}\right]+\left[\begin{array}{c}{z}_{1R}\\ z_{2R}\end{array}\right]$

$=\left[\begin{array}{cccc}{h}_{11R}& h_{12R}& -h_{11I}& -h_{12I}\\ h_{21R}& h_{22R}& -h_{21I}& -h_{22I}\end{array}\right]\left[\begin{array}{c}{x}_{1R}\\ x_{2R}\\ x_{1I}\\ x_{2I}\end{array}\right]+\left[\begin{array}{c}{z}_{1R}\\ z_{2R}\end{array}\right]$

$\left[\begin{array}{c}{y}_{1I}\\ y_{2I}\end{array}\right]=\left[\begin{array}{cccc}{h}_{11I}& h_{12I}& h_{11R}& h_{12R}\\ h_{21I}& h_{22I}& h_{21R}& h_{22R}\end{array}\right]\left[\begin{array}{c}{x}_{1R}\\ x_{2R}\\ x_{1I}\\ x_{2I}\end{array}\right]+\left[\begin{array}{c}{z}_{1I}\\ z_{2I}\end{array}\right]$

Above two formulas can merge into following expression:

2) when detector receives new data y, accounting equation y=Hx+n without constrained solution:

$\hat{x}=H^{-1}y$

Wherein y is for receiving data, and H is channel matrix.Claim

(referring to without constrained solution without constrained solution of equation y=Hx+n

be not defined as some in modulation constellation points).

3) given reference point: $x_{\mathrm{ref}}=\left[\begin{array}{c}1\\ 1\\ 1\\ 1\end{array}\right]$ (also can be made as other any constellation point), calculates current data without constrained solution

with given reference point x _refdifference value vector:

$d=\hat{x}-x_{\mathrm{ref}}$

4) distance of take between the adjacent constellation point of coordinate direction as unit value (if 16QAM modulation level direction point set is { 3,-1,1,3}, unit value is exactly 2 so), each element in above-mentioned difference value vector is rounded up and rounds and obtain following two vectors downwards:

Wherein

expression rounds up,

represent to round downwards.

If shift vector is $d=\left[\begin{array}{c}{d}_1\\ d_2\\ d_3\\ d_4\end{array}\right],$ Each element d in d _ibe vectorial d _ceilor vectorial d _floorthe element of corresponding line, combines and can obtain a plurality of candidate vector element.

5) by candidate vector, add respectively reference vector, the vector point after being shifted, newer point set is adjudicated to output court verdict as candidate symbol.

Beneficial effect of the present invention is mainly reflected in following two aspects:

1) propose to realize a kind of low complex degree detection method of ML performance, significantly reduced the detection algorithm complexity in high order modulation and multi-layer data stream situation, had certain practicality.

2) the method is compared with existing other low complex degree input scheme, can obtain the systematic function identical with ML algorithm, is better than other low complex degree detection algorithms in performance.

Accompanying drawing explanation

Fig. 1 is the schematic diagram (16QAM of take modulation, single layer data stream are example) being shifted according to input signal and reference signal.

When Fig. 2 has provided transmission two layer data, adopt the BER curve of displacement detection algorithm.

When Fig. 3 has provided transmission four layer data, adopt the BER curve of displacement detection algorithm.

Embodiment

Fig. 1 is the schematic diagram (16QAM of take modulation, single layer data stream are example) being shifted according to input signal and reference signal.After displacement, according to rounding and combining, determine current demand signal unconfinement solution four constellation point around.

The performance of this low-complexity MIMO detection algorithm is described by simulation result below, mainly considers the performance of BER under different signal to noise ratio conditions.Suppose that channel is flat fading.At transmitting terminal, the bit of transmission carries out 16QAM modulation without coding.Simulated conditions is as shown in table 1.

Table 1 simulation parameter table

Antenna configuration	2 * 2 or 4 * 4
		Channel	flat?fading
Modulation system	16QAM
		Every frame symbolic number	30×N t(antenna number)
Frame number under every SNR	10000
		Chnnel coding	Uncoded

When Fig. 2 has provided transmission two layer data, adopt the BER curve of displacement detection algorithm.Can see, while adopting this programme, can realize the systematic function identical with ML.

When Fig. 3 has provided transmission four layer data, adopt the BER curve of displacement detection algorithm.Can see, while adopting this programme, can realize the systematic function identical with ML.

In conjunction with detection algorithm proposed by the invention, ML algorithm and other two kinds of low complex degree ML algorithms, carry out analysis of complexity, obtain result as shown in table 2.

The comparison of table 2 analysis of complexity

Claims (2)

1. the low-complexity MIMO detection method based on displacement, is characterized in that: this detection method is carried out (16QAM of take modulation, 2 * 2MIMO system be example) according to the following steps:

1) this complex system is expressed as to real system of equal value.Make y _jRwith y _jIrepresent respectively to receive on j reception antenna real part and the imaginary part of signal, i.e. y _jR=Re{y _j, y _jI=Im{y _j.Similarly, the input signal on transmitting terminal i transmit antennas is expressed as: x _iR=Re{x _i, x _iI=Im{x _i.For 2 * 2MIMO channel, receive real part and imaginary part that signal can be expressed as it according to the method described above, expression is as follows:

$\left[\begin{array}{c}{y}_{1R}+{\mathrm{jy}}_{1I}\\ y_{2R}+{\mathrm{jy}}_{2I}\end{array}\right]=\left[\begin{array}{cc}{h}_{11R}+{\mathrm{jh}}_{11I}& h_{12R}+{\mathrm{jh}}_{12I}\\ h_{21R}+{\mathrm{jh}}_{21I}& h_{22R}+{\mathrm{jh}}_{22I}\end{array}\right]\left[\begin{array}{c}{x}_{1R}+{\mathrm{jx}}_{1I}\\ x_{2R}+{\mathrm{jx}}_{2I}\end{array}\right]+\left[\begin{array}{c}{z}_{1R}+{\mathrm{jz}}_{1I}\\ z_{2R}+{\mathrm{jz}}_{2I}\end{array}\right]$

H wherein _ijR=Re{h _ij, h _ijI=Im{h _ij, z _iR=Re{z _i, z _iI=Im{z _i.The real part of above formula and imaginary part can be expressed as:

$\left[\begin{array}{c}{y}_{1R}\\ y_{2R}\end{array}\right]=\left[\begin{array}{cc}{h}_{11R}& h_{12R}\\ h_{21R}& h_{22R}\end{array}\right]\left[\begin{array}{c}{x}_{1R}\\ x_{2R}\end{array}\right]-\left[\begin{array}{cc}{h}_{11I}& h_{12I}\\ h_{21I}& h_{22I}\end{array}\right]\left[\begin{array}{c}{x}_{1I}\\ x_{2I}\end{array}\right]+\left[\begin{array}{c}{z}_{1R}\\ z_{2R}\end{array}\right]$

$=\left[\begin{array}{cccc}{h}_{11R}& h_{12R}& -h_{11I}& -h_{12I}\\ h_{21R}& h_{22R}& -h_{21I}& -h_{22I}\end{array}\right]\left[\begin{array}{c}{x}_{1R}\\ x_{2R}\\ x_{1I}\\ x_{2I}\end{array}\right]+\left[\begin{array}{c}{z}_{1R}\\ z_{2R}\end{array}\right]$

$\left[\begin{array}{c}{y}_{1I}\\ y_{2I}\end{array}\right]=\left[\begin{array}{cccc}{h}_{11I}& h_{12I}& h_{11R}& h_{12R}\\ h_{21I}& h_{22I}& h_{21R}& h_{22R}\end{array}\right]\left[\begin{array}{c}{x}_{1R}\\ x_{2R}\\ x_{1I}\\ x_{2I}\end{array}\right]+\left[\begin{array}{c}{z}_{1I}\\ z_{2I}\end{array}\right]$

Above two formulas can merge into following expression:

2) when detector receives new data y, accounting equation y=Hx+n without constrained solution:

$\hat{x}=H^{-1}y$

Wherein y is for receiving data, and H is channel matrix.Claim (referring to without constrained solution without constrained solution of equation y=Hx+n be not defined as some in modulation constellation points).

3) given reference point: $x_{\mathrm{ref}}=\left[\begin{array}{c}1\\ 1\\ 1\\ 1\end{array}\right]$ (also can be made as other any constellation point), calculates current data without constrained solution

with given reference point x _refdifference value vector:

$d=\hat{x}-x_{\mathrm{ref}}$

4) distance of take between the adjacent constellation point of coordinate direction as unit value (if 16QAM modulation level direction point set is { 3,-1,1,3}, unit value is exactly 2 so), each element in above-mentioned difference value vector is rounded up and rounds and obtain following two vectors downwards:

Wherein

expression rounds up,

represent to round downwards.

If shift vector is $d=\left[\begin{array}{c}{d}_1\\ d_2\\ d_3\\ d_4\end{array}\right],$ Each element d in d _ibe vectorial d _ceilor vectorial d _floorthe element of corresponding line, combines and can obtain a plurality of candidate vector element.

5) by candidate vector, add respectively reference vector, the vector point after being shifted, newer point set is adjudicated to output court verdict as candidate symbol.

2. according to the low-complexity MIMO detection algorithm based on displacement described in right 1, it is characterized in that: during transmission multi-layer data stream, need judge current channel condition number, according to channel condition information, determine the quantity of transport stream, the quantity of convection current is adjusted dynamically.

Images