CN102833018B - SEESM (simplified exponential effective signal to interference and noise ratio mapping) implementation method based on order self-adaption mu law non-uniform quantizing - Google Patents

Abstract

The invention provides an SEESM (simplified exponential effective signal to interference and noise ratio mapping) implementation method based on order self-adaption mu law non-uniform quantizing. The method includes the following steps: firstly, determining maximum values and minimum values of sub-carriers SINR (signal to interference and noise ratio) in a system bandwidth, and adaptively acquiring a quantization order M applicable to a current channel condition based on the above maximum values and the minimum values; secondly, dividing subcarrier SINR range into M quantized intervals according to an M-fold mu law, and mapping subcarrier SINRs into corresponding quantized intervals; thirdly, uniformly quantizing the SINRs in a same quantized interval into an arithmetic average value of the SINRs in the quantized intervals; finally, using EESM(exponential effective signal to interference and noise ratio mapping)algorithm to acquire an equivalent SINR value according to the quantized value of the subcarrier SINRs. Index calculation times in the EESM algorithm can be reduced from subcarrier amount to quantization order, and on the premise of hardly losing system throughput rate performance, complexity of EESM realization is greatly lowered and timeliness of a system is improved.

Description

SEESM implementation method based on exponent number self adaptation μ rule non-uniform quantizing

Technical field

The present invention relates to wireless communication field, the implementation method that specifically a kind of effective Signal to Interference plus Noise Ratio of index of the simplification based on exponent number self adaptation μ rule non-uniform quantizing shines upon (Simplified Exponential Effective Signal to Interference and Noise Ratio Mapping, SEESM).

Background technology

Two-forty, large bandwidth are just becoming the core feature of future broadband wireless communication systems, OFDM (Orthogonal Frequency Division Multiplex, OFDM) with the performance of its good spectrum efficiency and antagonism intersymbol interference (Iter Symbol Interference, ISI), become the support technology in the 4th generation wireless communication standard.

For single-carrier system, can be easy to obtain Signal to Interference plus Noise Ratio (the Signal to Interference and Noise Ratio on every sub-channels, SINR), carry out accordingly adaptive coding and modulating (Adaptive Modulation and Coding, link circuit self-adapting (Link Adaptive, the LA) technology such as AMC).Yet for ofdm system, time delay expansion makes broadband wireless channel experience frequency selective fading, the SINR of OFDM different sub carrier is different, for the channel characteristics of a plurality of subcarriers on a frequency domain resource piece is integrated and obtains an equivalent SINR value, the effective Signal to Interference plus Noise Ratio mapping of a kind of index (Exponential Effective Signal to Interference and Noise Ratio Mapping, EESM) method has been carried out, its utilization index information measure function, by the subcarrier SINR sequence compaction of variation, it is an equivalent SINR value, thereby frequency-selective channel can be equivalent to flat fading channel and carry out various technical research.

The method that EESM calculates equivalent SINR value is as follows:

${\mathrm{\γ}}_{\mathrm{eff}}=-\mathrm{\β}\ln [\frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}\exp (-\frac{{\mathrm{\γ}}_i}{\mathrm{\β}})]---\left(1\right)$

In formula (1), N is ofdm system number of sub carrier wave, and i is subcarrier number, γ _ibe the SINR value of i subcarrier, β is the scale factor relevant with modulation coding mode (Modulation and Coding Scheme, MCS).

Although EESM has solved the mapping problems of frequency-selective channel to flat fading channel well, for LA technology provides application prerequisite, still, with reference to formula (1), we find that EESM need to carry out the exponent arithmetic of sub-carrier number contents, with third generation partner program (The3 ^rdgeneration Partnership Project, Long Term Evolution (Long Term Evolution 3GPP) formulating, LTE) R9 consensus standard is example, in standard, stipulated the situation of 20MHz bandwidth, if OFDM subcarrier spacing is 15KHz, now the number of subcarrier is 1200, applies EESM method and need to calculate index 1200 times, huge like this complexity, (especially the hardware of mobile terminal is realized) brought huge challenge to realize to the hardware of system.

Summary of the invention

The object of the invention is in order to overcome the defect of the high complexity of above-mentioned EESM method, a kind of simplification EESM method based on exponent number self adaptation μ rule non-uniform quantizing is provided.First the present invention determines maximum and the minimum value of each subcarrier SINR in system bandwidth, on this basis, obtains adaptively the quantification exponent number M that is applicable to current channel condition; Then use M folding μ rule subcarrier SINR scope is divided into M quantized interval, each subcarrier SINR is shone upon in corresponding quantized interval; To the SINR in same quantized interval, it by its unified quantization, is the arithmetic mean value of each SINR in this interval; Finally the quantized value for subcarrier SINR is used EESM algorithm, obtains equivalent SINR value.The present invention can be reduced to quantification exponent number from number of sub carrier wave by EESM algorithm Exponential calculation times, under the prerequisite of loss system throughput performance hardly, has greatly reduced the implementation complexity of EESM, has improved the real-time of system.

Particularly, the present invention is achieved by the following technical solutions, the present invention includes following steps.

Step 1, search subcarrier SINR sequence, maximum and the minimum value of acquisition subcarrier SINR;

Step 2, according to the maximum of subcarrier SINR and minimum value, determines adaptively and quantizes exponent number M;

Step 3, according to M folding μ rule, determine the separation of quantized interval;

Step 4, according to the separation of quantized interval, implements to quantize to each subcarrier SINR;

Step 5, with the subcarrier SINR value quantizing, is used EESM algorithm, obtains equivalent SINR.

Preferably, in step 2, described definite step that quantizes exponent number adaptively, adaptive approach is as follows:

M＝max{ρln[(exp(Γ _max)-exp(Γ _min))] ,1}

Wherein, Γ _max, Γ _minmaximum and the minimum value of the subcarrier SINR of difference Shi YidBWei unit.

Preferably, in step 3, described according to M folding μ rule determine the separation of quantized interval, concrete grammar is as follows:

$\left\{\begin{array}{c}{\mathrm{\λ}}_1={\mathrm{\γ}}_{\min },{\mathrm{\λ}}_{M+1}={\mathrm{\γ}}_{\max };\\ {\mathrm{\λ}}_I=(1-\mathrm{\κ}){\mathrm{\λ}}_1+\mathrm{\κ}{\mathrm{\λ}}_{i+1},i=M,M-1,\·\·\·\mathrm{3,2}\end{array}\right.$

Wherein, λ _iseparation, γ _max, γ _minbe respectively maximum and the minimum value of subcarrier SINR, κ is the scale factor of nonuniform quantiza, during iteration, according to i order from big to small, carries out.

Preferably, in step 4, described quantification, first need to be mapped into corresponding quantized interval by subcarrier SINR according to following rule:

For i subcarrier SINR γ _i, when it meets:

γ _i∈(λ _p,λ _p+1],p＝2,3,...Morγ _i∈[λ _p,λ _p+1],p＝1

Its mapping is entered to p quantized interval; Wherein, M is for quantizing exponent number, λ _pfor quantized interval separation.

Preferably, in step 4, described quantification, to the SINR in same quantized interval, usings their arithmetic mean value as the quantized value of this quantized interval.

Preferably, in step 5, described to quantizing subcarrier SINR application EESM algorithm, the final form that obtains equivalent SINR is:

${\mathrm{\γ}}_{\mathrm{eff}}=-\mathrm{\β}\ln \left(\frac{1}{N}\underset{p=1}{\overset{M}{\mathrm{\Σ}}}{M}_p\exp (-\frac{{\stackrel{\‾}{\mathrm{\γ}}}_p}{\mathrm{\β}})\right)$

Wherein, be the quantized value of p quantized interval, M _pbe the number of SINR value in p quantized interval, M is for quantizing exponent number, and N is total number of sub-carriers order, and β is the scale factor relevant to MCS.

Preferably, described segmentation factor ρ=0.25.

Preferably, scale factor κ=0.5 of described nonuniform quantiza.

EESM compares with prior art, the beneficial effect that SEESM of the present invention has is: under the prerequisite of loss system throughput performance hardly, the method can select to quantize exponent number adaptively according to channel conditions, by by the exact value of SINR, the quantized value with SINR replaces, make EESM algorithm Exponential calculation times be reduced to quantification exponent number from number of sub carrier wave, thereby greatly reduce system hardware implementation complexity, improved the real-time of system.

Accompanying drawing explanation

Fig. 1 is overall flow figure of the present invention;

Fig. 2 is in embodiment scene, under different channels bandwidth, SEESM algorithm of the present invention to EESM algorithm in prediction the approximate error aspect equivalent SINR;

Fig. 3 is in embodiment scene, under different channels bandwidth, and the contrast of the complexity of SEESM algorithm of the present invention and EESM algorithm;

Fig. 4 is in embodiment scene, and bandwidth is 1.4MHz, when time delay expands to 480ns, and SEESM algorithm of the present invention and the comparison of EESM algorithm in system throughput performance.

Embodiment

Below in conjunction with accompanying drawing, embodiments of the invention are elaborated, the present embodiment be take method of the present invention and is implemented as prerequisite, provided detailed execution mode and concrete operating process, but protection scope of the present invention is not limited to following embodiment.

In the present embodiment, inventor chooses the descending point-to-point communication link of LTE R9 agreement as application scenarios, subscriber equipment (User Equipment, UE) receive from base station (Base Station, BS) information, UE estimates channel condition information (Channel State Information, CSI), and being fed back to BS, BS is used LA technology accordingly.

The design parameter of this embodiment application scenarios is as shown in table 1.WINNER II channel model about occurring in table 1, puts forward in the WINNER project of Ta Shi European Union.WINNER II channel model utilizes the ray addition method, by shedding scattering object group according to statistical law around at BS and UE, simulates actual electromagnetic wave reflection, refraction etc., thereby realizes the simulation to actual channel.WINNER II channel model is measured by actual channel, has provided the statistical parameter of 15 kinds of scenes, and table 2 has provided the time delay spreading parameter of 4 kinds of scenes used in our emulation.

Table 1

Parameter name	Parameter value
		Link operational mode	Single-shot list is received pattern (SISO)
System bandwidth	[1.4、3、5、10、15、20]MHz

Subcarrier spacing	15KHz
		Downlink sub-carrier number	[721803006009001200]
Simulated channel model	WINNER II Model[refers to table 2]
		Channel estimation methods	Perfect channel estimating
UE translational speed	0m/s
		HARQ maximum retransmission	0

Table 2

Sequence number	Scene code name	RMS time delay expansion (ns)	Coherence bandwidth [MHz]
				1	A1	25	8.0
2	B1	76	2.6
				3	B2	480	0.4
4	C3	630	0.3

Particularly, the process of whole embodiment is: first, UE obtains the SINR value of each subcarrier by channel estimation technique; Then, UE is used the SEESM algorithm of the present invention's proposition by all equivalent SINR of subcarrier SINR boil down to _eff; Finally, UE is according to SINR _effdetermine channel quality indication (Channel Quality Indicator, CQI), fed back to BS as the foundation of AMC.

More specifically, the concrete steps of the present embodiment are as follows:

Step 1, UE obtains the SINR value of each subcarrier by channel estimation technique, and this step is for carrying before application the present invention, but non-the present invention's main contents, therefore it will not go into details.

Step 2, as shown in flow chart element in Fig. 11, search subcarrier SINR sequence { γ _i, (i=1,2...N), wherein γ _ibe the SINR of i subcarrier, N is total number of sub-carriers order, relevant to system bandwidth and subcarrier spacing, as shown in table 1.Obtain thus the maximum γ of subcarrier SINR _maxwith minimum value γ _min.

Described subcarrier SINR sequence refers to, in whole broad band multicarrier channel, by channel estimating, obtains the SINR on each subcarrier, and all subcarrier SINR arrange the sequence forming according to the frequency band sequence of positions at its place.

Step 3, as shown in flow chart element in Fig. 12, according to the fading range of the determined subcarrier SINR of step 2, determine to quantize exponent number M adaptively, and the method for determining is as formula (2):

M＝max{ρln[(exp(Γ _max)-exp(Γ _min))],1} （2）

In formula (2), ρ is the segmentation factor, and empirical value is ρ=0.25, Γ _max, Γ _minmaximum and the minimum value of the subcarrier SINR of difference Shi YidBWei unit, i.e. Γ _max=10lg (γ _max), Γ _min=10lg (γ _min), γ wherein _maxand γ _minmaximum and the minimum value of the determined subcarrier SINR of step 2.

Described definite exponent number that quantizes adaptively refers to, use the resulting quantification exponent number of formula (2) M, can be according to channel width, the expansion of channel RMS time delay, channel signal to noise ratio (Signal to Noise Ratio, SNR) automatically adjust, with minimum quantification exponent number, reach best the approaching of EESM algorithm performance.

Step 4, as shown in flow chart element in Fig. 13, use M folding μ rule, determine the separation of quantized interval.Because whole interval is quantified as M rank, when left and right end points is included, total M+1 separation, according to order from small to large, is designated as respectively λ by them _i(i=1,2 ... M+1).Definite method of this M+1 separation is as formula (3):

$\left\{\begin{array}{c}{\mathrm{\λ}}_1={\mathrm{\γ}}_{\min },{\mathrm{\λ}}_{M+1}={\mathrm{\γ}}_{\max };\\ {\mathrm{\λ}}_I=(1-\mathrm{\κ}){\mathrm{\λ}}_1+\mathrm{\κ}{\mathrm{\λ}}_{i+1},i=M,M-1,\·\·\·\mathrm{3,2}\end{array}\right.---\left(3\right)$

In formula (3), γ _max, γ _minbe respectively maximum and the minimum value of determined subcarrier SINR in step 2, κ is the scale factor of nonuniform quantiza, and empirical value is κ=0.5, and M is the determined quantification exponent number of step 3.Formula (3) is an iterative computation formula, is determined the initial value of iteration by the maximum declining and minimum value, and then, according to order from big to small, iteration is obtained each separation successively.

Step 5, as shown in flow chart element in Fig. 14, completes the quantification to subcarrier SINR.First each subcarrier SINR is mapped into corresponding quantized interval, then the subcarrier SINR in same quantized interval is asked for to arithmetic average, using the quantized value of this mean value as all SINR in this interval.

Described subcarrier SINR is mapped into corresponding quantized interval, and mapping ruler is wherein: for i subcarrier SINR γ _i, when it meets formula (4), its mapping is entered to p quantized interval.

γ _i∈(λ _p,λ _p+1],p＝2,3,...Morγ _i∈[λ _p,λ _p+1],p＝1 （4）

In formula (4), M is for quantizing exponent number, λ _pfor the determined quantized interval separation of step 4.

The described subcarrier SINR in same quantized interval asks for arithmetic average, refers to:

Suppose that p the SINR set of values in quantized interval is combined into { γ _p,j, j=1,2...M _p, M wherein _pbe SINR value number, the arithmetic mean of p quantized interval in p quantized interval calculate with the following method:

${\stackrel{\‾}{\mathrm{\γ}}}_p=\frac{1}{M_p}\underset{j=1}{\overset{M_p}{\mathrm{\Σ}}}{\mathrm{\γ}}_{p,j}---\left(5\right)$

Step 6, as shown in flow chart element in Fig. 15, implements EESM algorithm to the quantized value of subcarrier SINR, obtains equivalent SINR.Concrete grammar is as formula (6):

${\mathrm{\γ}}_{\mathrm{eff}}=-\mathrm{\β}\ln \left(\frac{1}{N}\underset{p=1}{\overset{M}{\mathrm{\Σ}}}{M}_p\exp (-\frac{{\stackrel{\‾}{\mathrm{\γ}}}_p}{\mathrm{\β}})\right)---\left(6\right)$

In formula (6), γ _efffor equivalent SINR value, for the quantized value of determined p the quantized interval of step 5, M _pbe SINR value number in p quantized interval, M is the determined quantification exponent number of step 3, and N is total number of sub-carriers order, and β is the scale factor relevant to MCS.

Due to subcarrier SINR is quantized, the SINR value in same quantized interval all uses same quantized value to represent,, for the SINR value application EESM algorithm in p quantized interval, exponent arithmetic number of times is by M _pinferior being reduced to 1 time; For the SINR value application EESM algorithm in all quantized intervals, exponent arithmetic number of times is reduced to M time by N time, and N is total number of sub-carriers order, and M is for quantizing exponent number, and this just greatly reduces computational complexity, hardware is realized more easy.

Step 7, according to resulting equivalent SINR, consults the mapping curve of SINR and CQI under additive white Gaussian noise (Additive White Gauss Noise, AWGN) channel, obtains CQI and is fed back to BS, as the use of AMC.The non-the present invention's of this step main contents, it will not go into details herein.

Fig. 2 has provided under the present embodiment scene, under 6 kinds of bandwidth of LTE R9 agreement regulation, applies the determined equivalent SINR of SEESM algorithm of the present invention, with the approximate error of the determined equivalent SINR of application EESM algorithm.Can find out, the approximation ratio of the present invention to EESM, reduces along with the increase of system bandwidth, along with the increase of Time Delay of Systems expansion, reduces.Even if but under worst condition (maximum bandwidth, maximum delay expansion), approximate error can not surpass 2dB, can complete preferably the prediction to equivalent SINR.

Fig. 3 has provided under the present embodiment scene, under 6 kinds of bandwidth of LTE R9 agreement regulation, and SEESM algorithm of the present invention and EESM algorithm, the comparison in index calculation times.Can find out, the outstanding advantage of the present invention aspect reduction computation complexity, and this advantage is more obvious along with the increase of system bandwidth.

Fig. 4 has provided under the present embodiment scene, in bandwidth, is 1.4MHz, and time delay expands in the situation of 480ns, and the SEESM algorithm that system is used the present invention to propose, with use EESM algorithm, the comparison in system throughput performance.Can find out, SEESM has approached well EESM algorithm aspect throughput.

Claims (8)

1. based on exponent number self adaptation μ, restrain a SEESM implementation method for non-uniform quantizing, wherein, SEESM refers to the effective Signal to Interference plus Noise Ratio mapping of the index of simplification, it is characterized in that, comprises the following steps:

Step 1, search subcarrier Signal to Interference plus Noise Ratio SINR sequence, maximum and the minimum value of acquisition subcarrier SINR;

Step 2, according to the maximum of subcarrier SINR and minimum value, determines adaptively and quantizes exponent number M;

Step 3, according to M folding μ rule, determine the separation of quantized interval;

Step 4, according to the separation of quantized interval, implements to quantize to each subcarrier SINR;

Step 5, with the subcarrier SINR value quantizing, is used the effective Signal to Interference plus Noise Ratio mapping of index EESM algorithm, obtains equivalent SINR.

2. the SEESM implementation method based on exponent number self adaptation μ rule non-uniform quantizing according to claim 1, is characterized in that, in step 2, and described definite step that quantizes exponent number adaptively, adaptive approach is as follows:

M＝max{ρln[(exp(Γ _max)-exp(Γ _min))],1}

Wherein, Γ _max, Γ _minmaximum and the minimum value of the subcarrier SINR of difference Shi YidBWei unit, ρ is the segmentation factor.

3. the SEESM implementation method based on exponent number self adaptation μ rule non-uniform quantizing according to claim 1, is characterized in that, in step 3, described according to M folding μ rule determine the separation of quantized interval, concrete grammar is as follows:

$\left\{\begin{array}{c}{\mathrm{\λ}}_1={\mathrm{\γ}}_{\min },{\mathrm{\λ}}_{M+1}={\mathrm{\γ}}_{\max };\\ {\mathrm{\λ}}_I=(1-\mathrm{\κ}){\mathrm{\λ}}_1+\mathrm{\κ}{\mathrm{\λ}}_{i+1},i=M,M-1,\·\·\·\mathrm{3,2}\end{array}\right.$

Wherein, λ _iseparation, γ _max, γ _minbe respectively maximum and the minimum value of subcarrier SINR, κ is the scale factor of nonuniform quantiza, during iteration, according to i order from big to small, carries out.

4. the SEESM implementation method based on exponent number self adaptation μ rule non-uniform quantizing according to claim 1, is characterized in that, in step 4, described quantification, first need to be mapped into corresponding quantized interval by subcarrier SINR according to following rule:

Value γ for i subcarrier SINR _i, when it meets:

γ _i∈(λ _p,λ _p+1],p＝2,3,...Morγ _i∈[λ _p,λ _p+1],p＝1

Its mapping is entered to p quantized interval; Wherein, M is for quantizing exponent number, λ _pfor quantized interval separation.

5. the SEESM implementation method based on exponent number self adaptation μ rule non-uniform quantizing according to claim 1, it is characterized in that, in step 4, described quantification, to the SINR in same quantized interval, using their arithmetic mean value as the quantized value of this quantized interval.

6. the SEESM implementation method based on exponent number self adaptation μ rule non-uniform quantizing according to claim 1, is characterized in that, in step 5, to quantizing subcarrier SINR application EESM algorithm, the final form that obtains equivalent SINR is:

${\mathrm{\γ}}_{\mathrm{eff}}=-\mathrm{\β}\ln \left(\frac{1}{N}\underset{p=1}{\overset{M}{\mathrm{\Σ}}}{M}_p\exp (-\frac{{\stackrel{\‾}{\mathrm{\γ}}}_p}{\mathrm{\β}})\right)$

Wherein, be the quantized value of p quantized interval, M _pbe the number of SINR value in p quantized interval, M is for quantizing exponent number, and N is total number of sub-carriers order, and β is the scale factor relevant to modulation coding mode MCS.

7. the SEESM implementation method based on exponent number self adaptation μ rule non-uniform quantizing according to claim 2, is characterized in that described segmentation factor ρ=0.25.

8. the SEESM implementation method based on exponent number self adaptation μ rule non-uniform quantizing according to claim 3, is characterized in that scale factor κ=0.5 of described nonuniform quantiza.