CN101420403A - Adaptive modulation method and apparatus - Google Patents

Abstract

The present invention discloses a self-adapting modulating method and a device thereof, wherein the method comprises the following steps: confirming a mode selection matrix according to a Markov Model of single input single output channel, and in the step the mode selection matrix provides a modulating mode which is adopted by the single input single output channel corresponding to each Markov state at the next time; and inquiring the mode selection matrix according to the Markov state in which the single input single output channel is for confirming the modulating mode that is adopted by the single input single output channel at the next time in the transmission process of information symbol. The self-adapting modulating method and device according to the invention can guarantee the transmission quality of system and have lower complexity at a state of deteriorated information quality of channel state caused by higher Doppler frequency of channel.

Description

Self-adaptive modulation method and device

Technical field

The present invention relates to wireless communication technology field, particularly be applicable to the self-adaptive modulation method and the device of high-speed radiocommunication system.

Background technology

Multiple-input and multiple-output (MIMO, Multiple-Input Multiple-Output) technology can increase substantially the spectrum efficiency of wireless communication system.OFDM (OFDM, Orthogonal FrequencyDivision Multiplexing) technology is converted into flat fading channel to frequency-selective channel, thereby can resists multidiameter delay effectively by increasing certain Cyclic Prefix.The MIMO-OFDM system that proposes in conjunction with the advantage of MIMO and two kinds of technology of OFDM has been subjected to extensive concern as a kind of high-speed radiocommunication system.

For the non-self-adapting system, adaptive modulation technology can rationally be adjusted the transmission parameter according to current state information of channel, thus the systematic function of significantly improving.In the MIMO-OFDM system, use adaptive modulation technology, rationally adjust modulating mode according to the MIMO-OFDM channel conditions and can reach the purpose that improves the availability of frequency spectrum, guarantees communication quality.

Traditional self-adaptive modulation method generally is to suppose by channel estimating or predicting under the situation that can accurately obtain channel condition information and carry out.But, the method of directly utilizing receiving terminal channel condition information estimated value to carry out Adaptive Modulation only is applicable to that channel fading changes enough slow situation, when Doppler frequency is higher, channel variation is too fast, this method will produce bigger performance loss, even causes self-adaptive modulation method unavailable owing to bit error rate is too high.In order to resist the performance loss that fast change of channel is brought, can in adaptive modulation system, introduce channel predictor channel variation is predicted.Yet this method by prediction acquisition channel condition information can only alleviate the performance loss that fast change of channel is brought to a certain extent.Along with the raising of Doppler frequency, the channel estimating error will increase, and the performance of BER of Adaptive Modulation algorithm will significantly worsen.In real system, in case bit error rate surpasses the patient maximum of system, message transmission will become invalid transmission, thereby limit the application of this self-adaptive modulation method.

This shows, when Doppler frequency is higher cause the channel condition information quality to drop to a certain degree after, existing self-adaptive modulation method substantially all can't satisfy system's bit error rate requirement, thereby has limited the application of adaptive modulation technology in practical communication system.

Summary of the invention

In order to solve the problems of the technologies described above, the invention provides a kind of self-adaptive modulation method and device, even make under the Doppler frequency condition with higher, still can satisfy system's bit error rate requirement.

The described self-adaptive modulation method of the embodiment of the invention comprises:

According to the Markov model deterministic model selection matrix of the single delivery channel of single input, described model selection matrix has provided the single delivery channel of single input of corresponding each markov state at next modulating mode that adopts constantly;

In the information symbol transmission course,, determine that the single delivery channel of described single input is at next modulating mode that adopts constantly according to the described model selection matrix of described single input single delivery channel present located markov status poll.

The self-adaptive modulation method of the described MIMO-OFDM of the embodiment of the invention comprises:

Be the K group of described MIMO-OFDM system channel equivalence, every group of N _TThe single output of individual single input parallel sub-channels is set up N _TIndividual Markov model, every group of N _TThe respectively corresponding Markov model of individual single input single output parallel sub-channels, wherein, K is the sub-carrier number of system, N _TNumber for MIMO-OFDM system transmitting antenna;

Markov model deterministic model selection matrix according to the single output of each single input parallel sub-channels, wherein, described model selection matrix has provided corresponding each markov state, and the single output of each single input parallel sub-channels is at next modulating mode that adopts constantly;

In the information symbol transmission course,, determine that the single output of each single input parallel sub-channels is at next modulating mode that adopts constantly respectively according to the described model selection matrix of each single input single output parallel sub-channels present located markov status poll.

The described Adaptive Modulation device of the embodiment of the invention comprises:

The mode matrix generation unit is used for the Markov model deterministic model selection matrix according to the single delivery channel of single input; Described model selection matrix has provided corresponding each the markov state of the single delivery channel of described single input at next modulating mode that adopts constantly;

The modulating mode selected cell, be used for transmission course at information symbol, according to the described model selection matrix of described single input single delivery channel present located markov status poll, determine the modulating mode that the single delivery channel of described single input selects next to adopt constantly.

By self-adaptive modulation method of the present invention and the device as can be seen, even under the higher situation that causes the channel condition information deterioration of channel Doppler frequency, self-adaptive modulation method that the embodiment of the invention provided and device still can satisfy the requirement of system's bit error rate, and through experiment showed, that the self-adaptive modulation method that the embodiment of the invention provides and the implementation complexity of device are starkly lower than traditional self-adaptive modulation method that utilizes channel predictor.

In addition, optimization aim difference according to Adaptive Modulation, self-adaptive modulation method that the embodiment of the invention provided and device can also be by selecting the modulating mode that satisfies system's bit error rate requirement, for example select the bigger modulating mode of information symbol transmission rate, thereby satisfying under the situation that the system bit error rate requires the further performance of optimization system message transmission.

Description of drawings

To make clearer above-mentioned and other feature and advantage of the present invention of those of ordinary skill in the art by describe exemplary embodiment of the present invention in detail with reference to accompanying drawing below, in the accompanying drawing:

Fig. 1 has shown the described pre-training flow process of the embodiment of the invention;

Fig. 2 has shown the described transfer process of the embodiment of the invention;

Fig. 3 has shown the described method according to the Markov model deterministic model selection matrix of being set up of the embodiment of the invention;

Fig. 4 has shown the example according to the determined model selection matrix of the embodiment of the invention;

Fig. 5 is the described Adaptive Modulation apparatus structure of an embodiment of the invention schematic diagram;

Fig. 6 (a)～(c) self-adaptive modulation method that adopts the embodiment of the invention to provide is provided, utilizes between traditional self-adaptive modulation method and the desirable self-adaptive modulation method, increase along with travelling carriage translational speed in the MIMO-OFDM system, the bit error rate of MIMO-OFDM system, effective information speed and throughput performance be schematic diagram relatively;

Fig. 7 is that the complexity of the described self-adaptive modulation method of the embodiment of the invention and traditional self-adaptive modulation method compares schematic diagram.

Embodiment

For making purpose of the present invention, technical scheme clearer, below with reference to the accompanying drawing embodiment that develops simultaneously, the present invention is described in further details.

Be that applied environment describes self-adaptive modulation method provided by the invention and device in detail with the MIMO-OFDM system below.

In the described MIMO-OFDM of embodiment of the invention system, total K subcarrier, N _TRoot transmitting antenna and N _RThe root reception antenna, wherein, N _R〉=N _TSuppose the Cyclic Prefix (CP of MIMO-OFDM system, Cycle Prefix) long enough, then in the MIMO-OFDM system, there is not intersymbol interference (ISI, Inter-Symbol Interference), yet, owing to be subjected to the influence of Doppler frequency, exist in the MIMO-OFDM system and disturb (ICI, Inter-carrier Interference) between subcarrier; The transmitting terminal of MIMO-OFDM system is multiplexed into information symbol on each antenna by serial/parallel conversion, and can be according to the adaptively selected m+1 kind of channel conditions modulating mode { M _i, i=0 ..., any one modulating mode among the m} sends described information symbol, wherein, and M ₀Represent that this antenna does not send information, { M in this mark space _i, i=1 ..., m} represents the modulating mode that the m kind is different.In addition, suppose that power is identical on every transmitting antenna of MIMO-OFDM system, and total transmitted power is constant; Receiving terminal can accurately be estimated this channel condition information constantly, and (t k) represents the channel fading coefficient matrix of k subcarrier in t mark space with H.

Based on the MIMO-OFDM system that satisfies above-mentioned condition, the optimization aim of the described self-adaptive modulation method of present embodiment is to select a kind of modulating mode under the prerequisite that satisfies system's bit error rate requirement, with the transmission rate of maximization information symbol, promptly satisfy following formula (1).

$\max \underset{k=1}{\overset{K}{\mathrm{\Σ}}}\underset{j=1}{\overset{N_T}{\mathrm{\Σ}}}{R}_j\left(k\right)---\left(1\right)$

BER≤BER _target

Wherein, R _j(k) transmission rate on the j root antenna of corresponding k the subcarrier of expression, 1≤k≤K and 1≤j≤N _T, BER represents system's bit error rate, BER _TargetIt is the patient maximum bit error rate of system.

By eliminating many inter-antenna interference, under the situation of not considering error propagation, the multi-input multi-ouput channel of above-mentioned MIMO-OFDM system can equivalence be organized every group of N for K _TThe parallel sub-channels of the single output of individual single input, wherein, K is the sub-carrier number of MIMO-OFDM system.In this case, in order to make the represented optimization aim of formula (1) be easier to calculate, the transmission rate of above-mentioned optimization aim Approximate Equivalent for the parallel sub-channels of the single output of each single input of maximization under the prerequisite that meets the demands in the parallel sub-channels bit error rate that guarantees the single output of each single input promptly can be satisfied following formula (2).

maxR _j(k)

(2)

BER≤BER _target

Wherein, R _j(k) transmission rate on j the parallel sub-channels of corresponding k the subcarrier of expression, 1≤k≤K, 1≤j≤N _TBER represents the bit error rate of the parallel sub-channels of the single output of each single input.

In order to realize above-mentioned optimization aim, the described self-adaptive modulation method of present embodiment can be realized by two processes, promptly pre-training process and transmission course.

Specifically, the described pre-training process of present embodiment as shown in Figure 1, mainly comprises:

Step 101: in the duration is in the time interval of T symbol, and transmitting terminal sends known symbol, and receiving terminal carries out channel estimating according to the symbol that is received;

Step 102: receiving terminal is set up Markov model respectively according to the single output of each the single input parallel sub-channels on channel estimation results each subcarrier that the multi-input multi-ouput channel equivalence obtains to the MIMO-OFDM system;

Step 103: receiving terminal is according to the Markov model deterministic model selection matrix of being set up.

Wherein, when described model selection matrix has provided respectively corresponding each the markov state of the single output of each single input parallel sub-channels on each subcarrier, at next modulating mode that constantly should adopt.

The described method of above-mentioned steps 102 and step 103 will be described in detail later.

The described transmission course of present embodiment as shown in Figure 2, mainly comprises:

Step 201: receiving terminal carries out channel estimating, and determines the single output of each single input parallel sub-channels present located markov state on each subcarrier according to channel estimation value.

Specifically, the Markov model that corresponding above-mentioned steps 102 is set up for the single output of each the single input parallel sub-channels on each subcarrier, in this step, receiving terminal will be determined the single output of each single input parallel sub-channels present located markov state on each subcarrier respectively according to channel estimation value.

Step 202: receiving terminal is according to each parallel sub-channels present located markov status poll model selection matrix on each subcarrier of determining, for each parallel sub-channels on each subcarrier is determined modulating mode, and this modulating mode is fed back to transmitting terminal.

Because each single input list that described model selection matrix has provided on each subcarrier is exported corresponding respectively each the markov state of parallel sub-channels at next modulating mode that constantly should adopt, therefore, according to the residing markov state of determined current this parallel sub-channels of step 201, can from described model selection matrix, search and obtain the single output of this list input parallel sub-channels at next modulating mode that constantly should adopt.

Step 203: transmitting terminal carries out Adaptive Modulation to each parallel sub-channels on each subcarrier respectively according to the modulating mode of receiving terminal feedback, thereby realizes Adaptive Transmission.

To describe present embodiment below in detail and in above-mentioned steps 102, set up the process of Markov model for the single output of each the single input parallel sub-channels on each subcarrier.

A known Markov model is made of three key elements: state set, Stationary Distribution sequence and state transition probability matrix.

In the present embodiment, the state set of the single output of a single input parallel sub-channels Markov model is be that the quantized interval uniform quantization should list be imported list and exported parallel sub-channels master signal to noise ratio (SNR with Δ (dB of unit), Signal Noise Ratio) interval [D, U] resulting N signal to noise ratio subinterval

Wherein, each subinterval is corresponding to a markov state S _n, n=1 ..., N.

In the present embodiment, the Stationary Distribution sequence of the single output of a single input parallel sub-channels Markov model is { π _n, n=1 ..., N}, wherein, π _nRepresent that the single output of this list input parallel sub-channels is at markov state S _nUnder the Stationary Distribution probability.In actual applications, each single single output parallel sub-channels Stationary Distribution probability under each markov state of importing can obtain according to the single probability density function of exporting the signal to noise ratio of parallel sub-channels of this list input, promptly calculates by following formula (3).

${\mathrm{\π}}_n={\∫}_{D_{S_n}}{p}_{r_j}\left(x\right)\mathrm{dx}---\left(3\right)$

Wherein,

Expression markov state S _nPairing signal to noise ratio subinterval,

(for example represent the single output of this list input parallel sub-channels, j subchannel) probability density function of signal to noise ratio, can obtain by mathematical derivation or emulation, for example in simulation process, can investigate the signal to noise ratio probability density function that each channel state of signal-to-noise that constantly presents of each parallel sub channel is determined the single output of each single input parallel sub-channels respectively according to ergodic theorem.

Except the described computational methods of above-mentioned formula (3), the Stationary Distribution sequence of each single input list output parallel sub-channels Markov model also can obtain by the number of times that this list input list output parallel sub-channels of statistics in pre-training process is in each markov state.

In the present embodiment, the state transition probability matrix of the single output of a single input parallel sub-channels Markov model can obtain by the number of times that the single output of this list input of statistics in pre-training process parallel sub-channels shifts between each markov state.In order to express easily, in the present embodiment, use p _{L, n}Represent that the single output of this list input parallel sub-channels is from markov state S _lTo markov state S _nTransition probability, obtain the state transition probability matrix P={p of the single output of this list input parallel sub-channels Markov model thus _{L, n}, l, n=1 ..., N}.

In pre-training process, can determine state set, Stationary Distribution sequence and the state transition probability matrix of the Markov model of the single output of each the single input parallel sub-channels on each subcarrier by aforementioned calculation or statistic processes, thereby determine the Markov model of the single output of each the single input parallel sub-channels on each subcarrier.

Known in above-mentioned MIMO-OFDM system, the channel on each subcarrier is with distributing, therefore, the multi-input multi-ouput channel of MIMO-OFDM system the parallel sub-channels of the equivalent single output of the single input of K group can shared Markov model.But, the N in each group _TThe single output of individual single input parallel sub-channels is different distributions often, therefore, is necessary for the N in each group _TThe single output of individual single input parallel sub-channels is set up Markov model respectively.That is to say, in the described MIMO-OFDM of present embodiment system, the single output of corresponding single input parallel sub-channels on the different sub carrier can a shared Markov model, promptly the 1st of this MIMO-OFDM system the subcarrier the 1st the single output of single input parallel sub-channels to K the subcarrier can shared the 1st Markov model, the 1st subcarrier the 2nd the single output of single input parallel sub-channels to K the subcarrier can shared the 2nd Markov model, ..., and the like, the 1st subcarrier N to K the subcarrier _TThe single output of individual single input parallel sub-channels can shared N _TIndividual Markov model.Thus, the MIMO-OFDM system need set up N altogether _TThe Markov model of the single output of individual single input parallel sub-channels, and need not to set up K * N _TIndividual Markov model, and, this N _TThe Markov model of the single output of individual single input parallel sub-channels can constitute the Markov model of a mimo channel.Thus, as can be seen, utilize the same distribution character of each subcarrier upper signal channel of MIMO-OFDM system, can be with the quantity of the Markov model set up from K * N _TThe individual N that is reduced to _TIndividual, thus the complexity of the described adaptive approach of present embodiment reduced greatly.

Further, utilize the same distribution character of each subcarrier upper signal channel of MIMO-OFDM system, can be K pre-training of mimo channel on the same subcarrier that continues the once training in advance that time span is the MIMO-OFDM channel of T mark space, regard persistence length as to be T.That is to say, the state transitions situation of observing the single output of each the single input parallel sub-channels on K the subcarrier in T symbol duration simultaneously can obtain K * T pre-training result altogether, be equivalent to pre-training K * T time, thus, the state of observing and add up the single output of each the single input parallel sub-channels on all subcarriers of MIMO-OFDM system in pre-training period simultaneously can obtain accurate relatively Markov model in the short pre-training time, shorten the required time of pre-training process greatly.

Describe in the present embodiment step 103 receiving terminal below again in detail according to the method for the Markov model deterministic model selection matrix of being set up, as shown in Figure 3, the described model selection matrix of present embodiment can be determined by following steps:

Step 1031: the bit error rate of corresponding each modulating mode when calculating each parallel sub-channels on each subcarrier respectively and be in any one markov state according to the Markov model of each parallel sub-channels on each subcarrier of having set up.

As mentioned above, because each subcarrier of the described MIMO-OFDM of present embodiment system can a shared Markov model, the bit error rate of corresponding each modulating mode when therefore, above-mentioned steps 1031 can be reduced to and calculate the single output of each single input parallel sub-channels respectively and be in any one markov state.

In the present embodiment, suppose that j the single output of single input parallel sub-channels is in markov state S _nAnd adopt M _iBit error rate when planting modulating mode is BER _j(M _i, S _n).BER _j(M _i, S _n) can adopt following formula (4) to calculate.

${\mathrm{BER}}_j(M_i,S_n)=\frac{{\∫}_{D_{S_n}}{p}_{r_j}\left(x\right)\mathrm{BER}(M_i,x)\mathrm{dx}}{{\mathrm{\π}}_n}---\left(4\right)$

Wherein, BER (M _i, x) M is adopted in expression _iKind of the modulating mode approximate bit error rate when signal to noise ratio is x, can by or measure or obtain by actual emulation according to passing through the signal to noise ratio under every kind of modulating mode that theoretical derivation goes out and the relation formula of bit error rate.

Except the computational methods shown in the above-mentioned formula (4), BER _j(M _i, S _n) can also be directly according to above-mentioned BER (M _i, x) obtain, for example, ask BER (M _i, x) exist $x\⋐D_{S_n}$ Maximum in the subinterval, minimum value or mean value or the like.

Step 1032: according to the bit error rate that step 1031 calculates, corresponding each markov state calculates the expection bit error rate of each single input single output parallel sub-channels when next adopts each modulating mode constantly on each subcarrier respectively.

In like manner, because each subcarrier of the described MIMO-OFDM of present embodiment system can a shared Markov model, above-mentioned steps 1032 can be simplified corresponding each markov state, calculates the expection error rate of each single input single output parallel sub-channels when next adopts each modulating mode constantly respectively.

Specifically, in the present embodiment, suppose that j the single output of single input parallel sub-channels is in markov state S _lAt present one adopt M constantly _iThe expection bit error rate of planting modulating mode is

Then

Can adopt following formula (5) to calculate.

${{\mathrm{BER}}_j}^{\mathrm{expect}}(M_i,S_l)=\underset{n=1}{\overset{N}{\mathrm{\Σ}}}{p}_{l,n}^j{\mathrm{BER}}_j(M_i,S_n)---\left(5\right)$

From above-mentioned formula (5) as can be seen, j current markov state S that is in of single input single output parallel sub-channels _lSituation under adopt M constantly at next _iExpection bit error rate when planting modulating mode and the single output of this list input parallel sub-channels are at next markov state S that constantly may transfer to _nAnd by markov state S _lTo markov state S _nBetween transition probability relevant, that is to say that should list importing list by current markov state transitions to the state transition probability weighted average of any markov state with the single output of this list input parallel sub-channels exports parallel sub-channels at following weighted average that the corresponding error rate obtains of the markov state of being transferred to.

Step 1033: corresponding each the markov state of each single input single output parallel sub-channels on each subcarrier of determining by above-mentioned steps 1032 is in next expection bit error rate constantly, corresponding each markov state, determine the modulating mode set that each parallel sub-channels on each subcarrier is formed at next satisfying of constantly can adopting modulating mode that system's bit error rate requires respectively, just determine institute corresponding expect bit error rate be less than or equal to system the set formed of the modulating mode of patient maximum bit error rate.

Corresponding above-mentioned steps 1032, the modulating mode M of satisfied following formula (6) _iThe modulating mode that satisfies system's bit error rate requirement that can select exactly.

${{\mathrm{BER}}_j}^{\mathrm{expect}}(M_i,S_l)\≤{\mathrm{BER}}_{t\arg \mathrm{et}}---\left(6\right)$

Wherein,

Be that j parallel sub-channels is from markov state S _lTransfer to markov state S _nProbability; BER _j(M _i, S _n) as previously mentioned, represent that j parallel sub-channels is in markov state S _nShi Caiyong M _iBit error rate when planting modulating mode.

Step 1034: from above-mentioned steps 1033 resulting modulating mode set, select a kind of modulating mode, for example select modulating mode with maximum information symbol transmission speed.

Need to prove that the system of selection of the modulating mode in the above-mentioned steps 1034 is relevant with system self-adaption modulator approach optimization aim.For example, corresponding present embodiment, as previously mentioned, the optimization aim of self-adaptive modulation method that present embodiment provides is to select a kind of modulating mode under the prerequisite that satisfies system's bit error rate requirement, transmission rate with the maximization information symbol, therefore, in above-mentioned steps 1034, specifically be the modulating mode that selection has maximum information symbol transmission speed from the modulating mode set.Therefore, corresponding different self-adaptive modulation method optimization aim, can adopt diverse ways from the modulating mode set, to select modulating mode, for example, reach under the situation of aims of systems transmission rate and minimize minimum bit error rate if the optimization aim of system self-adaption modulator approach is the transmission rate at information symbol, then above-mentioned steps 1034 can replace with, and selects information symbol rate more than or equal to aims of systems transmission rate and modulating mode with minimum expected bit error rate from the resulting modulating modes set of above-mentioned steps 1033.

Can obtain the modulating mode that corresponding each the markov state of each parallel sub-channels on each subcarrier should adopt in next moment, i.e. model selection matrix by above-mentioned steps 1031～1034.After determining described model selection matrix, receiving terminal will be stored described model selection matrix.

Fig. 4 shows the model selection matrix example of determining according to the described method of the foregoing description.Wherein, used channel model is COST207 channel 6 footpath models, and concrete parameter is as shown in table 1.

Path number	Relative time delay (μ s)	Power attenuation (dB)
			1	0	-2.5
2	0.3	0
			3	1.0	-3
4	1.6	-5
			5	5.0	-2
6	6.6	-4

Table 1

The centre frequency of each channel is 5GHz in above-mentioned channel model, and bandwidth is 2.5MHz, and average signal-to-noise ratio is 35dB, and there are 256 subcarriers in the MIMO-OFDM system, 4 transmitting antennas, and 4 reception antennas, the translational speed of travelling carriage is 40km/h.The mimo channel of 4 transmitting antennas and 4 reception antennas can form the single output of the single input parallel sub-channels of 4 equivalences altogether, therefore, need set up 4 Markov models altogether.Transmitting terminal has 6 kinds of optional modulating mode: BPSK, QPSK, 8PSK, 16QAM, 32QAM and 64QAM, in the model selection matrix, represent with 1,2,3,4,5,6 respectively, and because transmitting terminal can select not send information under the extremely low situation of signal to noise ratio, this situation is represented with 0.The main quantized interval of signal to noise ratio be 5dB to 30dB, quantized interval is 0.1dB, has 251 markov states.Here, select for use this less main quantized interval can reduce the channel status number on the one hand, can not influence the performance of algorithm on the other hand substantially.

Model selection matrix shown in Figure 4 is divided into 6 hurdles, and the label of each left side, hurdle one row is represented the numbering 1～251 of each markov state, and each hurdle top line is represented the numbering 1～4 of each parallel sub-channels.For example, model selection matrix the 150th row the 2nd column element is 4 expressions when if the single output of the single input parallel sub-channels of the 2nd equivalence of mimo channel is in the 150th markov state, at next delivery time, this subchannel should be selected the 4th kind of modulating mode for use, and promptly 16QAM comes transmission information.

After receiving terminal obtained above-mentioned model selection matrix, pre-training process finished.Carry out transmission course.As shown in Figure 2, in transmission course, determine the single output of each single input parallel sub-channels present located state by receiving terminal according to channel estimation value, inquire about above-mentioned model selection matrix, thereby determine the modulating mode of the next delivery time of this parallel sub-channels, and give transmitting terminal this feedback information.At transmitting terminal, according to the modulation parameter of feedback each subchannel is carried out Adaptive Modulation, thereby realize Adaptive Modulation.

The described self-adaptive modulation method of corresponding the foregoing description, embodiments of the invention give a kind of self adaptation debugging apparatus, and its internal structure mainly comprises as shown in Figure 5:

Mode matrix generation unit 501, be used for Markov model deterministic model selection matrix according to the parallel sub-channels correspondence of the single output of each single input on each subcarrier, promptly by the single output of each the single input parallel sub-channels on each subcarrier, corresponding each markov state, the matrix that the modulating mode that should adopt in next moment is formed;

Modulating mode selected cell 502, be used for transmission course at information symbol, according to each single described model selection matrix of list output parallel sub-channels present located markov status poll of importing, determine that each the single input list on each subcarrier is exported parallel sub-channels at next modulating mode that adopts constantly respectively.

Wherein, described mode matrix generation unit 502 further comprises:

First channel estimation module 5011 is used for carrying out channel estimating at pre-training process;

Markov model is set up module 5012, is used for setting up Markov model according to the channel estimation results of first channel estimation module respectively for the single output of each the single input parallel sub-channels on each subcarrier, and detailed process is referring to step 102;

Bit Error Rate Computation module 5013, be used for Markov model according to the single output of each the single input parallel sub-channels on each subcarrier of having set up, corresponding each markov state, each the single input list that calculates respectively on each subcarrier is exported the bit error rate that parallel sub-channels adopts each modulating mode, and detailed process is referring to step 1031;

The modulation pattern set symphysis becomes module 5014, be used for according to the bit error rate that calculates, corresponding each markov state, determine the expection error rate of each single input single output parallel sub-channels when next adopts each modulating mode constantly on each subcarrier respectively, and determine that according to the determined expection error rate the single output of each single input parallel sub-channels on each subcarrier can adopt at next constantly, satisfy the modulating mode set that system's bit error rate requires, detailed process is referring to step 1032 and 1033;

Modulating mode is selected module 5015, be used for corresponding each markov state, select a kind of modulating mode to import single output parallel sub-channels at next modulating mode that adopts constantly as each list on each subcarrier respectively from described modulating mode set, detailed process is referring to step 1034.

In specific operation process, modulating mode selects module 5015 to select for example modulating mode of transmission rate maximum from described modulating mode set, perhaps adopts other selection mode, to realize different Adaptive Modulation optimization aim.

Described modulating mode selected cell 502 comprises:

Second channel estimation module 5021 is used for carrying out channel estimating in the information symbol transmission course;

State determination module 5022 is used for the channel estimation results according to 5021 outputs of second channel estimation module, determines the single output of each the single input parallel sub-channels present located markov state on each subcarrier respectively;

Debugging mode determination module 5023, be used for according to the single described model selection matrix of parallel sub-channels present located markov status poll of exporting of each single input, determine the modulating mode that the single output of each the single input parallel sub-channels on each subcarrier should adopt in next moment.

By the foregoing description as can be seen, even under the higher situation that causes the channel condition information deterioration of channel Doppler frequency, self-adaptive modulation method that the embodiment of the invention provided and device still can guarantee the transmission quality of system, thereby have enlarged the scope of application of self-adaptive modulation method.

In addition, difference according to the optimization aim of Adaptive Modulation, self-adaptive modulation method that the embodiment of the invention provided and device can also be by selecting satisfying the modulating mode that the system bit error rate requires, thereby satisfying under the situation that system's bit error rate requires the further performance of optimization system message transmission.For example, if the optimization aim of self-adaptive modulation method is under the situation that satisfies system's bit error rate requirement, the transmission rate of maximization system, self-adaptive modulation method that the embodiment of the invention provided and device can satisfy the modulating mode that the system bit error rate requires from all selects the modulating mode of transmission rate maximum, thereby can significantly improve system's effective information speed and throughput.

Need to prove, above embodiment all is that the Adaptive Modulation with the MIMO-OFDM system is that example describes, it will be appreciated by those skilled in the art that, described self-adaptive modulation method of the application and device can also be applied in other any systems, for example, be applied in the single-input single-output system.

To be given in the method for carrying out Adaptive Modulation in the single-input single-output system below, mainly comprise:

At first according to single Markov model deterministic model selection matrix of importing single delivery channel;

In the information symbol transmission course,, determine the single delivery channel of described single input at next modulating mode that adopts constantly, and feed back to transmitting terminal according to the described model selection matrix of described single input single delivery channel present located markov status poll.

The method of above-mentioned deterministic model selection matrix specifically comprises:

At first, corresponding each markov state is determined the bit error rate when the single delivery channel of described single input adopts each modulating mode;

In step, can utilize formula ${\mathrm{BER}}^{\mathrm{expect}}(M_i,S_l)=\underset{n=1}{\overset{N}{\mathrm{\Σ}}}{p}_{l,n}\mathrm{BER}(M_i,S_n)$ Calculating is in markov state S _lThe single delivery channel of single input adopt M constantly at next _iExpection bit error rate BER when planting modulating mode _Expect(M _i, S _l); Wherein, p _{L, n}Represent that the single delivery channel of this list input is from markov state S _lTo markov state S _nTransition probability, BER (M _i, S _n) represent that the single delivery channel of single input is in markov state S _nShi Caiyong M _iBit error rate when planting modulating mode;

Then, corresponding each markov state according to determined bit error rate, is determined the expection bit error rate of described single single delivery channel of input when next adopts each modulating mode constantly;

At last, corresponding each markov state according to described expection bit error rate, determines that the single delivery channel of described single input is at next modulating mode that adopts constantly.

In this step, corresponding each markov state, determine the expection bit error rate of described single single delivery channel of input when next adopts each modulating mode constantly be less than or equal to system the modulating mode of patient maximum bit error rate; Corresponding each markov state, the modulating mode of selecting respectively to have maximum information symbol transmission speed from the modulating mode of determining is imported single delivery channel at next modulating mode that adopts constantly as described list.As previously mentioned, according to the difference of system self-adaption modulation optimization aim, can also from the modulating mode of determining, select a kind of modulating mode to import single delivery channel at next modulating mode that adopts constantly by other method as described list.

Below will be by the performance of the concrete example explanation self-adaptive modulation method that the embodiment of the invention provided.

Fig. 6 (a)～(c) is respectively at one has 256 subcarriers, and 4 transmitting antennas in the MIMO-OFDM system of 4 reception antennas, are 10 in the patient maximum bit error rate of system ^-4Situation under, adopt self-adaptive modulation method that the embodiment of the invention provides, utilize bit error rate, effective information speed and the throughput performance schematic diagram relatively between traditional self-adaptive modulation method and the desirable self-adaptive modulation method.The optimization aim of this MIMO-OFDM system self-adaption modulation is the transmission rate of selecting to satisfy the modulating mode of system's bit error rate requirement and maximizing information symbol.Traditional self-adaptive modulation method as a comparison is for utilizing the self-adaptive modulation method of normalization minimum mean-square (NLMS, Normalized Least MeanSquare) channel predictor.Described desirable self-adaptive modulation method is meant, in the ideal case, the self-adaptive modulation method during known accurate channel condition information, owing to can't know the precise channels state information in practice, therefore this method can't be used in reality.In Fig. 6 (a)～(c), the self-adaptive modulation method of representing the embodiment of the invention to adopt with foursquare curve; With the traditional self-adaptive modulation method that utilizes the NLMS channel predictor of leg-of-mutton curve representative; The curve of band asterisk is represented desirable adaptive approach.Fig. 6 (a)～(c) abscissa is represented the translational speed of travelling carriage, and ordinate is represented bit error rate, effective information speed and throughput respectively.Wherein, the transmission rate of significant bit in the effective information rate representation unit bandwidth unit interval.If the instantaneous bit error rate of a known MIMO-OFDM frame is not more than the patient maximum bit error rate of system, then all bits all are significant bits in this frame, otherwise all bits all are invalid bit in this frame.Instantaneous bit error rate is the transmission rate that comprises bit in 0 the frame in the throughput representation unit bandwidth unit time.

The main simulation parameter of the MIMO-OFDM system of institute's emulation comprises: channel center frequency is 5GHz, and bandwidth is 2.5MHz, and the channel average signal-to-noise ratio is 30dB, and each OFDM is a frame.Be respectively in the available modulation system of transmitting terminal: BPSK, QPSK, 8PSK, 16QAM, 32QAM and 64QAM.Receiving terminal adopts ORTHOGONAL TRIANGULAR (QR, Orthogonal-Triangular Decomposition) is olation to carry out demodulation.Channel adopts the rayleigh fading channel in 6 footpaths, and concrete channel parameter is also as shown in table 1.

Shown in Fig. 6 (a)～(c), on the one hand, raising (corresponding to the raising of Doppler frequency) along with the travelling carriage translational speed, the adaptive approach that the embodiment of the invention provided can satisfy the requirement of system's bit error rate all the time, utilizes the adaptive algorithm of NLMS fallout predictor then too high and unavailable because of bit error rate.On the other hand, utilize the effective information speed of adaptive algorithm provided by the invention and throughput apparently higher than the adaptive algorithm of utilizing the NLMS fallout predictor, and near desirable Adaptive Modulation algorithm.As seen, adaptive algorithm provided by the invention not only can guarantee system transmissions quality and the effective information speed and the throughput of the system that can significantly improve under higher Doppler frequency situation.

By concrete example the complexity of the described self-adaptive modulation method of the embodiment of the invention is analyzed again below.

Suppose that pre-training continues T OFDM mark space altogether, total n state in the Markov model, transmitting terminal has m kind modulating mode.Suppose that the method that adopts QR to decompose is converted into the parallel sub channel with mimo channel.For convenience, calculate the complexity of addition and multiplication at this.In the pre-training stage, at first, utilize QR to decompose and the MIMO-OFDM channel is converted into equivalent subchannel, this process need

Inferior floating-point operation; In second step, calculate the signal to noise ratio of equivalent subchannel and determine its state, this process need 2N _TKT floating-point operation; In the 3rd step, the computing mode transfer matrix needs [2n ²-n] * N _TInferior floating-point operation; In the 4th step, the deterministic model selection matrix needs [2mn ²-mn] * N _TInferior floating-point operation.So,, need the floating-point operation total degree to be altogether in the pre-training stage $3N_T^2(N_R-N_T/3)\mathrm{KT}+2N_T\mathrm{KT}+[2{\mathrm{mn}}^2-\mathrm{mn}+2n^2-n]N_T.$ In transmit stage, each sends in the mark space, only needs to calculate the signal to noise ratio of equivalent subchannel and determine its state, needs 2N altogether _T* K floating-point operation.

Suppose that the NLMS channel predictor contains N tap altogether, its initial phase needs the individual OFDM mark space of T '.Utilize the Adaptive Modulation algorithm initialization stage of NLMS channel predictor to need (22N+3) N _TN _RThe inferior floating-point operation of KT '.This algorithm needs in each mark space of transmit stage $[(22N+3)N_T{N}_R+3N_T^2(N_R-N_T/3)+2N_T]\×K$ Inferior floating-point operation.

Fig. 7 is that the complexity of the described self-adaptive modulation method of the embodiment of the invention and traditional self-adaptive modulation method compares schematic diagram.Wherein, the curve self-adaptive modulation method of representing the embodiment of the invention to adopt of band rhombus; The traditional self-adaptive modulation method that utilizes the NLMS channel predictor of curve representative of band asterisk.It is 1000 OFDM mark spaces that algorithm among the present invention is got pre-training length, and the tap number of NLMS channel predictor is 3, and initialization length is 10 OFDM mark spaces.Abscissa among Fig. 7 is represented the number of transmission symbol, and ordinate is total flops.As shown in Figure 7, transmit 100 symbols after, the complexity of algorithm provided by the invention is reduced to the Adaptive Modulation algorithm that adopts the NLMS channel predictor, and the gap between both complexities increases with the increase of transmission symbol number.

The above only is preferred embodiment of the present invention, and is in order to restriction the present invention, within the spirit and principles in the present invention not all, any modification of being done, is equal to replacement, improvement etc., all should be included within protection scope of the present invention.

Claims (14)

1, a kind of self-adaptive modulation method is characterized in that, comprising:

According to the Markov model deterministic model selection matrix of the single delivery channel of single input, described model selection matrix has provided the single delivery channel of single input of corresponding each markov state at next modulating mode that adopts constantly;

In the information symbol transmission course,, determine that the single delivery channel of described single input is at next modulating mode that adopts constantly according to the described model selection matrix of described single input single delivery channel present located markov status poll.

2, method according to claim 1 is characterized in that, the deterministic model selection matrix comprises:

Corresponding each markov state is determined the bit error rate when the single delivery channel of described single input adopts each modulating mode;

Corresponding each markov state according to determined bit error rate, is determined the expection bit error rate of described single single delivery channel of input when next adopts each modulating mode constantly;

Corresponding each markov state according to described expection bit error rate, determines that the single delivery channel of described single input is at next modulating mode that adopts constantly.

3, method according to claim 2 is characterized in that, utilizes formula ${\mathrm{BER}}^{\mathrm{expect}}(M_i,S_l)=\underset{n=1}{\overset{N}{\mathrm{\Σ}}}{p}_{l,n}\mathrm{BER}(M_i,S_n)$ Calculating is in markov state S _lThe expection bit error rate BER of single single delivery channel of input when next adopts Mi kind modulating mode constantly ^Expect(M _i, S _l); Wherein,

p _{L, n}Represent that the single delivery channel of this list input is from markov state S _lTransfer to markov state S _nTransition probability, BER (M _i, S _n) represent that the single delivery channel of single input is in markov state S _nShi Caiyong M _iBit error rate when planting modulating mode, N represents that the Markov model of the single delivery channel of single input comprises the number of markov state.

4, method according to claim 2 is characterized in that, each markov state of described correspondence according to described expection bit error rate, determines that the single delivery channel of described single input comprises at the modulating mode that next adopts constantly:

Corresponding each markov state, determine the single delivery channel of described single input next expection bit error rate constantly be less than or equal to system the modulating mode of patient maximum bit error rate;

Corresponding each markov state, the modulating mode of selecting respectively to have maximum information symbol transmission speed from the modulating mode of determining is imported single delivery channel at next modulating mode that adopts constantly as described list.

5, method according to claim 1 is characterized in that, determines that the single delivery channel of described single input comprises at the modulating mode that next adopts constantly:

Determine the current markov state of living in of described single single delivery channel of input according to channel estimation results;

According to the described model selection matrix of determined markov status poll, determine the single delivery channel of described single input at next modulating mode that adopts constantly, and feed back to transmitting terminal.

6, the self-adaptive modulation method of a kind of multi-input multi-output orthogonal frequency division multiplexing system MIMO-OFDM is characterized in that, comprising:

Be the K group of described MIMO-OFDM system channel equivalence, every group of N _TThe single output of individual single input parallel sub-channels is set up N _TIndividual Markov model, every group of N _TThe respectively corresponding Markov model of individual single input single output parallel sub-channels, wherein, K is the sub-carrier number of system, N _TNumber for MIMO-OFDM system transmitting antenna;

Markov model deterministic model selection matrix according to the single output of each single input parallel sub-channels, wherein, described model selection matrix has provided corresponding each markov state, and the single output of each single input parallel sub-channels is at next modulating mode that adopts constantly;

In the information symbol transmission course,, determine that the single output of each single input parallel sub-channels is at next modulating mode that adopts constantly respectively according to the described model selection matrix of each single input single output parallel sub-channels present located markov status poll.

7, method according to claim 6 is characterized in that, described Markov model deterministic model selection matrix according to the single output of each single input parallel sub-channels comprises:

Corresponding each markov state is determined the bit error rate when the single output of each single input parallel sub-channels adopts each modulating mode respectively;

Corresponding each markov state according to determined bit error rate, is determined the expection bit error rate of each single input single output parallel sub-channels when next adopts each modulating mode constantly;

Corresponding each markov state according to described expection bit error rate, determines that respectively the single output of each single input parallel sub-channels is at next modulating mode that adopts constantly.

8, method according to claim 7 is characterized in that, utilizes formula ${\mathrm{BER}}_j^{\mathrm{expect}}(M_i,S_l)=\underset{n=1}{\overset{N}{\mathrm{\Σ}}}{p}_{l,n}^j{\mathrm{BER}}_j(M_i,S_n)$ Calculate j the single output of single input parallel sub-channels and be in markov state S _lThe time adopt M constantly at next _iPlant the expection bit error rate of modulating mode ${\mathrm{BER}}_j^{\mathrm{expect}}(M_i,S_l),$ Wherein,

Represent that the single output of this list input parallel sub-channels is from markov state S _lTransfer to markov state S _nTransition probability, BER _j(M _i, S _n) represent that j the single output of single input parallel sub-channels is in markov state S _nShi Caiyong M _iBit error rate when planting modulating mode, N represents that the Markov model of the single delivery channel of single input comprises the number of markov state.

9, method according to claim 7 is characterized in that, each markov state of described correspondence according to described expection bit error rate, determines that respectively the single output of each single input parallel sub-channels comprises at the modulating mode that next adopts constantly:

Corresponding each markov state, determine respectively the single output of each single input parallel sub-channels next expection bit error rate constantly be less than or equal to system the modulating mode of patient maximum bit error rate;

Corresponding each markov state, the modulating mode of selecting respectively to have maximum information symbol transmission speed from determined modulating mode is imported single output parallel sub-channels at next modulating mode that adopts constantly as each list respectively.

10, method according to claim 6 is characterized in that, the single output of described definite each single input parallel sub-channels comprises at the modulating mode that next adopts constantly:

Determine the single output of each single input parallel sub-channels present located markov state respectively according to channel estimation results;

According to the described model selection matrix of determining of markov status poll, determine each parallel sub-channels respectively at next modulating mode that adopts constantly, and feed back to transmitting terminal.

11, method according to claim 6 is characterized in that, the described N that sets up _TIndividual Markov model comprises:

In pre-training process, add up the state of the single output of each the single input parallel sub-channels on K subcarrier of MIMO-OFDM system simultaneously, determine described N _TIndividual Markov model.

12, a kind of Adaptive Modulation device is characterized in that, comprising:

The mode matrix generation unit is used for the Markov model deterministic model selection matrix according to the single delivery channel of single input; Described model selection matrix has provided corresponding each the markov state of the single delivery channel of described single input at next modulating mode that adopts constantly;

The modulating mode selected cell is used for the transmission course at information symbol, according to the described model selection matrix of described single input single delivery channel present located markov status poll, determines next modulating mode that adopts constantly of the single delivery channel of described single input.

13, device according to claim 12 is characterized in that, described mode matrix generation unit comprises:

First channel estimation module is used for carrying out channel estimating at pre-training process;

Markov model is set up module, and the channel estimation results that is used for according to the output of first channel estimation module is that the single delivery channel of described single input is set up Markov model;

The Bit Error Rate Computation module, the bit error rate when being used for Markov model according to the single delivery channel of having set up of described single input and calculating the single delivery channel of the described single input of corresponding each markov state and adopt every kind of modulating mode;

The modulation pattern set symphysis becomes module, be used for according to the bit error rate that calculates, corresponding each markov state, determine the expection bit error rate of described single single delivery channel of input when next adopts each modulation system constantly, and determine next modulating mode that constantly can the adopt set of the single delivery channel of described single input according to the determined expection error rate;

Modulating mode selection module is used for corresponding each markov state, selects a kind of modulating mode to import single delivery channel at next modulating mode that adopts constantly as described list from described modulating mode set respectively.

14, device according to claim 12 is characterized in that, described modulating mode selected cell comprises:

The second channel estimation module is used for carrying out channel estimating in the information symbol transmission course;

The state determination module is used for the channel estimation results according to the output of second channel estimation module, determines the single delivery channel present located markov state of described single input;

The modulating mode determination module is used for according to the described model selection matrix of described single input single delivery channel present located markov status poll, determines that the single delivery channel of described single input is at next modulating mode that adopts constantly.

Images