Embodiment
Do further detailed explanation below in conjunction with the Figure of description specific embodiments of the invention.
As shown in Figure 1, the present invention has designed a kind of radio communication and apparatus measures field of being applied in, and is applicable to the WLAN product test platform of commercial wireless network performance of products test.
Corresponding therewith, another technical problem that will solve of the present invention provides a kind of radio communication and apparatus measures field of being applied in, and is applicable to the method for testing of the WLAN product of commercial wireless network performance of products test.
The present invention adopts following technical scheme in order to solve the problems of the technologies described above: the present invention has designed the present invention and has designed a kind of WLAN product test platform, comprises Wireless LAN testing emission system and Wireless LAN testing receiving system; Said Wireless LAN testing emission system comprises the radio frequency sending set test module and transmits hardware processing module; The radio frequency sending set test module carries out conversion to data to be transmitted; Form baseband complex signal; Transfer to the hardware processing module that transmits and handle the generation radio magnetic wave, and launch through antenna; Said Wireless LAN testing receiving system comprises the radio-frequency transmitter test module and receives the signal hardware processing module; Receive the signal hardware processing module and receive radio magnetic wave through antenna; And handle and obtain receiving sequence; Pass to the radio-frequency transmitter test module again and carry out corresponding inverse transformation, restore information.
As a kind of optimal technical scheme of the present invention: the said hardware processing module that transmits comprises radio-frequency signal generator, and said radio-frequency signal generator comprises vector modulator, signal generator and continuous wave generator; Said reception signal hardware processing module comprises the rf signal analysis appearance, and said rf signal analysis appearance comprises low-converter, 16Bit if digitization appearance and continuous wave generator.
A kind of WLAN product test platform of the present invention's design is in practical implementation; In said transmit hardware processing module and the reception signal hardware processing module, radio-frequency signal generator is that NI PXIe-5673E, rf signal analysis appearance are that NI PXIe-5663, vector modulator are that NI PXIe-5611, signal generator are that NI PXIe-5450, continuous wave generator are that NI PXI-5652, low-converter are that NI PXIe-5601,16Bit if digitization appearance are NI PXIe-5622.
NI PXIe-5673E is the RF Vector Signal Generator that a centre frequency reaches as high as 6.6GHz, can produce the instant bandwidth of 100MHz and the transmitting power of 10dBm.By means of NI PXIe-5673E, can generate different waveforms, comprise all kinds of modulated waveforms, multi-tone signal and other random waveforms.It is made up of NI PXIe-5611 vector modulator, NI PXIe-5450 signal generator and NI PXI-5652 continuous wave generator.NI PXIe-5663 rf signal analysis appearance can provide highest frequency to reach the vector signal analysis ability of 26.5GHz and 50MHz instant bandwidth; Powerful protection against noise interference characteristic also makes it win good reputation: when frequency was in 800MHz, phase noise was merely-129dBc/Hz; When being positioned at 1GHz, make an uproar and also can not surpass-165dBm/Hz in the end.Identical with signal generator, it also is made up of three radio frequency integrated circuit boards, is respectively NI PXIe-5601 low-converter, NI PXIe-562216Bit if digitization appearance and NI PXI-5652 continuous wave generator.Here NI PXI-5652 continuous wave generator and NI PXIe-5673E radio-frequency signal generator all are used to provide local oscillation signal, and require the same homophase frequently of carrier wave of the generation on both sides.
Corresponding therewith, the present invention adopts following technical scheme in order to solve the problems of the technologies described above: the present invention has designed a kind of method of testing of WLAN product, comprises the steps:
Steps A. generate the frame that a group data stream is simulated the MAC layer through the PN sequence generator, as armed information, and the working frequency range of selected data transmission rate and radiofrequency signal;
Step B. calculates the bit information error rate with data flow described in WLAN product emission to be measured and the receiving step A;
Step C. launches also data flow described in the receiving step A with the WLAN product test platform of the present invention's design, calculates the bit information error rate equally, and notes the radio frequency parameter in the radio-frequency transmitter test module;
Step D. is with data flow described in the Wireless LAN testing emission system step of transmitting A of the present invention design, and receives signal, the variation of observation bit information error rate with WLAN product to be measured;
Step e. with data flow described in the WLAN product step of transmitting A to be measured, and, observe the bit information error rate, and note the radio frequency parameter in the radio-frequency transmitter test module with Wireless LAN testing receiving system reception signal;
Step F. at last related data is compared,, obtain the data analysis result according to the quality of protocol of wireless local area network design code evaluation WLAN product to be measured.
Wherein, calculating the bit information error rate is that transmitting terminal passes through global variable and writes down information to be launched; Receiving terminal draws information sequence after resolving and receiving signal, and the two is isotopic number more not, draws bit information error rate BER.
As a kind of optimal technical scheme of the present invention: related data comprises bit information error rate, radio frequency parameter, frame synchronization position, spectrogram and planisphere constantly in the said step F.
As a kind of optimal technical scheme of the present invention: said radio frequency parameter comprises that low-converter gain, outside gain, frequency mixer level, intermediate frequency power output, rf attenuation and intermediate frequency decay.
As a kind of optimal technical scheme of the present invention: said data analysis result comprises frame synchronization position, bit information error rate, time domain waveform figure, spectrogram, planisphere and error message constantly.
Shown in Fig. 2-1,2-2, as a kind of optimal technical scheme of the present invention: transmitting-receiving data flow is the ERP-OFDM pattern in the said WLAN product test method; Wherein data streams comprises: at first successively the frame of MAC layer is carried out framing, scrambling, convolutional encoding, interweaves and mapping, inverse fast Fourier transform and interpolation protection interval by the radio frequency sending set test module; By the hardware processing module that transmits signal is put in order waveform, the modulation of IQ road, high power amplification then, and go out through antenna transmission; Receiving data stream comprises: receive the signal hardware processing module and receive radio magnetic wave through antenna; And radio magnetic wave is carried out linear power successively amplifies with the IQ road and detect, then through the radio-frequency transmitter test module successively to signal carry out synchronously, channel estimating, equilibrium, deletion protect that interval, fast Fourier transform, inverse mapping deinterleave, decoding, descrambling and parse for frame.
" PLCP " of data streams refers to framing, and the frame that is the MAC layer adds frame head, comprises receiver in the frame head and is used for synchronously and the lead code of channel estimating, in addition information such as the speed of mac frame, length; " PLCP " of receiving data stream refers to parse for frame, promptly removes the frame head that adds before, obtains the frame of MAC layer again; " HPA ": high power amplifier is amplified to the desired high power levels of long-distance transmissions with low level signal, to prevent in transmission course, to occur the excessive depletion of signal power on transmitter; " LPA ": linear power amplifier, refer to export the amplifier that the amplitude of signal is directly proportional with input signal amplitude, be used for input signal is amplified to the power level place that needs; " add GI " and promptly add protection at interval.
As a kind of optimal technical scheme of the present invention: said interpolation protection is spaced apart the front that the data of each 16 bit in OFDM symbol back is copied to this symbol, forms prefix, unremitting connection naturally at the interface point place; Said removal protection is spaced apart the data of initial 16 bits of removing each OFDM symbol.Because the transmission signals of ERP-OFDM emission mode is a series of OFDM symbol, transmitted in mutually orthogonal subchannel by 64 number of sub-carrier successively, and each OFDM symbol has 64.In order farthest to eliminate intersymbol interference, (Guard Interval, GI), but the interchannel interference that can bring like this destroys the orthogonality between subchannel at interval as protection can before each OFDM symbol, to insert a segment information.So adopt the protection interval of Cyclic Prefix (Cyclic Prefix) form; Shown in figure 21, the data of 16 bits of each OFDM symbol back are copied to the front of this symbol, form prefix; The unremitting connection naturally at the interface point place; Like this, OFDM symbol is 64+16=80 bit (1 bit=1) altogether, removes the data that GI just removes initial 16 bits of each OFDM symbol.
Wherein, for the WLAN product test platform data streams of ERP-OFDM pattern:
(1) scrambling and descrambling
Data field will pass through the scrambling of scrambler before encoding, the generator polynomial of scrambler is following, and structure is as shown in Figure 3.
S (x)=x
7+ x
4+ 1 formula 1
XOR represented in plus sige among the figure.Before the scrambling, the initial condition of scrambler should be made as the pseudo-random non-zero value, needs a descrambler to come restoring data equally at receiving terminal, and the structure and the scrambler of descrambler are identical, do not make mistakes in order to make data, also will guarantee the synchronism of scrambling, descrambling.
(2) convolutional encoding
The purpose of convolutional encoding is to make that position half the, 1/3 or 1/4 is arranged in each OFDM symbol is the redundant digit that is used for error correction.Encoding rate depends on the speed of data, sees table 2.1 for details.The generator polynomial that convolution coder adopts is g
0=O133, g
1=O171, wherein O representes octal number, the structure of encoder such as Fig. 4.The data that are labeled as " A " are being positioned at data " B " before when encoder is exported.When encoding rate was 1/2, the data of output were coded data; If 2/3 or 3/4, then need to adopt " deleting surplus " technology can obtain coded data; Accordingly, receiver is wanted to insert dummy argument " 0 " in the position of originally being deleted earlier when decoding, then decoding.The general viterbi algorithm that adopts of decoding, Fig. 5 have been explained when encoding rate is 3/4, delete step surplus and the insertion dummy argument.
(3) data interlacing
Data behind the coding also will be called data interlacing through twice displacement.Each coding figure place N with single OFDM symbol
CBPSBe the length that interweaves.Displacement for the first time guarantees that adjacent bits of coded is mapped on the non-conterminous subcarrier; Displacement for the second time guarantees that adjacent bits of coded alternately is mapped on the high significance bit and low order of planisphere, with the long-time appearance of the mapping (enum) data of avoiding having only low order.Equally, receiving terminal also needs the contrary interleaving process of two steps displacement.
Twice displacement that interweaves depends on following two formula respectively:
I=(N
CBPS/ 16) (kmod 16)+floor (k/16) k=0,1 ..., N
CBPS-1 formula 2
J=s * floor (i/s)+(i+N
CBPS-floor (16 * i/N
CBPS)) mod si=0,1 ..., N
CBPS-1 formula 3 in above two formulas, the numbering before k representes to interweave, i is the sequence number after the displacement for the first time, j is the order after accomplishing of interweaving, s is then determined by following formula.
S=max (N
BPSC/ 2,1) formula 4
(4) modulation mapping
The subcarrier of OFDM can adopt the modulation system of BPSK, QPSK, 16-QAM, 64-QAM, and speed is depended in concrete selection equally.The planisphere that modulation is adopted, shown in Fig. 6-1,6-2,6-3,6-4, these mapping relations are that data can be mapped as a sequence of complex numbers after ovennodulation, i.e. { the form of d|I+jQ} according to the principle design of Gray code.In order to obtain identical average power, need multiply by normalized factor k to sequence of complex numbers to all mappings
MOD, k
MODDecide according to modulation system, see table 1 for details.
The corresponding relation of table 1 normalization factor and modulation system
As a kind of optimal technical scheme of the present invention: under the said ERP-OFDM pattern, in the data streams step, also comprise the pilot sub-carrier operation between map operation and inverse fast Fourier transform operation.
Under the situation that perhaps exists phase noise to disturb in the occurrence frequency skew, receiver still can carry out stable coherent detection, need in the OFDM symbol, insert 4 pilot signals.These 4 pilot signals are inserted in label respectively on-21 ,-7,7,21 the sub-carrier positions.For several continuous OFDM symbols, the polarity of the pilot tone that adjacent-symbol inserted also has nothing in common with each other, and the polarity of pilot tone is determined that by the output sequence of the scrambler of " 1 " this sequence is a periodic sequence that length is 127 points entirely by initial condition.Certainly, this sequence will be passed through necessary cyclic extensions, with the number of coupling OFDM symbol.After inserting pilot tone, the data that obtain just can have been carried out inverse fast Fourier transform IFFT by 64 bit groupings then.
Shown in Fig. 7-1,7-2; As a kind of optimal technical scheme of the present invention: transmitting-receiving data flow is the ERP-DSSS/CCK pattern in the said WLAN product test method; Wherein data streams comprises: at first successively the frame of MAC layer is carried out framing, scrambling, coding and spread spectrum by the radio frequency sending set test module; By the hardware processing module that transmits signal is carried out the modulation of IQ road, high power amplification then, and go out through antenna transmission; Receiving data stream comprises: receive the signal hardware processing module and receive radio magnetic wave through antenna; And radio magnetic wave is carried out linear power successively amplifies and the detection of IQ road, then through the radio-frequency transmitter test module successively to signal carry out synchronously, channel estimating, equilibrium, despreading, decoding, descrambling and parse for frame.
Wherein, for the WLAN product test platform data streams of ERP-DSSS/CCK pattern:
(1) scrambling and descrambling
The frame of physical layer needed through scrambler before band spectrum modulation, and structure is as shown in Figure 8, and the generator polynomial of scrambling is
G (x)=z
-7+ z
-4+ 1 formula 5
For long PLCP, the initial condition of scrambler is [110 1100], wherein z
1=1 ..., z
7=0.And the PLCP of short lead code is in contrast, with [0,011 011] to scrambler initialization.
Because the feedback configuration between scrambler and the descrambler is synchronous, so receiving terminal needn't be known the initial condition of scrambler in advance.The descrambling multinomial is identical with scrambling, the structure of descrambler such as Fig. 9, and certainly, this is not unique structure.
For the scrambling/descrambling operation under ERP-OFDM pattern and the ERP-DSSS/CCK pattern; Transmission rate according to selected data; Promptly confirmed which kind of pattern of this The data, the concrete principle of correspondence is: " ERP-OFDM pattern " comprises 6,9,12,18,24,36,48,54; " ERP-DSSS/CCK pattern " comprises 1,2,5.5,11, and unit is Mbps.Come to use respectively the scrambling/descrambling operation under the different mode.
(2) basis modulation and spread spectrum
ERP-DSSS/CCK has stipulated four kinds of speed and corresponding modulation mode.Basic access speed (1Mbps, 2Mbps) has adopted the modulation system of differential phase keying (DPSK) and high speed chip spread spectrum; The data rate of spreading rate then has been based on CCK modulation technology 5.5Mbps and 11Mbps will be explained mode separately respectively below.
Differential phase keying (DPSK) is the basis of DSSS system.It is encoded to data with the phase difference of transmission signals.Therefore the absolute phase of waveform is unimportant, importantly the difference of phase place.It can overcome the problem of separating timing phase inversion.In the DSSS system, 1Mbps has adopted the modulation system of DBPSK, and 2Mbps then is DQPSK.The coding schedule of the two is respectively shown in table 2 and the table 3.
Table 2 1Mbps DBPSK coding schedule
Input bit | Phase difference | |
0 |
0 |
Table 3 2Mbps DQPSK coding schedule
The dibit pattern | Phase difference | |
00 |
0 |
01 |
π/2 |
11 |
π |
10 |
3π/2(-π/2) |
Data behind the coding are utilized its frequency spectrum of Barker code expansion of 11:
+1,-1,+1,+1,-1,+1,+1,+1,-1,-1,-1
Output is from left to right in proper order, and first code element is alignd with a starting point of sending symbol, and the duration of each spread spectrum is 11 Baud Lengths.
(3) CCK modulation
The extended code length of CCK modulation is 8 and based on complement form, in 802.11g, obtains these 8 plural code words through formula:
C is a code word in the formula, c={c
0..., c
7, c
0Send at first, therefore, as long as obtain
The extended code that can obtain exporting.Wherein
Be to adjust the whole sequence initial phase, by the front two d of input information
0, d
1Decision, d
0, d
1Encode based on DQPSK, obtain
Coding rule such as table 4 definition, the variation of phase place is by rotation counterclockwise.Because be differential coding, so the phase change of
is relevant with
of previous symbol.For first CCK symbol of PSDU, its
just depends on the phase place of last symbol of front PLCP head.
Table 4 DQPSK coding schedule
d
0,d
1 |
Idol sign change phase place |
The odd symbol variation phase |
00 |
0 |
π |
01 |
π/2 |
3π/2(-π/2) |
11 |
π |
0 |
10 |
3π/2(-π/2) |
π/2 |
The value of all the other of input information three phase places of common decision residue.The rule as follows:
For 5.5Mbps, data are encoded with 4 bit groupings, and each grouping is d
0, d
1, d
2, d
4d
0, d
1As stated, d
3, d
4Through following three formula, calculate
With
formula 7
Table 5QPSK coding schedule
Data dibit (d
i,d
i+1)
|
Phase place |
00 |
0 |
01 |
π/2 |
10 |
π |
11 |
3π/2(-π/2) |
When speed is 11 Mbps, produce the information data that each CCK symbol needs 8 bits, i.e. d
0, d
1, d
2, d
3, d
4, d
5, d
6, d
7, front two produces phase place according to table 4 equally
Dividing into groups in twos in six of backs, forms data dibit (d
2, d
3), (d
4, d
5), (d
6, d
7), the QPSK that stipulates according to table 5 respectively again is encoded to
With
Note the difference of table 5 and table 3, what adopt here is binary coding, but not Gray code.
After calculating four phase places according to table 4 and table 5, in the substitution formula 6, just can draw the spread spectrum code character of CCK, after head and the tail connect successively, form the plural bit stream before the emission.
For " detection of IQ road ", corresponding to " modulation of IQ road " of radio frequency sending set test module, what the radio frequency sending set test module produced is baseband complex signal; " modulation of IQ road " then is to extract real part and imaginary part; Constitute I road and Q road respectively, again the IQ two paths of signals is modulated respectively, form electromagnetic wave; Still stack up, then emission according to plural form.The radio-frequency transmitter test module will carry out " detection of IQ road " to received signal, demodulation again behind the electromagnetic wave on separation I road and Q road, reduction baseband complex signal.
Shown in Fig. 2-2,7-2, for WLAN product test platform receiving data stream:
1. simultaneous techniques
(1) synchronously thick
In realization, frame synchronization can be divided into thick synchronously with synchronously smart again.Slightly do not need reference sequences synchronously, the principle of employing remains the variation according to list entries power, promptly when useful signal gets into receiver, because the power of its power noise in the surrounding environment, so the performance number of list entries can be undergone mutation.At this moment just can confirm the noise sequence that is sampled as before the performance number sudden change, both dwindle the scope in the initial moment of frame after the removal, reduce smart synchronous computing, also improve the latter's accuracy.
Slightly can pass through two synchronous slide window implementation synchronously, shown in figure 10.Two identical windows of size keep constant window gap, slide backward with the starting point of identical step-length from receiving sequence, whenever move and once just calculate the performance number of the interior short sequence of window separately, and two values are compared.If in advance sampled point is abundant, then the front end of receiving sequence must have noise, and also the power ratio of two windows is infinitely close to 1 when meaning beginning; When useful signal progressively got into the tail end of window 2, the power ratio of two windows no longer was 1.Experiment shows that as long as there is the sampled point of 1,2 useful signal to get into window 2, its power can far exceed the performance number of window 1 short-and-medium sequence at once.Because two windows have an overlay segment, so difference power must be the sequence generation in the first two window gaps, therefore a back scope that the window gap is exactly the useful signal original position.
Thick synchronization program based on LabVIEW is shown in figure 11.The length of two sliding windows is 50, and step-length is 10, and the window gap also is 10; When the power ratio of two windows during less than decision value, can confirm that the initial moment of frame has got in the preset range, stop circulation this moment; Sequence behind the intercepting window 1 is further obtained concrete original position.
(2) synchronously smart
In the frame of ERP-OFDM, comprise 10 Short Training periodic sequences, 16 points of each sequence amount to 160 points.Here utilized the periodicity and the autocorrelation of short training sequence to realize smart synchronous computing just.
Noise signal is a random signal, and random signal be owing to can not repeat to produce, thus noise signal can not and other sequences between produce correlation, promptly do not have cross-correlation.So when carrying out related calculation with the noise sequence of 16 points or noise, useful signal mixed sequence and short training sequence, its correlation is inevitable less than the cross correlation value between the short training sequence.Therefore, can progressively remove remaining noise samples point through sliding window equally.The smart synchronization program based on ERP-OFDM frame preamble sign indicating number according to this principle design can be referring to Figure 12.
In the program, the length of two sliding windows all is 144, is the total length of 9 short training sequences; The window gap is 16; Sliding step is 1 in addition.Obviously, when first window has loaded preceding 9 of 10 Short Training periodic sequences just, when another must load the sequence in back 9 cycles, the cross correlation value of two windows was maximum.This moment, the original position of first window was exactly the beginning of complete frame data.Certainly, this is to remove the relative position after the partial noise, and " rough position " of thick output synchronously is the accurate moment that simultaneous operation draws before therefore adding.In the drawings; Two windows begin to slide backward from the thick first place of the sequence as a result of output synchronously; Each slip all will be calculated the cross correlation value of two short sequences in the window; Pass through and previous comparison again, higher value is left in the register, and be recorded in the original position of big correlation correspondence first window constantly in the variable " relative position ".After the several times slip, the value of staying at last in the variable has been exactly smart synchronous result.According to relative position, remove noise residual in the receiving sequence once more, output signal sequence at last.
No matter the frame of ERP-DSSS/CCK form adopts the modulating mode of which kind of speed, and its lead code all is with 11 Barker code expansion, therefore when receiving, will utilize the autocorrelation of Barker code well.The essence of ERP-DSSS/CCK pattern just is based on this principle realization synchronously.
The checking correlation just need be used correlator, and is shown in figure 13 based on the correlator configuration of 11 Barker codes.Receiving sequence is exported 1bit in shift register at every turn, and data and the contraposition of Barker code word in back 11 registers that at every turn are shifted are multiplied each other, and the summation that adds up again is as this output.
When in correlator, aliging from beginning to end with 11 Barker codes through the lead code SYNC field of spread spectrum; Cross-correlation obtains maximum; And this maximum can be whenever at a distance from 11 appearance once; This be because lead code through 11 translations after, 11 bits that formed by character late expansion can the situation that head and the tail align occur once more with Barker code, therefore appear periodically.If contain noise in 11 registers of correlator, the Barker code frequency expansion sequence that is not complementary with fixed codeword in other words conj.or perhaps, the output of correlator crest can not occur.So, when a plurality of all after dates appear continuously in correlation peak, just can judge that first of correlator list entries of first peak value of generation is exactly first Bit data of PPDU.Note its label in receiving sequence, certainly, this also is " relative position " of a relative and thick synchronized result.
Shown in figure 14, this section program is divided into three parts, and pseudo-peak is adjudicated and gone to correlator, peak value.Correlator comprises related operation and asks mould, related operation to utilize loop body to get " thick synchronized result " 11 successively that summation adds up after multiplying each other with 11 Barker code step-by-steps; Delivery is asked energy more subsequently.Observe the waveform of correlation this moment, has the crest of plurality of continuous.Then judge peak value, and be recorded in the sequence number of peak value in the array 4 according to threshold value.Module has been used the sequential organization body " to remove pseudo-peak ", and totally two frames after the program in having accomplished first frame, just can move back one frame here.In first frame, take out adjacent two numbers in the array 4 with the while loop structure at every turn, to ask difference back to remove 11 and get surplusly, remainder is that 0 counter adds 1, circulates otherwise directly get into next time.Behind the each counting of counter, check whether its value has reached C, and C gets 10 here.After the counter meter is reach 10 times, stop circulation, use the value in the array 4 this moment to deduct 10 * 11=110, " relative position " asked.Afterwards need be to counter O reset, so that carry out computing next time, Here it is is provided with the purpose of second frame." relative position " adds that " rough position " is exactly synchronous result.
2. channel estimation and equalization algorithm
(1) channel estimation and equalization of ERP-OFDM pattern
The ERP-OFDM pattern has been taked the way of frequency domain equalization, and through LS (Least Square) algorithm for estimating, promptly least square criterion is released the estimated value of channel, and designed frequency-domain equalizer on this basis, the overall characteristic of adjustment channel.
Because transmitting of ERP-OFDM system is a series of OFDM symbol, and in the moment of signal emission, the channel that transmits each symbol is almost consistent.Therefore can select for use a known OFDM symbol of receiver that the ERP-OFDM system is carried out channel estimating, pass through the symbol of equalizer correction loading section then.This known symbol generally uses the long training sequence of preamble field.
Be X if transmit; Receiving signal is Y; The impact of wireless channel is corresponding to be H;
that goes out for the LS algorithm computation requires it can make the square error that receives the reception signal
behind signal and the estimation model minimum, and promptly
is satisfied
Formula 8
So establish cost function J do
Formula 9
After deriving, have
Formula 10
Shown in figure 15, in program, at first two long sequence symbols in the intercepting synchronization frame divided by 2, are tried to achieve mean value after the addition.Because the frame of ERP-OFDM form has been done IFFT one time when emission, transformed to time domain to the frequency signal of script again, and the LS algorithm is based on that frequency domain derives, so the FFT that will try again returns frequency domain to the signal transformation that receives.According to formula 10, on each subcarrier, done division one time then, the long symbol divided by before the emission can obtain the estimated channel parameter.
After trying to achieve the parameter of channel, other OFDM symbols that receive in the signal carry out equilibrium according to following formula:
Formula 11
Balanced purpose is will be at the signal before receiving terminal restores emission as much as possible, so the signal after the equilibrium
will be infinitely close to emission preface X.During programming, remain and on each subcarrier, do division one time, i.e. step-by-step is divided by the impulse response of channel.
(2) channel estimation and equalization of ERP-DSSS/CCK pattern
The person of being different from, the ERP-DSSS/CCK pattern is in time domain, to adjust the characteristic of channel, the realization equilibrium.Through modeling to wireless channel, can know that multi-path channel can be approximated to be a time-limited FIR filter, the impulse response of finding the solution channel is exactly the coefficient of finding the solution each tap of filter.
Ask the tap coefficient of filter will use sef-adapting filter, so-called sef-adapting filter is meant: the tap coefficient ω of N rank FIR filter
0ω
1..., ω
N-1Can regulate automatically according to the size of evaluated error e (n), make certain cost function minimum.Figure 16 is the schematic diagram of auto-adaptive fir filter.
The present invention is provided with cost function according to the LMS algorithm, promptly requires the output signal of sef-adapting filter and the mean square error between the reference signal minimum.So cost function J is defined as:
J=E{|e (n) |
2}=E{|d (n)-y (n) |
2 Formula 12
According to steepest descent method, the iterative formula of obtaining the tap coefficient of LMS adaptive algorithm does again
ω (n)=ω (n-1)+2 μ (n) e (n) x (n) formula 13
μ in the formula (n) is the constant sequence of control evaluated error e (n) convergence rate and filter stability, is called convergence factor.
For the selection of convergence factor, become the thought of convergence factor when that the present invention has adopted is a kind of " search earlier, back convergence ":
Formula 14
μ in the formula
0Be a fixing convergence parameter, τ express time search constant.
LabVIEW program according to above principle is write can be referring to shown in Figure 17.Programming be one 11 rank sef-adapting filter, have two nested loop bodies.Outer circulation is got 11 bit data at every turn and is sent into filter, and interior circulation then is used to calculate the output y (n) of each iteration postfilter.Deduct output valve with reference signal and be evaluated error, be recorded in the array " error ", and can watch the convergence situation of error through oscilloscope.Realize upgrading through the relevant parameter of stack when the whole zero setting of coefficient of filter when initial, each iteration, numerical value is kept in the shift register, and supplying next time, circulation time uses.Here the convergence factor parameter μ that chooses
0=0.1, time search constant τ is 20.
In addition, LabVIEW provides the driver of various modular instruments, through cooperatively interacting of multiple driver, and the correct setting of instrument parameter, just can control controller and call radio-frequency module work.The general flow figure of modular instrument work is shown in figure 18.Equally, the configuration of software also should be followed this principle.
1. transmitter driver
Shown in figure 19, table 6 is the explanations to the driver module of using in the program, mainly the input parameter of each VI and the function of realization is described.
Module in the table 6 transmitter driver
Associative list 6 can find out that with Figure 19 driver module is interconnective through instrument handle (instrument han dle), is comprising the input parameter of all VI of front in the handle.The instrument operating procedure is following:
(1) program is at first used " niRFSA Initialize " VI initialization instrument." resource name " refers to the instrument title that needs startup, and the input here is the logical resource name of PXIe-5673E, i.e. " RFSG ";
(2) configure instrument parameter then.Set required centre frequency, transmitting power and the I/Q speed of radiofrequency signal through three parameter configuration module, and generate required initial phase, sampling interval and the baseband complex signal etc. of analog waveform.802.11g the radiofrequency signal of system works in the ISM band of 2.4GHz, promptly keeps the licensed band of exempting from of giving industry (Industrial), scientific research (Scientific) or Medical Devices (Medical) use.The concrete channel distribution and the visible table 7 of centre frequency of each channel.
The channel plan of table 7 2.4GHz frequency band
(3) then write data in the instrument.Need to load and start I/Q modulation module (niRFSG Configure IQ Enabled), resampling module (rfsg_Resample and Write) and electromagnetic wave emission (niRFSG Initiate) and stop module (niRFSG Abort); Make the electromagnetic wave can repeat its transmission through a loop structure body here; Certainly, this is as far as same Frame;
(4) need close instrument (niRFSG close) at last, or when data exception, show run-time error.
2. receiver driver
The hardware module of receiver so the driver somewhat complicated is mainly reflected in the configuration of parameter, is seen Figure 20 because will accomplish work such as signal reception, data sampling, frequency spectrum demonstration.Table 8 is functional descriptions of driver module in the program.
The workflow of receiver hardware is following:
(1) instrument initialization.Receiving unit uses " niRFSA Initial " to start instrument, and is corresponding, the logical resource name " RFSA " that its " resource name " should be PXIe-5663, and generate handle with this;
(2) configure instrument parameter.The parameter in this stage mainly is divided into three parts, is respectively the trigger parameter of signal reception, the sampling parameter and the desired parameter of demonstration frequency spectrum of A/D conversion.
1) trigger parameter of signal reception
At first be the receive mode of instrument, PXIe-5663 provides two kinds of receive modes, is respectively IQ and Spectrum.If will carry out demodulation analysis to the received signal, then need instrument is set at the former; If will carry out power spectrumanalysis to the received signal, then need instrument is set at the Spectrum mode.Native system mainly carries out Base-Band Processing to the received signal, thereby should be IQ with this parameter setting; Be maximum power and the centre frequency that expectation receives signal then.Centre frequency need be identical with being provided with of transmitter, import in the interface of receiver through " IQ carrier Frequency " input frame, and maximum power is the parameter of reserving in order to guarantee to receive the integrality of signal, is set at here-45dB; Need select to trigger the type of subsystem at last; This parameter can be divided into three kinds of digital edge-triggered subsystem, software trigger subsystem and power trigger subsystems again; Here adopted power trigger subsystem pattern; Promptly when begin to receive signal and sampling, minimum noise samples point is stored in the memory space of instrument guaranteeing, thereby has saved valuable internal memory according to power trigger levels decision instrument.When the performance number that receives signal begins to rise and arrives trigger value, then triggering system work.
2) sampling parameter of A/D conversion
Comprise to received signal sample frequency and sampling number etc.The IQ speed of ERP-OFDM signal is 20Mbps, the ERP-DSSS/CCK signal then be 11Mbps, according to the nyquist sampling law, sample rate should be the twice of speed at least, so here according to the system pattern that do not pass through, corresponding sample rate is set respectively; Sampling number is represented the number of the sampled point that instrument provides altogether, and this wherein also comprises the data of noise signal naturally.Because which kind of frame format no matter, the maximum of its PSDU part can not surpass 4095 eight hytes, so as long as the assurance sampling number greater than 40000.In addition, also have a reference clock to choose, this parameter has four kinds of selections: Onboard clock, and RedIn, ClkIn and PXI_Clk, these different selections are corresponding to the connection method of different composition PXIe-5663 three sub-module input ports.Native system is set at PXI_Clk with this.
3) show the frequency spectrum desired parameters
In order on receiver, to observe the spectrogram that receives signal, need use following parameter: carrier frequency, demonstration bandwidth, spectrum mode and resolution bandwidth.That mentions in carrier frequency and the preceding text is identical, sets according to table 5.2; Show the spectrum width that bandwidth is meant on receiver to be demonstrated, if be provided with little, then frequency spectrum show imperfect, otherwise signal spectrum relative narrower then can't observe and analyze it.Here according to the bandwidth of ERP-OFDM signal be 20MHz, ERP-DSSS/CCK signal be 22MHz and be set at 40MHz; Average mode has been chosen RMS (root mean square) pattern, i.e. signal effective value representation; Resolution bandwidth is an important parameter of spectrum analyzer, in order to characterizing the minimum frequency space that instrument can identify two signals, if the sigtnal interval less than the value of resolution bandwidth, then instrument can't be differentiated out.So to two tight two adjacent signals, its resolving power depends on the resolution bandwidth of spectrum analyzer.Here be set to 150kHz.
4) transmit parameter
Calling " niRFSA commit " writes above-mentioned parameter in the internal memory of PXIe-5663.
(3) data read and module loading.Opposite with transmitting terminal, the hardware module of receiver need be taken out the I/Q data read in the instrument, so that carry out follow-up operation, as sampling and consigning to processing such as band receiver of base carries out synchronously, equilibrium, finally obtains initial PSDU information.Simultaneously, signal can also be observed its spectrogram at receiving terminal after handling through frequency domain, uses the reason of loop structure the same with transmitter here, is in order to observe waveform times without number.Specifically inform instrument parameter configuration completion, can begin to receive data through " niRFSA Initate " VI; The data that will be stored on the PXIe-5663 internal memory after will sampling through " niRFSA Fetch IQ " module are transferred in the internal memory of controller, and timeout parameter value wherein is set at-1, represent that this operation will be waited for up to trigger event always to produce; Re-use " niRFSA Abort " VI and stop initialized I/Q data read operation by " niRFSA Initate ".
(4) merge the error message of band receiver of base and driver at last, close instrument through " niRFSA close " VI.
The module of table 8 receiver driver
Combine accompanying drawing that execution mode of the present invention has been done detailed description above, but the present invention is not limited to above-mentioned execution mode, in the ken that those of ordinary skills possessed, can also under the prerequisite that does not break away from aim of the present invention, makes various variations.