CN1553601A - Method for eliminating self generated interference in time division duplex system - Google Patents

Abstract

This invention discloses a method for overcoming the self interference in the time division duplexing system. It utilizes the sending and receiving signal known by common consent to calculate channel impulse response, then remove the new impulse response from the receiving signal and the convolution from sending signal. The simplest method is in term of assuming the channel impulse response is known to build a system of equations describing how the reflected signals compose, make determination accordingly. Then Utilizing the channel impulse response processed and the above system of equations to calculate the signal distortion. If existing pretty strong up stream synchronization code word signal, then rebuilding the up stream synchronization code word tested, then for removing infection of the strong up stream synchronization code word to the mean of the next intra frame, needs to feedback the rebuilt up stream synchronization code word to the second stage disturbance algorithm of the next frame signal.

Description

Eliminate the method for spontaneous interference in the tdd systems

1, technical field

The present invention relates to a kind of method of improving time division duplex (TDD) system communication quality, more specifically to the method that reflects caused spontaneous interference in a kind of TDD of the elimination system by downstream signal.

2, background technology

The upward signal and the downstream signal of typical case TDD system all transmit by a carrier wave.That is to say that related equipment (as base station in the mobile radio system and travelling carriage) must switch between receiving and sending.Such as upward signal can only just can be removed to receive in the base station after having transmitted downstream signal.

Because rf wave can reflect by surrounding environment, therefore such situation can take place, promptly be reflected and during through bigger propagation delay, turn back to the antenna that is receiving the transmitting base station of upward signal at that time when a downstream signal.In receiver, this downstream signal that is reflected can reduce the performance of receiver greatly, and blocks the communication of upward signal.

Up to now, people are just considering the TDD system such as the TDD (HCR TDD) that is defined as the high chip propagation rate of standard by the 3G (Third Generation) Moblie Committee of Experts (3GPP) and Digital European Cordless Telephone/digitally enhanced cordless telephone (DECT) is used for indoor environment and Microcell.The height that this means the base station can be lower.Adopt the coverage of base station of mobile radio system of this scheme very little, and its signal delay (signal is from passing the propagation time to reverberation and reflected back transmitting base station) is also very little.This effect is in most of the cases all set up concerning mobile radio station.Therefore, this effect is little of being actually negligible, and system does not need it is handled.

Now, Zhong Guo TD-SCDMA (time division multiplexing S-CDMA) is just in the China Wireless Telecommunication Standar committee (CWTS), International Telecommunication Association (ITU) be confirmed as standard in the mobile communication Committee of Experts (3GPP) for the third time.What TD-SCDMA was different with other TDD system is that it also is considered to be particularly suitable for macrocell and disposes and studied by this direction.That is to say to be applicable to the very large sub-district of some scopes, and because above-mentioned effect, signal bigger delay can occur in these sub-districts.Detailed description to the TD-SCDMA system is found in following TD-SCDMA standard: CWTS TS C101v3.0.0, TS C102 v3.0.0, TS C103 2.2.0, TS C104 v3.0.1, frame structure in the TS C105 v3.0.0.TD-SCDMA mobile communication system is compatible mutually with GSM's, and they all utilize training sequence (midamble) to do pulse detection.The duration of each wireless sub-frame is 5ms, is made up of 7 duration business time-slot that is 675 μ s and 3 special time slot: DwPTS (descending pilot frequency time slot), GP (protection time slot) and UpPTS (uplink pilot time slot).

Fig. 1 shows TDD from the influence of disturbing the TD-SCDMA system.Label (1) expression base station A, (2) expression base station B, (3) expression reflector.As seen from the figure, because propagating, signal needs the time, downstream signal TX is not overlapping aspect the time relationship with the signal RX that same base station is received, initial TX signal will could be received by same base station after after a while, when base station A received the downstream signal of final stage, that frame downstream signal at that time is battery has fired in addition.In TD-SCDMA, all base stations are synchronous working all, but has influence each other after all in each base station.When system works, base station A launches a downstream signal (TX), and this signal is reflected and is received (RX) by same base station subsequently.Because the base station still transmits continuing at this moment, can not received signal, so the signal section of dash area can not be received (but signal still can arrive this base station) by this base station.Have only the non-shaded portion of reflected signal (RX) really to be received by the base station subsequently.

Because the TD-SCDMA network all is synchronous usually, so the reflected signal of the TX signal of base station A also can receive by base station B, so base station B also can be affected.Because its distance and radio propagation conditions, so the signal amplitude that receives less than base station A place usually of the amplitude of the received reflected signal of base station B.

In general, above-mentioned reflection process relates to a more than reflector, and it is unusual complicated that this makes that follow-up signal processing work becomes, and needs more suitably algorithm and more powerful processor.

In TD-SCDMA, only the conversion of the upward signal (UL) of above-mentioned influence to downstream signal (DL) once may appear in existence, promptly appears at descending pilot frequency time slot (DwPTS) afterwards.Fig. 2 shows this situation.

Specifically, Fig. 2 shows near the time slot descending in a TD-SCDMA system/up switching point.After descending pilot time slot, the base station receives uplink pilot time slot.This uplink pilot time slot is utilized in random access procedure, and according to the simultaneous bias of transmitting mobile stations, can follow descending pilot frequency time slot closely and arrive this base station.Above-mentioned effect can seriously influence the reception of uplink pilot time slot, thereby reduces the performance of whole system.In this case, distortion is mainly from descending pilot frequency time slot.Reflected signal is to arrive in the protection interval (G) of TD-SCDMA system.

This problem is found in the main on-the-spot test (master field trail) at TD-SCDMA, and has measured at base station A and base station B.Fig. 3 and Fig. 4 provide measurement result respectively, wherein:

Fig. 3 shows transmitting of the descending pilot frequency time slot that receives in A place, base station,

Fig. 4 shows the reflected signal (this moment, transmitter was closed) of the descending pilot frequency time slot that receives in B place, base station.In Fig. 3, used wave spectrum analyzer to measure.In the middle of measuring, for making the instrument operate as normal, the employing measure TX signal of decaying, and amplify the RX signal.In once measuring, can see original TX signal (downlink business time slot and descending pilot frequency time slot) and RX signal (reflected signal of descending pilot frequency time slot) like this.Can see that from measurement result the reflected signal of descending pilot frequency time slot can be 1000 times (30 decibels) of ambient noise intensity.Can determine to occur the position of distortion near the reflector of base station.Reflector away from the base station is numerous, but these reflectors are equally also away from the base station that transmits.In this embodiment, the distance of the quantity of reflector and itself and base station is proportional, and power that is received simultaneously and 4 powers of above-mentioned distance are inversely proportional to.Correspondingly, reflection power reduces manyly more, and the time lag between descending pilot frequency time slot and the time of reception is also big more.The length of distorted signals is approximately 100 microseconds, and this is equivalent to the distance of reflector from the base station is 15 kilometers.Also the maximum magnitude that this base station can cover about 15 kilometers.

Measurement result among the B of base station has been shown among Fig. 4.In current the measurement, the antenna of wave spectrum analyzer and this base station is coupled together.Have only when the base station is in the position of " pass " and just can receive.Can see the reflected signal of downlink business time slot and descending pilot frequency time slot this moment simultaneously.The waveform of reflected signal is different with intensity.Its reason is: the ripple that is reflected not only will pass to reflector, also will return base station B along different paths.At base station A, the return path of reflected wave is easy to determine that in fact it is exactly the path of signal biography to reflector.This just why the intensity of the reflected signal shown in Fig. 3 can descend reason faster.

The influence of this effect is very serious, to the feasibility of entail dangers to TD-SCDMA in the macrocell deployment.

Up to now, also have no talent and when implementing the TDD system, consider this effect, thereby at present also without any the solution at this effect.

Similar problem is also arranged in Radar Technology: radar also is to send signal by radar station earlier in use, and radar station receives reflected signal again in follow-up time then.

In multiple radar application, have only moving target to be only that it is interested, and building, mountain peak etc. only can produce fixing reflection.In this case, reflected signal is averaged, from the signal that receives, deduct then, so just can see the signal of expression moving target.This process is called as the fixed target null method.

By above process, varying signal not in the time of can eliminating, and receive only that those are sent by same reflector, reflected signal the time become part.

3, summary of the invention

The present invention solves this problem by eliminating reflected signal, makes to receive the signal that is generated by different transmitters.

In a word, the present invention adopts following technical scheme, consider to be sent out the characteristic of new-type and modulation of the signal of signal and related mobile radio channel, go to discern the reflected signal or the part that are sent out signal and be sent out signal, and from received signal, they are eliminated.

This can finish in the following way: utilize generally acknowledged known transmission signal and received signal to remove the calculating channel impulse response, eliminate (deducting) then newly to impulse response and the convolution that sends signal from received signal.

When only coming calculating channel impulse response and handling identical received signal with a frame, then need to determine channel impulse response which partly belong to reflected signal, which part has nothing to do with reflection.For accomplishing this point, the simplest mode is to suppose to set up the equation group how a relevant reflected signal constitutes under the known condition of channel impulse response.Find the solution this equation group according to the channel impulse response that sets, then less channel impulse response tap is made as zero, thereby makes judgement.Then, utilize channel impulse response and above-mentioned equation group, calculate the distortion of signal through reprocessing.

When the channel impulse response of each frame changes when little, then can the channel impulse response of each frame be averaged, the channel pulse of frame rings before perhaps adopting, and comes the reconstruct reflected signal.

When relevant with the major part of the reflected signal that is received that part of between frame and frame, the variation when little in the transmission signals, and channel impulse response also changes between frame and frame when little, the reflected signal that is received is averaged, and from instantaneous received signal, deduct the received signal of institute's average reflection, just be enough to eliminate the influence of reflected signal.Become part during non-in this way can erasure signal, and in the signal that is received the time become part and do not have remarkable influence.

In the TD-SCDMA system, descending pilot frequency time slot is the principal element that produces reflected signal.The identical modulation of 64 chips processes of descending pilot frequency time slot, and have 4 kinds of different phase places.With regard to basic waveform, channel impulse response remains unchanged between frame and frame.So here the algorithm of Cai Yonging is changed the reception reflected signal after modulating, and specifically, is according to phase modulation signal to be carried out the counter-rotating of phase place, and then resulting signal is averaged.

For the recipient, must modulate again the signal after average, and from instantaneous received signal, deduct this average signal, to detect uplink pilot time slot.

Yet, if there is very strong uplink synchronous code word (uplink synchronization code, be called for short SYNC1) signal, this will cause tangible influence to mean value, makes the existence that will detect SYNC1 in not having the next frame signal of uplink synchronous code word mistakenly.Therefore, under the condition of the influence of considering strong uplink synchronous code word signal, must revise interference cancellation algorithm described above.Thus, interference elimination method described above can be taken as first order interference eliminated, also will consider second level interference compensation method simultaneously.

The advantage of this algorithm of the present invention is that its complexity is very low, and it can implement on a position, and other algorithms remain unchanged.

Another advantage of this algorithm of the present invention is, as long as according to model identical, the phase place of the descending pilot frequency time slot signal in all relevant base stations modulated, and then this algorithm just can be eliminated the reflected signal of adjacent base station.

In following embodiment, will carry out detailed narration to this algorithm.

In each modification of the method for the invention,, can improve the performance of receiver by eliminating reflected signal.

The creative step of the present invention is embodied in:

(1) make full use of the characteristics of related transmission signals and mobile radio channel, eliminate above-mentioned TDD from disturbing,

(2) at the characteristics of TD-SCDMA and adopt and simplify this general-purpose algorithm,

(3) with the fixed target null method correct adopted in the Radar Technology and with in it TDD communication system of migrating.

4, description of drawings

Each accompanying drawing that the present invention is relevant with body part shows background of the present invention and invention scheme, wherein:

The TDD that Fig. 1 shows in the TD-SCDMA system disturbs certainly.

Fig. 2 shows near each time slot the descending and up switching point among the TD-SCDMA.

Fig. 3 shows the reflected signal of the descending pilot frequency time slot among the A of base station.

Fig. 4 shows the reflected signal (this moment, reflector was closed) of the descending pilot frequency time slot among the B of base station.

Fig. 5 is illustrated in the TD-SCDMA on-the-spot test, the result who disturbs elimination certainly of descending pilot frequency time slot.

Fig. 6 is the block diagram of subframe structure involved in the present invention.

Provide the block diagram of the two-stage interference cancellation algorithm of implementing in the present invention among Fig. 7.

The reprocessing principle that the uplink synchronous code word detects has been described among Fig. 8.

Provide among Fig. 9 and be used to illustrate under the condition that comprises SYNC1 in the current received frame simulation result of two-stage interference cancellation algorithm effect.

Provide among Figure 10 and be used to illustrate under the condition that comprises SYNC1 in the last received frame simulation result of two-stage interference cancellation algorithm effect.

5, embodiment

With the TD-SCDMA system is example, and following algorithm has been used to on-site test system.Digital signal processor (DSP) internal measurement in the base station obtains as following result.These results can be used for the correctness of check algorithm.

Fig. 5 shows and adopts technical scheme of the present invention, the relative power schematic diagram of protection time slot and uplink pilot time slot in the TD-SCDMA on-the-spot test, and disturbing certainly that descending pilot frequency time slot produces as seen from the figure is eliminated.Transverse axis among the figure is represented chip position (is unit with the microsecond), and the longitudinal axis is relative power (unit is decibel); Black block curve is represented initial received power, the received power after darker block curve is represented to handle.As seen from the figure, initially receive and be positioned at before about 100 microseconds the signal fluctuation that is reflected by descending pilot frequency time slot signal and causes in the curve and eliminated effectively.

Another problem that produces from solution of the present invention is the complexity influence of algorithm used in the present invention.Following content of the present invention is to being reflected concise and to the point narration and the complexity analyzing that the algorithm that recurrence eliminates is carried out in caused interference by descending pilot frequency time slot, and the result shows that the complexity of the used algorithm of the present invention compares with joint detection algorithm, is inappreciable.

5.1 background introduction

Base station sub-system can produce some and disturb certainly in protection at interval, thereby the reception of uplink pilot time slot is caused interference.This interference is to be caused by the reflection of surrounding environment to descending pilot frequency time slot.Concerning the higher base station sub-system of height above sea level, the distorted signals time can reach 100 microseconds.The used algorithm of the present invention is suitable for overcoming this ill effect, and guarantees the reliability that uplink pilot time slot detects.

5.2 arthmetic statement

5.2.1 structured flowchart

In the TD-SCDMA system, descending pilot frequency time slot and protection at interval intersection can from descending switch to up.Because barrier on every side can cause the reflection and the scattering of process time delay to descending pilot frequency time slot, their can cause interference to the reception of institute's uplink pilot time slot, thereby reduce the performance of system.Fig. 6 shows the frame structure of TD-SCDMA of the present invention system.

Fig. 6 is the frame structure in the TD-SCDMA system

5.2.2 the block diagram of algorithm interference cancellation algorithm of the present invention is divided into two-stage, wherein first order algorithm is used for eliminating

The interference that causes owing to the emission of self descending pilot frequency time slot (DwPTS), and second level computing method is used to compensate the adverse effect that causes owing to strong SYNC1.

Described second level computing method is reconstructed (wherein adopt with predominating path and detect the same door limit value that uses) according to relevant multipath distribution character to detected SYNC1, and averages according to the average mode identical with the first order.Output signal to the first order compensates then.Like this, can detect from SYNC1 and eliminate the influence that SYNC1 caused the input of digital signal processor.

The benefit of this process is: this algorithm can be restrained sooner, but also can the SYNC1 that really receives be made a response.

In a word, algorithm can be divided into for two steps:

A. do not consider possible uplink synchronous code word, in first order interference cancellation algorithm, implement the elimination of reflected signal;

B. the detected uplink synchronous code word of reconstruct, then in order to eliminate of the influence of strong uplink synchronous code word to average in the next frame, need be in the interference cancellation algorithm of the second level the uplink synchronous code word signal feedback of reconstruct.

Provide the block diagram of two-stage interference cancellation algorithm among Fig. 7.

5.2.3 algorithmic elaboration

5.2.3.1 the embodiment of first order algorithm

Signal has 256 chips, comprising the protection interval of 96 chips behind the descending pilot frequency time slot and the uplink pilot time slot of 160 chips.According to analyzing and measuring, being evenly distributed of interference is stable and predictable, therefore can be easily deducts because the descending pilot frequency time slot adverse effect the received signal during uplink pilot time slot.

Adopt algorithm of the present invention that the interference signal of each chip in 256 the chip of each antenna is averaged, this algorithm can be as described below:

${\mathrm{av}}_m\left(i\right)=\frac{255\·{\mathrm{av}}_{m-1}\left(i\right)+\mathrm{inst}\_{\mathrm{value}}_m\left(i\right)}{256},i=\mathrm{1,2}...,256-----\left(1\right)$

Av wherein _m(i) and inst_value _m(i) represent in m the frame (burst) average interference on i the chip and received may comprise the uplink synchronous code word respectively, and 1/P is the forgetting factor that uses in the averaging process, for example P=256 in interior instantaneous received signal.In fact, the parameter P that uses in the described weighted average process can be 2 power, for example 2,4,8...128,256 ... in one.Consider the phase modulation of the descending pilot frequency time slot that can learn by frame number, must eliminate the influence of the phase modulation of descending pilot frequency time slot by selecting "+" or "-" in the formula (1) be respectively applied for real part and imaginary part in averaging process.

And last, the phase modulation of the descending pilot frequency time slot of current impulse can be recovered by deduct the average that calculates in the former frame from the signal that is received, and promptly obtains the output through the first order interference eliminated of revising:

modi_value_1 _m(i)＝inst_value _m(i)-av _m-1(i)，i＝1，2，...，256 (2)

Wherein choosing of the real part and the imaginary part of symbol will depend on corresponding downstream pilot time slot phase modulation in this subtraction process.These additions or subtraction have guaranteed not have in complexity the stability of descending pilot frequency time slot phase place under the substantive prerequisite that increases.

Should be noted that, in above-mentioned formula (2), need from instantaneous received signal, deduct previous mean value signal av _M-1Rather than the current mean value signal av that calculates (i), _m(i).

5.2.3.2 second level interference eliminated embodiment

In the second level, need to calculate the average av_sync1_recon of uplink synchronous code word _m(i), utilize phase add operation compensation then because the influence that strong uplink synchronous code word is caused.The SYNC1 average av_sync1_recon that former frame calculates _M-1(i) can be compensated in the output of the first order, that is:

modi_value_2 _m(i)＝modi_value_1 _m(i)+av_sync1_recon _m-1(i) (3)

Then, utilize the output of second level interference cancellation algorithm to go to detect and reconstruct uplink synchronous code word, it is exported by inst_sync1_recon _m(i) expression, it can be used to remove computation of mean values av_sync1_recon according to following formula _m(i):

$\mathrm{av}\_\mathrm{sync}1\_{\mathrm{recon}}_m\left(i\right)=\frac{(P-1)\·\mathrm{av}\_\mathrm{sync}1\_{\mathrm{recon}}_{m-1}\left(i\right)+\mathrm{inst}\_\mathrm{sync}1\_{\mathrm{recon}}_m\left(i\right)}{P},i=\mathrm{1,2},...,256-----\left(4\right)$

If do not detect the SYNC1 code word, the inst_sync1_recon of reconstruct then _m(i) should equal zero.

Because of what this was noticed be, in above-mentioned two formula, oeprator+or-and selects real part or imaginary part must depend on the phase place of downlink pilot frequency channel (DwPCH), and identical in use therein transformation rule and the first order algorithm.And same, the parameter P that uses in the weighted average process can be 2 power, for example 2,4,8...128,256 ... in one.

5.2.3.3 the reconstruct of detected uplink synchronous code word

If in the uplink synchronous code word detects DSP, implement the circular correlation computing, can for example be the correlation of 8 road detected uplink synchronous code character code words and the power of correlation then according to following formula output multichannel (way can be antenna number) herein:

FFT_code[j][1:256]＝FFT(sync1_code_256)；

Rx_value_corre[ante][j][1:256]＝

IFFT(FFT_code ^*[j][1:256].*FFT(modi_value_2[ante][1:256])) (6)

Rx_power_corre[ante][j][k]＝(abs(Rx_value_corr[ante][j][k])) ² (7)

The related power value is carried out reprocessing, promptly utilize threshold value to go to eliminate wherein because the influence that noise caused.The correlation that is lower than threshold value is set to zero, that is:

The basic principle of reprocessing has been described among Fig. 8.

Then, utilize following formula reconstruct uplink synchronous code word:

inst_sync1_recon[ante][j][1:256]＝

IFFT(FFT(Post_value_corre[ante][j][1:256])./FFT_code ^*[j][1:256])) (9)

Finally, the uplink synchronous code character code word inst_sync1_recon of reconstruct is fed in the middle of the second level interference cancellation algorithm to the next frame signal.

The simulation result of above-mentioned two-stage interference cancellation algorithm has been described among Fig. 9 and Figure 10, and wherein Fig. 9 and Figure 10 provide the example that receives the uplink synchronous code word in present frame and the former frame respectively.

A 5.3 example of the MATLAB language description of first order interference eliminated

％This is brief illustration about algorithmAntenna_Nbr＝8；％antenna numberChip_Nbr＝253；％chip number，the other 3 chips are used for delivering otherinformationBurst_Nbr＝256；％average burst numberMax_Burst_Nbr＝10000；％the maximum received burst number which is determinedby the number of message given by the DSP developersp＝1/Burst_Nbr；％forgetting factor％initialization of some bufferaverage_value＝zeros(Antenna_Nbr，Chip_Nbr)；％average interference valueinstant_value＝zeros(Antenna_Nbr，Chip_Nbr，Max_Burst_Nbr)；％received dadavalue per antenna per burst per chip        <!-- SIPO <DP n="11"> -->        <dp n="d11"/>DwPTS_phase＝zeros(Max_Burst_Nbr)；％modulation phase of DwPTSmodified_value＝zeros(Antenna_Nbr，Chip_Nbr，Max_Burst_Nbr)；％output data％end of initialization％initialization of instant_value and DwPTS_phaseinstant_value＝；％current received dataDwPTS_phase＝[0，2，2，2，0，1，1，2，0，1，2，1，0，3，2，3，0，2，1，3，0，2，3，3，0，2，2，1，0，2，2，3，0，1，2，2，0，3，2，2，0，2，1，2，0，2，3，2]；％current phase％end initializationfor k＝1：Max_Burst_Nbrinstant_work(1：Antenna_Nbr，1：Chip_Nbr)＝instant_value(1：Antenna_Nbr，1：Chip_Nbr，k)；modified_work＝zeros(Antenna_Nbr，Chip_Nbr)；％the algorithms to be implemented in the BTSC start here！！！  switch DwPTS_phase(k)％to determine the modulation phase of the current burst based on Frame Number  case 0％++(a+jb)*1＝a+jb，pi/4    for i＝1：Antenna_Nbr      for j＝1：Chip_Nbr        ％to restore phase        modified_work(i，j)＝instant_work(i，j)-average_value(i，j)；        average_value(i，j)＝average_value(i，j)*(Burst_Nbr-1)+instant_work(i，j)；        average_value(i，j)＝average_value(i，j)＞＞8；      end        <!-- SIPO <DP n="12"> -->        <dp n="d12"/>    end  case 1％-+(a+jb)(-j)＝b-ja，3*pi/4    for i＝1：Antenna_Nbr      for j＝1：Chip_Nbr           ％to restore phase，avrage_value multiply j，(a+jb)*j＝-b+jareal(moddified_work(i，j))＝real(instant_work(i，j))+imag(average_value(i，j))；imag(modified_work(i，j))＝imag(instant_work(i，j))-real(average_value(i，j))；real(average_value(i，j))＝real(average_value(i，j))*(Burst_Nbr-1)+imag(instant_work(i，j))；imag(average_value(i，j))＝imag(average_value(i，j))*(Burst_Nbr-1)-real(instant_work(i，j))；         real(average_value(i，j))＝real(average_value(i，j))＞＞8；          imag(average_value(i，j))＝imag(average_value(i，j))＞＞8；        end      end    case 2％--(a+jb)(-1)＝-a-jb，5*pi/4      for i＝1：Antenna_Nbr         for j＝1：Chip_Nbr        <!-- SIPO <DP n="13"> -->        <dp n="d13"/>          ％to restore phase，avrage_value multiply(-1)，(a+jb)*(-1)＝-a-jb          modified_work(i，j)＝instant_work(i，j)+average_value(i，j)；          average_value(i，j)＝average_value(i，j)*(Burst_Nbr-1)-instant_work(i，j)；          average_value(i，j)＝average_value(i，j)＞＞8；        end      end    case 3％+-(a+jb)(j)＝-b+ja，7*pi/4      for i＝1：Antenna_Nbr        for j＝1：Chip_Nbr           ％to restore phase，avrage_value multiply(-j)，(a+jb)(-j)＝b-jareal(modified_work(i，j))＝real(instant_work(i，j))-imag(average_value(i，j))；imag(modified_work(i，j))＝imag(instant_work(i，j))+real(average_value(i，j))；real(average_value(i，j))＝real(average_value(i，j))*(Burst_Nbr-1)-imag(instant_work(i，j))；           imag(average_value(i，j))＝imag(average_value(i，j))*(Burst_Nbr-1)+real(insant_work(i，j))；           real(average_value(i，j))＝real(average_value(i，j))＞＞8；           imag(average_value(i，j))＝imag(average_value(i，j))＞＞8；         end        <!-- SIPO <DP n="14"> -->        <dp n="d14"/>     end  otherwise     disp(′Wrong phase of DwPTS！′)；  end％this is the end of the algorithm to be implemented in the BTSCmodified_value(1：Antenna_Nbr，1：Chip_Nbr，k)＝modified_work(1：Antenna_Nbr，1：Chip_Nbr)；end％this is the output of this algorithm，1：Max_Burst_Nbr are the effective value

5.4 the complexity analyzing of algorithm

5.4.1 first order interference eliminated

The first step of algorithm deducts previous average exactly from received signal, owing to be to carry out complex operation, therefore need respectively real part and imaginary part to be operated.

Then, because variable av _mAnd inst_value _m(i) be plural number, then required amount of calculation is (real multiplication) computing of 2 multiplication and 2 additions (real addition) computing in a circulation (iteration), be that 2 basic multiplications (multiplication and additioncalculation is hereinafter to be referred as MAC) that add are arranged in each chip.

Next procedure carries out the divide operations in the dyadic shift operation.

So needed amount of calculation is exactly 4 MAC in each chip.

As a result, the computation complexity of whole algorithm is:

Multiplexing value=4*253* antenna number MAC/0.005 second (2)

As antenna number is 8, and then multiplexing value was 1,620,000 MAC/ seconds.

Each recurrence averager (recursive averager) needs the memory capacity of 2 integers (real part and imaginary part), just 4 bytes.

Therefore, total storage demand amount is:

Storage capacity requirement amount=4*253* antenna number (byte) (3)

As antenna number is 8 o'clock, and the memory capacity that then needs is 8096 bytes.

5.4.2 second level interference eliminated

Its computational complexity and desired memory cell all with the first order in require identical.

(5) conclusion

The required huge operand of the multiplexing number of this algorithm and joint detection algorithm is compared, and in fact can ignore.Therefore, this algorithm extremely is easy to implement.

Claims (15)

1, a kind ofly in tdd systems, overcome from the method for disturbing, described tdd systems comprises more than one base station, and each base station can send respectively and received signal, and described method comprises the following steps: each base station

(1) transmitting downstream signal,

(2) received signal, and discern the reflecting part of above-mentioned downstream signal, and from total received signal, identify upward signal,

(3) eliminate its original transmitted and go out reflected signal in the signal, and

(4) residual signal is handled, to detect upward signal.

2, the method for claim 1, the mode of wherein eliminating reflected signal is:

(1) utilize signal of being launched and the signal that is received, determine the channel impulse response of received signal, and

(2) from the signal that is received, deduct the convolution of channel impulse response and the signal launched.

3, the method for claim 1, wherein:

(1) come the calculating channel impulse response with a pulse,

(2) handle same received signal,

(3) determine where partly being caused by reflection of channel impulse response, which part has nothing to do with reflection.

4, method as claimed in claim 3, wherein adopt following step to discern and the channel impulse response that reflects relevant channel impulse response and have nothing to do with reflection:

(1) under the known condition of channel impulse response, sets up the equation group how a relevant reflected signal constitutes;

(2) find the solution this equation group according to channel impulse response;

(3) be zero with less channel impulse response tap setting, make judgement;

(4) then, utilize, recomputate the distortion of signal through the channel impulse response of reprocessing and above-mentioned equation group;

(5) repeat aforesaid step (1) to step (4), to discern reflecting part and non-reflective portion in the above-mentioned channel impulse response.

5, method as claimed in claim 3, wherein the variation of the channel impulse response between frame and the frame is not remarkable, and it is characterized in that: the channel impulse response to former frame averages.

6, method as claimed in claim 3, wherein the variation of the channel impulse response between frame and the frame is not remarkable, it is characterized in that:

Channel impulse response with former frame is constructed the signal that is reflected.

7, method as claimed in claim 4, wherein the variation of the channel impulse response between frame and the frame is not remarkable, it is characterized in that:

Channel impulse response to former frame averages.

8, method as claimed in claim 4, wherein the variation of the channel impulse response between frame and the frame is not remarkable, it is characterized in that:

Channel impulse response with former frame is constructed the signal that is reflected.

9, method as claimed in claim 3 is a non-modulated to the transmission signals relevant with major part in the reflected signal that is received wherein, and the channel impulse impulse response between frame and the frame changes not remarkable, it is characterized in that:

(1) reflected signal that receives is averaged, and

(2) from instantaneous received signal, deduct the average reflection signal that is received, become part in this signal when non-to eliminate.

10, as the described method of claim 1-7, tdd systems wherein is the TD-SCDMA system, and the most of reflected signals that received cause by descending pilot frequency time slot, it is characterized in that:

(1) 64 chips of descending pilot frequency time slot is modulated, made the phase place of these 64 chips identical, and, have 4 kinds of different phase places for all 64 chips of descending pilot frequency time slot;

(2) waveform of the channel impulse response between frame and the frame remains unchanged substantially;

(3), use the method that signal is rotated to go the modulation of conversion reception reflected signal, and then resulting signal is averaged according to above-mentioned phase modulation;

(4) the average received signal is modulated again, and from instantaneous received signal, deducted this signal, then uplink pilot time slot is detected.

11, method as claimed in claim 10, it is characterized in that for eliminating of the influence of possible uplink synchronous code word average in the next frame, implement following second level interference cancellation algorithm: the detected uplink synchronous code word of reconstruct, SYNC1 signal to reconstruct is averaged then, the output signal of feeding back whole interference cancellation algorithm then with the second level interference cancellation algorithm of replenishing the next frame signal input signal.

12, method as claimed in claim 11 is characterized in that the step of described second level interference cancellation algorithm is as follows:

(1) calculates the correlation of the detected uplink synchronous code word of multichannel and the power of correlation;

(2) these correlations are carried out reprocessing, go these related power values of comparison with threshold method, the correlation that is lower than threshold value be set to zero, to eliminate interference of noise;

(3) average (av_sync1_recon of calculating uplink synchronous code word _m(i)): the uplink synchronous code word of current reconstruct and uplink synchronous code word average last time are weighted on average; Obtain the average of current uplink synchronous code word, use for the second level interference cancellation algorithm of next frame signal;

(4) the uplink synchronous code word average (av_sync1_recon that former frame is obtained _M-1(i)) be added to the output (modi_value_1 of the first order of next frame signal _m(i)) in, obtain partial output valve, this output valve can be used as the output valve of whole algorithm, simultaneously the input value that detects as new uplink synchronous code word again;

(5) to later each frame detection signal, repeat (1) calculating process to (4).

13. method as claimed in claim 12 is characterized in that in the described weighted average process that the weight of uplink synchronous code word average last time is made as (P-1)/P, the weight of the uplink synchronous code word of current reconstruct is 1/P.

14. method as claimed in claim 12, the way that it is characterized in that the uplink synchronous code word can be the number of antenna.

15. method as claimed in claim 13 is characterized in that the parameter P that uses in the weighted average process can be 2 power, for example 2,4,8...128,256 ... in one.

Images