Summary of the invention
(1) technical problem that will solve
In view of this, main purpose of the present invention is to provide a kind of method that reduces the ofdm system power consumption of synchronous module, with in the safeguards system net synchronization capability, reduces the work power consumption of system synchronization module.
(2) technical scheme
For achieving the above object, the invention provides a kind of method that reduces the ofdm system power consumption of synchronous module, this method comprises:
Step 101: on time-domain, the sign bit of the synchronizing sequence that receives and the sign bit of local known array are carried out the frequency sweep related operation, obtain the estimated value of integer frequency offset;
Step 102: on the time-domain sign bit of the synchronizing sequence behind the correcting frequency deviation that receives and the sign bit of local known array being carried out related operation, obtain the exact position that FFT windows.
In the such scheme, described step 101 comprises:
The sign bit of the synchronizing sequence that A1, calculating receive;
The sign bit of A2, the local known array of calculating;
A3, the sign bit of the synchronizing sequence that receives and the sign bit of local known array are carried out the frequency sweep related operation.
In the such scheme, described steps A 1 comprises:
Suppose that ofdm system slightly finishes synchronously, and finish the fractional part of frequency offset compensation that find transmitting terminal to send the approximate location of known array, at this moment, the sign bit of the synchronizing sequence of reception can be expressed as:
Wherein, z (k) expression transmitting terminal sends signal; Δ f
IntBe the integer frequency deviator, be positive negative integer; f
CarrierBe ofdm system intercarrier distance; T
sBe the ofdm system sampling period.
In the such scheme, described steps A 2 comprises:
Suppose that ofdm system comprises local known array c (k), then the frequency sweep sequence can be expressed as:
In the present invention, c (k) gets sign bit, and following formula can be expressed as:
Wherein sign () can be expressed as
sign(r)=real_r+j□imag_r
In the such scheme, described steps A 3 comprises:
Utilization receives the sign bit of data, and can get according to maximum likelihood principle:
I ∈ [k wherein
Max, k
Max], k
MaxBe the patient maximum integer frequency deviation value of system, make formula (5) at [k like this
Max, k
Max] to reach peaked i in the scope be exactly integer frequency bias.
In the such scheme, among described steps A 2~A3 the signal of the known synchronizing sequence in this locality and reception is got sign bit and carry out the frequency sweep related operation, needed hardware comprises:
2048 * 2 RAM are used for storing real part and the imaginary part of the synchronizing signal of receiving, wherein each 1 bit wide of I, Q and being stored in the same address in hardware designs;
2048 * 2 ROM are used for storing known synchronizing sequence real part and imaginary part, wherein each 1 bit wide of I, Q and being stored in the same address in hardware designs;
Value adopts the mode of look-up table in this modular design, store with ROM
M=-k
Max..k
Max, because
Value smaller, so need more bit wide to store, real part and imaginary part are represented with 22bit respectively, k in the design
Max=30, therefore need DROM64 * 44; And
Two 1 * 1 multiplier and two s' 22 * 22 multiplier calculates real part and imaginary part respectively.
In the such scheme, described step 102 comprises:
B1, finish synchronously smart by the relevant algorithm of sliding window;
B2, passing threshold judgement in the relevant result of sliding window, the exact position of finding FFT to window;
B3, utilize the sign bit computing to finish sliding window related operation.
In the such scheme, described step B1 comprises: smart synchronous algorithm will be by seeking first directly the position of multipath, the exact position of finding FFT to window, this algorithm are that finish data behind the compensate of frequency deviation and the local known array that utilize to receive carry out matched filtering and finish; Algorithm is seen formula (9):
L is the length that transmitting terminal sends known array in the formula, and r is the signal after the compensating through carrier wave frequency deviation of receiving, and c (k) is local known signal, and l is the hunting zone; Through type (9) passing threshold in the hunting zone judges that the n value that finds is thick synchronous deviate.
In the such scheme, described step B2 comprises: the hunting zone is decided to be [l, l], the window front and back l subcarrier of position of the thick FFT that determines synchronously of i.e. search, what all were tried to achieve in this scope adding up and being averaging, draw mean value mean after, judge as threshold value with α * mean, α obtains by ofdm system emulation, the exact position that can find FFT to window.
In the such scheme, on the time-domain sign bit of the synchronizing sequence behind the correcting frequency deviation that receives and the sign bit of local known array being carried out related operation, obtain the exact position that FFT windows among described step B1~B3, needed hardware comprises:
2272 * 2 RAM, after being used for finishing little several times and integer frequency offset compensation, to use real part and the imaginary part of 2248 * 26 data in thick front and back, the synchronizing sequence positions 100 subcarrier scopes of estimating synchronously of storage, I, each 1 bit wide of Q symbolization position in hardware designs, this sequence is the computing of window and local known array perfect (10) with 2048 after the storage;
224 * 24 RAM are used for the value of 201 Corr (n) after memory-type (10) computing, average and judge smart sync bit being used for, and correlated results is represented with 24 bit wides in hardware designs;
A ROM, DROM2048 * 2 are used for storing the sign bit of local known synchronizing sequence, each 1 bit wide of local known synchronizing sequence I, Q in the hardware designs; And
Two 1 * 1 multiplier is used for finishing complex multiplication operation.
In the such scheme, the hardware that described step 101 and step 102 need jointly comprises:
One of 2272 * 2 single port RAM;
One of 2048 * 2 single port ROM;
One of 64 * 44 single port ROM;
One of 224 * 24 single port RAM;
Two of 1 * 1 multipliers; And
Two of 22 * 22 multipliers.
(3) beneficial effect
From technique scheme as can be seen, the present invention has following beneficial effect:
1, the method for this reduction ofdm system power consumption of synchronous module provided by the invention, when not reducing the system synchronization performance, significantly reduced the bit wide of memory and multiplier in the related operation, thereby reduced the power consumption of system synchronization, be highly suitable for the design of hand-held mobile terminal device chip.
2, the method for this reduction ofdm system power consumption of synchronous module provided by the invention is applied to can effectively reduce the power consumption of synchronous module of this system's base band demodulator in China Mobile multimedia broadcasting (CMMB) system.
3, the method for this reduction ofdm system power consumption of synchronous module provided by the invention, the method for sign bit computing also is applicable to the low power dissipation design of other each key modules of OFDM band receiver of base.
Embodiment
For making the purpose, technical solutions and advantages of the present invention clearer, below in conjunction with specific embodiment, and with reference to accompanying drawing, the present invention is described in more detail.
The method of reduction ofdm system power consumption of synchronous module provided by the invention, be based on that the synchronization module low power consumption design method of China Mobile multimedia broadcasting (CMMB) system and emulation proposes, by to the emulation of sign bit related operation on the time-domain with to the extraction of formula (8) related operation common factor, can optimize the circuit design of synchronization module, reduce the bit wide of memory and multiplier in the computing, thereby effectively reduce the power consumption of system synchronization module.
As shown in Figure 2, Fig. 2 is the method realization flow figure of reduction ofdm system power consumption of synchronous module provided by the invention, and this method comprises:
Step 101: on time-domain, the sign bit of the synchronizing sequence that receives and the sign bit of local known array are carried out the frequency sweep related operation, obtain the estimated value of integer frequency offset;
Step 102: on the time-domain sign bit of the synchronizing sequence behind the correcting frequency deviation that receives and the sign bit of local known array being carried out related operation, obtain the exact position that FFT windows.
Above-mentioned steps 101 specifically comprises:
The sign bit of the synchronizing sequence that A1, calculating receive.Suppose that ofdm system slightly finishes synchronously, and finish the fractional part of frequency offset compensation that find transmitting terminal to send the approximate location of known array, at this moment, the sign bit of the synchronizing sequence of reception can be expressed as:
Wherein, z (k) expression transmitting terminal sends signal; Δ f
IntBe the integer frequency deviator, be positive negative integer; f
CarrierBe ofdm system intercarrier distance; T
sBe the ofdm system sampling period.
The sign bit of A2, the local known array of calculating.Suppose that ofdm system comprises local known array c (k), then the frequency sweep sequence can be expressed as:
In the present invention, c (k) gets sign bit, and following formula can be expressed as:
Wherein sign () can be expressed as
sign(r)=real_r+j□imag_r
A3, the sign bit of the synchronizing sequence that receives and the sign bit of local known array are carried out the frequency sweep related operation.Utilization receives the sign bit of data, and can get according to maximum likelihood principle:
I ∈ [k wherein
Max, k
Max], k
MaxBe the patient maximum integer frequency deviation value of system.
Like this, make formula (5) at [k
Max, k
Max] to reach peaked i in the scope be exactly integer frequency bias.
The integer frequency offset of CMMB system estimates just to be based on the integer frequency deviation estimation algorithm of synchronizing sequence frequency sweep, and this algorithm is at k
Max=100 o'clock simulation result is seen shown in Figure 3, can be clear that in frequency deviation be under the 40ppm, after finishing fractional part of frequency offset estimation and compensation, carries out the integer frequency deviation estimation algorithm of synchronizing sequence frequency sweep, and there is obvious peak value at the place at subcarrier-43.
A4, based on the specific design of CMMB band receiver of base, the hardware of this method is realized.
The algorithm that steps A 1-A3 of the present invention introduces is applied in the design of CMMB system.The beginning of each time slot of CMMB system comprises a beacon, and beacon is made up of structure shown in Figure 4.
Two synchronizing signals are that identical length is 2048 sequence.Integer frequency offset is estimated to finish the main application of synchronized sequence of its algorithm in time domain in the CMMB system:
k=0,1,...,2047,m=-k
max,..k
max (6)
k
MaxBe the integer frequency offset swept frequency range, sync (k) is synchronizing sequence;
But in actual hardware designs, need to calculate
Two variable m and k are arranged in this formula, and this will bring very big design complexities to hardware designs, and especially k changes in 0-2048, so whenever finishes the calculating of a m, need to calculate
Totally 2048 times, need expend a large amount of power consumptions and time, when the present invention designs this hardware module, computing formula is opened, suppose that at first m fixes:
Can find that like this index unique in the formula is
And no longer changed along with the variation of k, so just can use the method for look-up table to calculate
The core concept of this modular design that Here it is.
According to the method for steps A 2-A3, the signal of local known synchronizing sequence and reception is got sign bit and is participated in computing, can draw needed hardware like this from formula:
1. 2048 * 2 RAM are used for storing the real part of the synchronizing signal of receiving and imaginary part (in hardware designs each 1 bit wide of I, Q and be stored in the same address);
2. 2048 * 2 ROM are used for storing known synchronizing sequence real part and imaginary part (in hardware designs each 1 bit wide of I, Q and be stored in the same address);
3.
Value adopts the mode of look-up table in this modular design, store with ROM
M=-k
Max..k
Max, because
Value smaller, so need more bit wide to store, real part and imaginary part are represented with 22bit respectively, k in this modular design
MaxTherefore=30, needing DROM64 * 44, (ROM address, integrated circuit technology storehouse can only be 2
n, this DROM can extension storage to-32~32 swept frequency ranges);
4. also need the expense multiplier in addition: two 1 * 1 multiplier and two s' 22 * 22 multiplier calculates real part and imaginary part respectively;
Hardware configuration as shown in Figure 5.
In hardware designs shown in Figure 5, insert two registers, carry out two-stage pipeline structure, i.e. 1 * 1 complex multiplier and 22 * 22 complex multiplier flowing water computings, thus accelerate the speed of integer frequency offset estimation and reduce the hardware power consumption;
Above-mentioned steps 102 specifically comprises:
B1, finish synchronously smart by the relevant algorithm of sliding window.The system that finished thick synchronously after, find the approximate location that the FFT data begin after, need carry out the smart synchronized algorithm of symbol, with the exact position of finding FFT to window.
Smart synchronous algorithm will be by seeking first directly the position of multipath, the exact position of finding FFT to window.This algorithm is that finish data behind the compensate of frequency deviation and the local known array that utilize to receive carry out matched filtering and finish.
Algorithm is seen formula (9):
L is the length that transmitting terminal sends known array in the formula, and r is the signal after the compensating through carrier wave frequency deviation of receiving, and c (k) is local known signal, and l is the hunting zone.
Through type (9) passing threshold in the hunting zone judges that the n value that finds is thick synchronous deviate.
B2, passing threshold judgement in the relevant result of sliding window, the exact position of finding FFT to window.When algorithm of the present invention and hardware are realized the hunting zone is decided to be [l, l], the window front and back l subcarrier of position of the thick FFT that determines synchronously of i.e. search, what all were tried to achieve in this scope adding up and being averaging, after drawing mean value mean, judge the exact position that (α obtains by ofdm system emulation) can find FFT to window with α * mean as threshold value.
B3, utilize the sign bit computing to finish sliding window related operation.Step B1 Chinese style of the present invention (9) can be carried out computing in the symbolization position, and the matched filter hardware configuration as shown in Figure 6.
Algorithm is seen formula (10):
B4, in the practical application of CMMB base band demodulating chip synchronization module.The algorithm that step B1-B3 of the present invention introduces is applied in the design of CMMB system.Local known array among Fig. 6 still adopts the synchronizing sequence in Fig. 4 beacon.If l=100, site error is 2 after supposing slightly to finish synchronously, and namely sync bit is offset two subcarriers of physical location, and simulated environment is the exemplary dynamic multipath channel, utilizes the matched filtering algorithm can obtain peak value shown in Figure 7.
From Fig. 7, can clearly be seen that the relation of position between multipath, when this paper algorithm and hardware are realized the hunting zone is decided to be [100,100], window 100 subcarriers in front and back (l=100) of position of the thick FFT that determines synchronously of i.e. search, what all were tried to achieve in this scope adding up and being averaging, draw mean value mean after, judge as threshold value with 64 * mean, can find out multipath first footpath is the peak value at 2 places, position, thus the exact position of finding FFT to window.
On hardware designs, because this module and integer frequency deviation estimation algorithm all need to carry out related calculation with local known array, so in hardware designs, can share a hardware module, two block RAMs are arranged in this module:
1. 2272 * 2 RAM:
After finishing little several times and integer frequency offset compensation, to use real part and the imaginary part (I, each 1 bit wide of Q symbolization position in hardware designs) of 2248 * 26 data in thick front and back, the synchronizing sequence positions 100 subcarrier scopes of estimating synchronously of storage, this sequence is the computing of window and local known array perfect (10) with 2048 after the storage.
2. 224 * 24 RAM:
This RAM is used for the value of 201 Corr (n) after memory-type (10) computing, averages and judges smart sync bit (correlated results is represented with 24 bit wides in hardware designs) being used for.
Also have a ROM in addition in the module, DROM2048 * 2 are used for storing the sign bit (each 1 bit wide of local known synchronizing sequence I, Q in the hardware designs) of local known synchronizing sequence; Also need two 1 * 1 multiplier to finish complex multiplication operation.
In addition, step 101 and step 102 comprise the main hardware expense altogether:
One of 2272 * 2 single port RAM; One of 2048 * 2 single port ROM; One of 64 * 44 single port ROM; One of 224 * 24 single port RAM; Two of 1 * 1 multipliers; Two of 22 * 22 multipliers.
It is comprehensive that Design Compiler is carried out in employing 0.13 μ m SMIC storehouse after the RTL design of intact back, and work clock is 80MHz, and synthesis result is as shown in table 1.The sign bit computing will reduce about half than the power consumption of traditional bit wide computing (receiving data, each 13 bit wide of local known array I, Q).
RASP2272X2M16 |
0.9460mW |
DROM64X44M8 |
3.631e-5mW |
DROM2048X2M8 |
0.7514mW |
Multss_short (multiplier, number: 2, figure place: 1 * 1) |
5.689e-03×2=1.1378e-02mW |
Multss (multiplier, number: 2, figure place: 22 * 22) |
0.0752×2=0.1504mW |
RASP224X24M16 |
4.169e-4mW |
Cell internal power |
3.2054mW |
Net switching power |
23.3920uW |
Table 1 sign bit computing DC synthesis result
In addition, step 101 also comprises: the compensate of frequency deviation circuit, this circuit is mainly finished by cordic algorithm, finishes in CMMB system design adopted 15 stage pipeline structure.
The method of the described reduction ofdm system power consumption of synchronous module based on the CMMB system of step 101 and step 102 in addition, the bit wide size of the memory in the described hardware designs and multiplier can change (address size and data I, Q represent bit wide) according to side circuit design and integrated circuit technology storehouse.
Above-described specific embodiment; purpose of the present invention, technical scheme and beneficial effect are further described; institute is understood that; the above only is specific embodiments of the invention; be not limited to the present invention; within the spirit and principles in the present invention all, any modification of making, be equal to replacement, improvement etc., all should be included within protection scope of the present invention.