CN1677877A - Assembly structure of time-domain synchronous orthogonal frequency-division multiplex receiver - Google Patents

Abstract

The invention comprises following structures: digital baseband signal is obtained from analog signal in high frequency after passing through tuner and two times of digital frequency conversion including steps of automatic gain control, frequency evaluation, and time evaluation etc. Then, PN codes are captured. Digital baseband signal is divided into two parts: data and signals of PN codes. After phase correction is carried out for data part, channel equalization is carried out based on channel evaluation obtained from PN codes. The result is sent to forward error correction for decoding. Finally, sending sequence is recovered. The invention makes TDS - OFDM receiver well adapt ideal or non-ideal baseband model. Through theoretical analysis, computer simulation, FPGA prototype, ASIC products in small order etc. main performance reaches or exceeds system requirement and better than other digital TV transmission system.

Description

The time-domain synchronization OFDM receiver general structure

Technical field

The invention belongs to digital information transmission technical field, particularly a kind of time-domain synchronization OFDM (Time D omainSynchronous OFDM, the TDS-OFDM) general structure of digital television receiver.

Background technology

TV all is most important consumption electronic product for any country in the world today.Experienced the mechanical television epoch, black and white electronic television and color TV be after the epoch, TV has entered a New Times now: the Digital Television epoch.Digital Television is meant the making of whole employing digital form, transmission and received television program, and image, the sound that the beholder is watched be equivalent to TV station's studio program quality do not have ghost image and " snowflake ".It is the high-tech product that integrates hi-tech developments such as digital signal and the information processing technology, digital communication technology, computer and network technology, microelectric technique.Digital television broadcasting mainly via satellite, three kinds of transmission meanss of cable TV and terrestrial wireless realize.It is generally acknowledged that satellite broadcasting focuses on that solving large tracts of land covers.Cable tv broadcast focuses on solution " information is to the family ", particularly at population inhabitation congested areas such as cities and towns.And terrestrial wireless broadcast is as the traditional means of television broadcasting and since its unique simple reception and move the ability that receives, can satisfy present information society desired " information be'ss to the people " primary demand.So (Digital TV terrestrial broadcasting DTTB) will possess great commercial value to ground digital television broadcast in following many decades.Along with the development of digital television broadcasting over past ten years, the basic transmission technology of satellite and cable tv broadcast is ripe.And the transmission environment of ground digital television broadcast is abominable, frequency spectrum resource is limited, application demand is disperseed, its technology still has a lot of places that need improvement, especially potentiality are significantly improved at the stable and mobile aspects such as performance that receive that improve fixing reception, the present invention satisfies the DTTB demand condition with regard to being based on the design that such consideration brings forward the receiver general structure of DTTB, and main performance index arrives requirement.

Have only which demand of clear and definite system to satisfy, could seek corresponding technically or through compromise relative best solution.

For ground digital television broadcast, at first require Digital Television that enough good receptivity is arranged, simple, the small-sized and stable reception of low-gain antenna realization in indoor employing.Even in the environment of strong static state and dynamic multi-path, system still can steady operation.

Secondly, sufficiently high transmission code rate is arranged,, consider the broad space that Digital Television develops in the future and the diversity of service application in single 8MHz channel, the demand of transmission capacity is constantly increased so that high-quality high definition program (approximately 20Mbps) is provided.

Also will help frequency planning, use transmission DTV program in the existing television channel that distributes, realization and the same of analog television program are broadcast; When not having extra channel allocation, can use forbidding (Taboo) channel (, can not be used for the channel of simulated television), and have and the suitable coverage of existing simulated television platform because interference is excessive.Other requirement comprises: chnnel coding and channel estimation scheme that needs are advanced, so that reduce the C/N of system thresholding, reduce transmitting power, and reduced interference existing analog television program with this, and resist various interference/distortions.The operator scheme of high flexible, by selecting different modulation schemes, system can support to fix, portable, walking or move receives.Be easy to the interface with other media or server, support Polymera/business, can obtain class-of-service, have interactivity by hierarchical modulation.The frequency planning of high flexible and overlay area can use single frequency network and co-channel covering expansion/slit to fill.And the receiver realization of many kinds of cost prices of system's assentment, comprise low-cost realization or the like.

Sum up the demand condition, the main design object of visible DTTB system is to realize the efficient utilization of frequency spectrum, in the fixing and mobile receiving ability that guarantees to provide stable under the enough big message transmission rate.

In three kinds of modes of digital television broadcasting, terrestrial broadcasting is to use the most widely, is specially adapted to the country that the region is vast, radio network is complicated.But the interference that the terrestrial broadcasting channel faces is maximum, and is also the most serious, and especially the pace of change of the time delay of multipath and amplitude is more than satellite and wire cable channel complexity.The broadcast environment of satellite and wire cable and desirable awgn channel are very approaching, adopt outstanding chnnel coding and signal modulation system generally can make satellite and wire cable broadcast system performance near theoretical value.And in ground environment, the environment of broadcasting does not obviously satisfy awgn channel, and what system can steady operation is regional limited.Add terrestrial broadcasting requirement and existing analog television broadcast compatibility, high-power non-linear emission makes the interference aggravation between adjacent channel, the stability of a system is required harsh.Therefore, under abominable like this terrestrial broadcasting environment, how designing the system of each functional module operate as normal, is the essential technique difficult point of digital tv ground broadcasting.

The development of DTTB both at home and abroad.Digital Television has only time of short more than ten years till now from end of the eighties development, but development speed make us surprised soon.Through insistent research and development these years, DTTB has obtained a lot of achievements, and having reached can implementation phase.From in November, 1998 North America and the Europe DTTB program that started broadcasting, many countries have announced that their DTTB standard selects and realize a plan.At present, mainly contain three kinds of DTTB transmission standards in the world:

1. United States advanced television systems committee (Advanced Television Systems Committee, ATSC) eight level vestigial side bands of Yan Fa lattice code (Trellis-Coded 8-Level Vestigial Side-Band, 8-VSB) modulating system.

2. European digital video broadcast-terrestrial (Digital Video Terrestrial Broadcasting-Terrestrial, DVB-T) Coded Orthogonal Frequency Division Multiplexing (COFDM) of standard employing (Coded Orthogonal Frequency DivisionMultiplexing, COFDM) modulation.

3. Japanese floor synthetic service digital broadcasting (Integrated Service Digital Broadcasting-Terrestrial, ISDB-T) the frequency band segment transmissions of Cai Yonging (Bandwidth Segmented Transmission, BST) orthogonal frequency division multiplex OFDM.

Since these three DTTB systems had been arranged, many countries and regions were all in the DTTB system of selecting oneself.But along with the development of technology and deepening continuously of research, people recognize at signal peak/average power ratio, C/N thresholding, move receive, indoor/outer reception, spectrum efficiency, HDTV transmittability, with frequency/adjacent channel disturb, aspect the flexibility of interference, single frequency network and co-channel forwarding, impulse disturbances and continuous wave interference to having simulated television now, phase noise, static state/dynamically multipath distortion, system or the like, respectively there are its pluses and minuses in above-mentioned three systems, and each country all needs according to the national conditions selection of this country or formulates the Digital Television standard of oneself.At present, the ATSC standard has been adopted in the U.S., Korea S, Canada and China's Taiwan a few countries and area, the most of country in Europe is as Britain, Germany, Spain etc., and plurality countries and regions such as Australia, Brazil, Mexico, Singapore, China Hong Kong have adopted the DVB standard, and only Japan adopts the ISDB standard.

China also began the research work of high definition TV from 1994.And succeeded in developing Chinese first generation functional prototype of high definition TV in 1998.Organized the expert of China that Digital Television and digital high-definition television standard are formulated by radio and television responsible institution, SARFT(The State Administration of Radio and Television), (DigitalMultimedia Broadcasting for Terrestrial, DMB-T) host-host protocol is one of testing scheme to Tsing-Hua University's T-DMB.

The TDS-OFDM that adopts among the DMB-T of Tsing-Hua University belongs to multi-transceiver technology, but it is different with the COFDM in Europe, in TDS-OFDM, do not insert the pilot tone signal, but utilized spread spectrum technic, inserted the PN sequence time domain carry out frame synchronization, Frequency Synchronization, regularly synchronously, channel transfer characteristic is estimated and follow the tracks of phase noise etc.Owing to used spreading code, make DMB-T have the superior function of spread spectrum communication, thereby the system that makes both had the advantage of COFDM, had avoided its shortcoming again.

DMB-T receiver general structure proposed by the invention is at this scheme.In order more clearly to explain the present invention, below the structure of general introduction DTTB transmission plan earlier, on this architecture basics, introduce modulator approach and simultaneous techniques thereof that external three kinds of DTTB transmission plans adopt.

The DTTB system all is made up of compression layer, transport layer and transport layer three parts, and its system construction drawing is seen Fig. 1.Wherein compression layer is exactly a source encoding, comprises the compressed encoding of sound and image, mainly contains the MPEG series standard of ISO/IEC at present.Along with the development of technology, can adopt other new compression algorithm, for example wavelet coding, fractal image etc.The code stream multiple connection belongs to transport layer, mainly single channel or multichannel elementary stream are carried out multiple connection according to standard or other corresponding standard of MPEG-2 system layer, multiple a plurality of code streams such as audio code stream, video code flow, data code flow are combined into a transmission code stream, its length is fixed, so that Channel Transmission is inserted various time tags simultaneously and is used for indication and synchronous.

Chnnel coding and channel-decoding belong to transport layer, and the superiority of digital television transmission system performance is mainly derived from chnnel coding and signal modulation system.Compression layer in the DTTB scheme all is the same with transport layer basically at present, and difference is just on transport layer.According to different channel situation and different application demands, digital television system has adopted different error correction codings and modulation technique scheme, and its formation is seen Fig. 2.Error correction coding part has all adopted the cascade error correcting code basically, comprises that the error correction of outer sign indicating number, time domain interweave, ISN error correction, frequency-domain-interleaving.And modulation technique mainly contains two kinds of scheme one single-carrier modulated and multi-carrier modulation at present, belong to the ATSC that the U.S. is arranged and the 8VSB of the Shanghai Communications University modulation of single-carrier modulated, and DVB-T COFDM, the Japanese ISDB-T BST OFDM that Europe is arranged of multi-carrier modulation and the DMB-T TDS-OFDM of Tsing-Hua University.Because selection of technical scheme and specific implementation parameter is different, caused the difference of performance between the different system.

The ATSC digital television standard is the ATSC of Advanced Television Systems Committee exploitation.Initial design object is to be used for terrestrial broadcasting and the cabled distribution system that fixed outdoor receives, and does not support portable and mobile reception, and indoor reception is also bad.

Modulation scheme has adopted the single-carrier modulated with pilot signal, i.e. eight level vestigial side bands modulation (8-VSB), and (multiple frequency network, MFN Multi-Frequency Network MFN) realizes to be used for the single transmit machine.The design of this system allows to distribute the extra digital transmitter with suitable coverage on existing NTSC transmitter, and minimum to existing NTSC program influence aspect zone and population covering.It is the high development of the existing AM modulation technique of maturation, and its performance height depends on adaptive equalizer, therefore in order to resist multipath echo and various interference, needs very complicated equalizer.

U.S. VSB system has added the pilot signal of 0.3dB, is used for subcarrier and recovers.Transmission signals adopts section, field structure, and framing sends, and frame structure as shown in Figure 3.The segment sync signal that adds is used for system synchronization and clock recovery; And length reaches 511 two level field sync signals, is used for system synchronization and equalizer training, eliminates the echo interference that multipath fading causes by adopting well-designed adaptive decision feedback equalizer.Be anti-NTSC co-channel interference, adopt in receiver, to add interference suppression filter, also claim comb filter (because of its amplitude-frequency characteristic is pectination).

In addition, system is equipped with stronger inside and outside chnnel coding error correction safeguard measure.So the 8-VSB system of design compares with ISDB-T with DVB-T, for additive white Gaussian noise (AWGN) and little ghost image channel have stronger repellence, higher spectrum efficiency, lower peak value-average compares and the higher reliability of paired pulses noise.

But there are a series of problems in U.S. ATSC system, most importantly tackles strong dynamic multi-path difficulty.This is that defective by original design philosophy of system, frame structure itself causes.Under the little strong multipath situation of time delay, pilot signal can be had a strong impact on, and encounters difficulties synchronously.Especially equalizer performance sharply descends: though system has used training sequence, but the 24ms of being separated by between two training sequences, the quick variation of multipath can't be tracked during this time, though use complex structure decision-feedback (DFE), the error signal of utilizing data itself to produce is carried out self adaptation and is regulated, change fast multipath in order to follow the tracks of, but DFE needs channel to be balanced to a certain degree (mistaken verdict is less than 10%) ability operate as normal, and DFE is infinite impulse response structure (IIR), under strong multipath, system is unsettled.In addition, for the comb filter that tackles the co-channel interference use makes the thresholding rising 3dB of system, and realize complicated.

The DVB-T system is European digital video broadcasting (Digital Video Broadcasting, DVB) organization development.Design object is the fixing reception in indoor and outdoor, and provides portable reception and non-moving reception.Therefore, its mobile reception is also bad.The terrestrial broadcasting in the existing UHF frequency spectrum of distributing to analog television transmission of this system.The effectively clean bit code check of transmission depends on chnnel coding parameter, modulation type and protection selection at interval in the 8MHz channel in 4.98～31.67Mbit/s scope.

Aspect terrestrial transmission, it adopts the modulation technique COFDM different with U.S. 8-VSB, and OFDM belongs to multi-carrier modulation technology (being called the DMT modulation in the similar technology of ADSL).DVB-T has used 1705 (2k patterns) or the individual subcarrier of 6817 (8k patterns) in each television channel.The frame structure of DVB-T as shown in Figure 4.The basic principle of OFDM is divided into a broad frequency band some subchannels (Sub Channel or Subcarrier) exactly.If the shared bandwidth of each subchannel is enough narrow, they will experience flat fading respectively.In this case, the equalizer of receiver is easy to realize.And in order to improve system spectral efficiency, the frequency spectrum of each subchannel is overlapping in the ofdm system, but is quadrature again between them, the origin of its OFDM (OFDM) title that Here it is.Under the multipath channel situation, in order to keep the orthogonality between its each subchannel, must add protection at interval (Guard Interval, GI).DVB-T uses Cyclic Prefix, and (Cycle Padding, CP) protection copies to each code element front end with OFDM code element last part at interval exactly.

We know, the DFT of the circular convolution of two signals equals their products behind the DFT respectively:

DFT{d _nh _n}＝DFT{d _n}×DFT{h _n} (1)

Relation between signal and the channel impulse response is the linear convolution relation; under the situation of Cyclic Prefix (suffix) expansion; as long as the length of channel impulse response is less than protection length at interval; linear convolution is equal to circular convolution; just can overcome because the intersymbol interference that brings of channel multi-path (Inter Symbol Interference, ISI).Finish synchronous task for auxiliary, DVB-T has placed a large amount of pilot signals at frequency domain, interts among data subcarrier, and sends with the power that is higher than data 3dB.As shown in Figure 5.These pilot signals can finish system-frame synchronously, carrier wave recovers, clock adjustment and channel estimating.Because changes in channel characteristics faster can be followed the tracks of and estimate to the use of scattered pilot.

In addition, length selectable protection interval and subcarrier pattern allow the DVB-T system to support different network configuration, and the 8k pattern is applicable to that large-scale single frequency network (SFN) or 2k pattern are used for moving reception.Simultaneity factor can be selected different QAM number of modulation levels and different ISN code checks, can provide two layer services in a DTTB channel.

Compare with the 8-VSB modulation technique of the U.S., the COFDM technology belongs to the new technology that can access continuous development and raising, especially has performance advantage aspect opposing high level (high), long delay static state and the dynamic multi-path distortion to 0dB, this system can overcome the interference of time delayed signal reliably, the echo that comprises the reflection of physical features or building, the perhaps signal of distant place transmitter emission in the single frequency network environment.What point out any is that U.S. Sinclair broadcasting group has carried out the acceptance test of ATSC and DVB-T in the area, Baltimore of the U.S. 7～August in 1999, test result is very unfavorable to ATSC, this group has united the FCC that submits a written statement to a higher authority of tame TV station more than 200, requirement allows the TV station of the U.S. also can use the COFDM technology when using ATSC.Except European countries, Australia, Singapore, India successively announce to adopt the DVB standard in Europe at present.

Also there are some defectives in DVB-T.At first be that too much pilot tone makes that the frequency band loss is serious.Analyze the pilot tone of DVB and be not difficult to find that pilot signal and protection have occupied about 14% of effective bandwidth at interval at least, if adopt big protection at interval, this numerical value will be above 30%.Therefore, DVB-T is to be the anti-multipath performance that cost exchanges system for the undue valuable system transmission capacity that descends.Secondly, the thin synchronized algorithm of COFDM will utilize pilot signal to realize on frequency domain, and pilot tone was inserted before DFT, and carrying out DFT calculating needs (just can carry out correct demodulation afterwards) at first synchronously again.Therefore, COFDM adopts synchronous the needs and uses the iterative approach algorithm, so just has a convergence error and convergence time problem.And compare with single carrier 8-VSB, COFDM is responsive more to synchronous error, the bad meeting of net synchronization capability causes receiver performance obviously to descend, this also is that COFDM is short at AWGN and time delay, under the little multipath of energy under the situation performance not as the reason of 8-VSB.When using COFDM to carry out channel estimating in addition, regard multidiameter fading channel as on time and frequency a 2D signal, utilize sample interpolation can obtain the frequency response values of whole channel.Obtain an all channel and estimate to need continuous 4 COFDM frames, estimated time is also longer, makes bad to the tracking performance that changes channel in time fast.

The ISDB-T system is that (Association of Radio Industries andBusinesses ARIB) develops Japanese association of radio industrie and business.Design object is that the indoor and outdoor of integrated service is fixed, portable and mobile broadcast reception.The modulator approach that system adopts is called frequency band segment transmissions (BST) OFDM, is made up of one group of common fundamental frequency piece that is called the BST section.Except that the OFDM characteristic, BST-OFDM adopts different carrier modulation scheme and ISN encoder bit rate to different BST sections, and the hierarchical transmission characteristic is provided according to this.Each data segment has its oneself error protection scheme (ISN encoder bit rate, the time-interleaved degree of depth) and modulation type (QPSK, DQPSK, 16-QAM or 64QAM), and every section can be satisfied different business demands like this.Many sections can be grouped together neatly, and broadband services (for example HDTV) is provided.This system is derived from European system, and main change is at multimedia broadcasting and moves the demand that receives, and frequency band has been carried out segmentation, and the interleave depth (reaching 0.5 second most) that extended greatly, to improve mobile reception.Though Japanese system shows certain superiority (the also corresponding European system that proved needs improvement) in mobile test, it does not solve the question of substance among the COFDM.

Ground digital multimedia TV broad cast (the Digital MultimediaTV Broadcasting-Terrestrial that Tsing-Hua University proposes, DMB-T) purpose of scheme provides a kind of digital information transmission method, it uses the television broadcasting frequency spectrum, and the message transmission code check of effective payload of each channel can be up to 33Mbps under the bandwidth of 8MHz.The core of system has adopted the time-domain synchronization OFDM of mQAM/QPSK, and (its spectrum efficiency can be up to 4Bit/s/Hz for Time Domain Synchronous OFDM, TDS-OFDM) modulation technique.System uses more optimal forward error correction coding FEC to resist error burst, for example Reed-Solomon (Reed-Solomon, RS) or Turbo code, convolution code and their cascade etc.Realize hierarchical modulation and coding, class-of-service is provided, can realize multimedia service simultaneously.

Seeing grant number for details about the correlation circumstance of DMB-T, TDS-OFDM is that 00123597.4 " ground digital multimedia TV broad cast system " by name, grant number are 01115520.5 " time-domain synchronous orthogonal frequency division multiplex modulation method " by name, and grant number is the Chinese invention patent that 01124144.6 " protection fill method at interval in the orthogonal FDM modulation system " by name waits Tsing-Hua University to apply for.

For realize quick and stable synchronously, the DMB-T transmission system has adopted hierarchical frame structure.It has periodically, and can and Absolute Time Synchronization.The elementary cell of frame structure is called signal frame, as shown in Figure 6.253 signal frames are defined as a frame group, and 512 frame groups are defined as a superframe.The top layer of frame structure is called a day frame, is made up of superframe.The frame group of DMB-T transmission system is made of a control frame and 252 signal frames subsequently.Each frame group's duration is 139.15ms.First signal frame among the frame group is defined as frame group head (control frame), is used to transmit this frame of control group's signaling.Each signal frame among the frame group has unique frame number, and it is coded in the PN sequence of frame head.Each frame group is by the frame group number sign of a 9bit.The frame group number is coded in the transmission parameter signaling (TPS) of signal frame.TPS repeats in each signal frame in the frame group, only could change when new frame group begins.The superframe of DMB-T transmission system comprises 512 frame groups.Each frame group in the superframe is by its frame group number unique identification.First frame group number of superframe is 0, and last frame group number is 511.The duration of each superframe is 71.2448s.The day frame of DMB-T transmission system is made up of 1213 superframes, and is to carry out periodicity repetition the cycle with a consecutive days.In Beijing time 0:0:0AM or other selected reference time, the frame structure of DMB-T transmission system is reset and begins a new day frame.Each day, last superframe of frame was incomplete.

The signal frame of DMB-T transmission system uses the OFDM modulation of Domain Synchronous, and perhaps being called with the PN sequence is protection OFDM modulation at interval.A signal frame is made up of frame synchronization and frame two parts, and they have identical baseband signalling rate 7.56 MS/s (1/T).Consider the time selectivity (Doppler distributes and is about 100Hz) of channel, the length of each signal frame is defined as＜600 μ s.A signal frame can be used as an OFDM (OFDM) piece.An OFDM piece further is divided into a protection interval and an inverse discrete Fourier transform (IDFT) piece.For TDS-OFDM, frame synchronization sequence is as the protection interval of OFDM, and frame is as the IDFT piece, as shown in Figure 7.

Frame synchronization in the signal frame is made of preamble, synchronous three parts of 8 rank PN sequences and back.The PN sequence definition is 255 symbols, preamble and after be defined as the cyclic extensions of PN sequence synchronously, account for 512 symbols altogether with the PN sequence.8 rank PN sequence definitions are proper polynomial x ⁸+ x ⁶+ x ⁵The m sequence of+x+1, its initial condition will be determined the phase place of the m sequence that generated.The frame number of each signal frame determines the initial condition of its m sequence.Each signal frame among the signal frame group is assigned unique frame synchronizing signal, with the recognition feature as signal frame.

Another explanation of TDS-OFDM is the OTDM of PN sequence and IDFT piece.Because the PN sequence is a known array for receiving terminal, PN sequence and IDFT piece are can be separated at receiving terminal.

The PN sequence is except as the protection of OFDM piece at interval, can also be used as the purposes such as frame synchronization, carrier wave recovery and automatic frequency tracking, symbol clock recovery, channel estimating of signal frame at receiving terminal.The signal frame of receiving terminal remove and can be regarded as after the PN sequence and have at interval OFDM of zero padding protection (Zero padding OFDM, ZP-OFMD).People are verified, and having zero padding protection OFDM at interval is of equal value with having Cyclic Prefix (CP) protection OFDM at interval in theory.DVB-T and ISDB-T adopt to have Cyclic Prefix protection COFDM modulation at interval, no matter adopt CP or ZP as protection at interval, and what it transmitted in fact all is redundant information, frequency spectrum and power resource that occupying volume is outer.In order to obtain good synchronization performance, except using protection at interval, also will insert a large amount of pilot signals at frequency domain simultaneously, pilot signal accounts for data symbol 10% greatly.Hence one can see that, and TDS-OFDM will protect at interval and pilot signal lumps together, and can improve about 10% channel capacity.

Certainly, adopting PN protection preceding topic at interval is to guarantee that the PN sequence is enough little to the influence of data, and this can accomplish under the PN of known transmission sequence and characteristic of channel situation.This paper below will discuss and how use the PN sequence to finish the TDS-OFDM system synchronization, use the characteristic of channel of estimating to eliminate the influence of PN sequence to data in emulation.

The frame of signal frame adopts multi-carrier modulation, and the frequency domain sub-carrier number of frame is 3780, adjacent sub-carrier be spaced apart 2kHz, each sub-carrier adopts 64QAM planisphere.The frame of signal frame also has transmission parameter signaling (TPS) except normal data flow, in order to the transfer system configuration information.It is made up of 36 bits, and is mapped as 18 subcarriers with QPSK.The frame of signal frame is with 3762 data bits at 18 TPS symbol heels.

Mainly describe existing external three kinds of DTTB and the DMB-T of Tsing-Hua University system transmissions scheme above, mainly comprised the main simultaneous techniques and the existing problems thereof of modulator approach, corresponding frame structure and employing.Below we mainly discuss the basic principle of general ground digital television receiver, be divided into inner receiver and outer receiver and realize, and major function and the respective algorithms structure that inner receiver will be finished is described.

The is-symbol sequence that information theory is mainly studied.At first the source symbol is mapped to channel symbol sequence X=(x ₁..., x _n...), then generate channel output sequence Y=(y by these sequences ₁..., y _n. ...).Output sequence is at random, and list entries is depended in its distribution, the message that we will recover to send from output sequence.Any one communication system sent be with sequence X corresponding continuous time of waveform s (t, x), rather than sequence itself.The channel symbol sequence is finished by modulator to the distribution of channel waveform.Except that channel sequence X, waveform also depends on parameter set θ={ θ _T, θ _C.Subclass θ _TBe transmitter parameter, subclass θ _CBe channel parameter, these parameters are unknown for receiver.For recovering the symbol sequence X, receiver must estimate these unknown parameters from received signal, and these estimated values will be used as actual value then.Even these parameters are strict not corresponding with chronomere, we still say that becoming is " synchronous detecting " on the accurate meaning to it.

In the physical communication model of digital receiver, provide the division of inner receiver and external receiver, as shown in Figure 8 like this.The special role of inner receiver is to create " well " channel into external receiver, synchronization parameter is estimated its output sequence Y{ θ _T, θ _C, this sequence Y and transmitter, channel " synchronous ", the performance that makes external receiver is as far as possible near the ideal communication channel condition.And the main task of external receiver is to finish the best decoding that sends sequence.

Under the simplest awgn channel situation, parameter θ={ θ _T, θ _CMainly comprise phase theta or time delay ε.In this case, the main task of inner receiver is to estimate θ and ε from the known signal that is subjected to noise jamming.Then carry out parameter adjustment, as regulating θ (t) by the phase place that changes voltage controlled oscillator (VCO) according to estimated value.In multi-path environment, the estimation under finishing the additive noise situation, also to use channel estimator that channel is followed the tracks of.

On notion, the estimation standard of these synchronization parameters be maximum posteriori criterion (Maximum posterioriProbability, MAP).When the input data satisfy to wait general the distribution, MAP was consistent with maximum-likelihood criterion.So with the angle of mathematics, algorithm for estimating is exactly received signal r in fact _fLikelihood function ρ (r _f| α, ε, θ) maximization problems.As (θ ε) unites when estimating ρ (r when carrying out _f| α, ε θ) becomes

$\mathrm{\ρ}\left(r_f\right|\mathrm{\θ},\mathrm{\ϵ})=\underset{a}{\mathrm{\Σ}}P\left(a\right)\mathrm{\ρ}\left(r_f\right|\mathrm{\θ},\mathrm{\ϵ},a)---\left(2\right)$

But find the solution likelihood function generally speaking is very difficult, according to different approximation methods, has obtained various synchronous algorithm for estimating:

1) according to synchronized algorithm needs data message whether, can be divided into following two classes:

The judgement sensing (Decision Directed, DD)/data auxiliary (Data Aided, DA) auxiliary (Non-Data Aided, NDA) method of method and free of data.When the given data sequence, for example in acquisition procedure, send a known array a ₀, be called the synchronized algorithm that data are assisted (DA).The NDA algorithm is not know under the situation of sequence a, to various may sequence a average and obtain.

2) position of extracting from received signal according to the synchronous error signal can be divided into algorithm for estimating two classes: forward direction (Feedforward, FF) and the feedback (Feedback, FB).The forward direction class estimates it is to extract error signal before the recovery unit synchronously.Feedback sort is estimated to extract error signal after synchronous recovery unit, then corrected signal feedback is given the unit of front end.Feedback arrangement itself has the ability that automatic trace parameters slowly changes, and therefore also is known as the Error Feedback synchronizer.

3), algorithm can be divided into and other parameter correlation is estimated and irrelevantly estimated according to estimating whether other synchronization parameter information of needs of synchronization parameter.Irrelevant estimation does not need other synchronizing informations, and such algorithm uses with regard to cocoa before other parameter synchronization.To timing estimation, just can be divided into and phase-independent and relevant algorithm for estimating.When using with the phase-independent algorithm, the user does not need phase information, and such algorithm for estimating can just use before carrier synchronization, generally signal will be carried out delivery and handle.

4) kind according to synchronous digital signal can be divided into continuous signal estimation and burst estimation.Digital signal can be divided into two classes according to the length that sends data: continuous signal and burst.Continuous signal will send several thousand characters at least before interruption.Burst generally is based on packet, and data packet length can be selected (generally from tens characters to several thousand characters).For the synchronous recovery algorithms difference of this two classes signal, continuous signal requires algorithm that follow-up control will be arranged, and can follow the tracks of in a long time regularly to change; And burst requires the algorithm pull-in time short, can finish synchronously within a short period of time.

Since the sixties till now, a lot of scholars have been arranged at the top dissimilar ML algorithm for estimating that proposed, as shown in Figure 9, and the performance of these algorithms analyzed.We propose to be fit to the algorithm of DMB-T needs analyzing on these algorithm bases.

Summary of the invention

The present invention is according to the characteristic of the TDS-OFDM of ground digital multimedia TV broad cast DMB-T system of Tsing-Hua University, taken all factors into consideration the influence of the ideal basis band model of DMB-T system and various non-ideal factors such as timing is synchronous, Frequency Synchronization, channel estimating, elimination phase noise to the ground digital television receiver general structure design, at the characteristics of received terrestrial digital broadcasting channel, a kind of whole implementation structure of DMB-T receiver has been proposed.

The whole implementation structure of a kind of TDS-OFDM receiver that the present invention proposes is through stages such as theory analysis, Computer Simulation, FPGA model machine and ASIC be pilots, main performance all meets and exceeds system requirements, verified the novelty of DMB-T transmission system aspect signal modulation and net synchronization capability, performance is better than existing other digital television transmission system.

The invention provides a kind of TDS-OFDM receiver general structure, it is characterized in that, comprise with the lower part:

1, the time-domain synchronization OFDM receiver general structure is characterized in that, it realizes that in digital circuit it contains:

Once simulate frequency changer circuit, it has a reception time-domain synchronization OFDM is the antenna of TDS-OFDM signal;

The secondary digital frequency converter, it contains A/D translation circuit, Hilbert filter circuit, lower frequency changer circuit, sample value interpolating circuit and square root raised cosine (SRRC) low-pass filter circuit of mutual serial connection successively; Wherein, A/D translation circuit input links to each other with the output of free-running clock circuit, and its another input links to each other with above-mentioned output of once simulating frequency changer circuit;

Frame synchronization is PN sign indicating number capture circuit, and its input links to each other with the output of above-mentioned SRRC low pass filter;

Symbol timing recovery is the STR circuit, also claims the STR clock recovery circuitry, and its input links to each other with the PN sign indicating number output of vertical sync circuit;

Carrier recovery circuit, its input links to each other with the output of above-mentioned SRRC low-pass filter circuit, and its output links to each other with another input of low-converter in the above-mentioned secondary digital frequency converter;

Automatic gain control is agc circuit, its two inputs link to each other with the TDS-OFDM signal output part of above-mentioned Hilbert circuit and PN sequence, the frame data output end of following separator respectively, and its output links to each other with the input of once simulating frequency changer circuit;

PN synchronizing signal and data extractor, its input links to each other with the output of above-mentioned SRRC low-pass filter circuit;

Phase noise is removed circuit, and it is made of phase noise estimation, phasing and compensation and DFT translation circuit serial connection successively; Wherein another input of phase-correcting circuit links to each other with the frame data-signal output of above-mentioned separator; The output of DFT translation circuit links to each other with the input of phase noise circuit;

Channel estimating and equalizing circuit, it contains channel estimation circuit and the channel equalization circuit that is in series successively; Wherein two of channel estimation circuit inputs are that PN sign indicating number, frame data-signal output link to each other with the frame synchronization of above-mentioned separator respectively; The output of channel equalization circuit links to each other with channel estimation circuit input in the channel estimation circuit frequency domain decision-feedback circuit;

The decoding FEC circuit is the fec decoder circuit, and its input links to each other with the output of above-mentioned channel equalization circuit, and its output is data-signal.

2, time-domain synchronization OFDM receiver general structure according to claim 1, it is characterized in that: described vertical sync circuit is formed by interpolation filter, 4/T times multiplier, correlator, peak detector, sequence number detection switch, phase matcher and local PN code generation circuit serial connection successively, and the output of local PN code generation circuit links to each other with the input of above-mentioned correlator; The input that wherein inserts filter links to each other with above-mentioned SRRC low-pass filter circuit output.

3, time-domain synchronization OFDM receiver general structure according to claim 1, it is characterized in that: described STR clock recovery circuitry is made of timing error extractor, error normalization circuit, loop filter, one times of sampling interval delayer, digital controlled oscillator of being connected in series successively, and wherein the input of timing error extractor links to each other with the output of correlator in the vertical sync circuit; The output of digital controlled oscillator links to each other with the input of insertion filter in the vertical sync circuit.

4, time-domain synchronization OFDM receiver general structure according to claim 1, it is characterized in that: described carrier recovery circuit is connected in series successively by deaccentuator and restore circuit and forms, wherein the output of restore circuit links to each other with another input of low-converter in the secondary digital frequency conversion circuit, and the output of the phase matched circuit in the input of deaccentuator and the above-mentioned vertical sync circuit links to each other.

5, time-domain synchronization OFDM receiver general structure according to claim 1, it is characterized in that: described agc circuit contains: the time domain agc circuit, it is formed by accumulator and low pass filter serial connection, and the input of accumulator links to each other with the PN sequence output of above-mentioned separator; The frequency domain agc circuit, it is connected in series successively by DFT circuit, smoothing circuit, accumulator, low pass filter and forms; The input of DFT circuit links to each other with the frame data output end of above-mentioned separator; Adder, its two inputs link to each other with the output of two low pass filters among above-mentioned time domain AGC, the frequency domain AGC respectively, and its output links to each other with the above-mentioned frequency conversion input of once simulating.

6, time-domain synchronization OFDM receiver general structure according to claim 1, it is characterized in that: described phase noise is removed circuit, wherein noise estimating circuit contains: Chuan Jie compare selection circuit, band pass filter, second multiplier, IDFT circuit and the second complex conjugate computing circuit, Chuan Jie channel estimation circuit, first multiplier and the first complex conjugate computing circuit successively successively; Wherein another input of first multiplier links to each other with the output of compare selection circuit, the output of the first complex conjugate computing circuit links to each other with another input of second multiplier, and the input of channel estimation circuit links to each other with the PN sign indicating number output of above-mentioned separator; Described phase compensating circuit input links to each other with the output of the above-mentioned second complex conjugate computing circuit, and its another input links to each other with the frame data output end of above-mentioned separator; The input of described DFT circuit also links to each other with the frame data output end of above-mentioned separator, and the output of DFT circuit links to each other with the input of compare selection circuit, and its input also links to each other with the frame data output end of above-mentioned separator.

7. time-domain synchronization OFDM according to claim 1 receives general structure, it is characterized in that, and described channel estimating and equalizing circuit, it contains:

Adpative equalizing circuit, it is to be connected in series successively by a DFT circuit, adaptive equalizer and level decision circuit to form, wherein, the DFT circuit is that the DFT data output end links to each other with the frame data of above-mentioned separator;

Channel estimation circuit, it contains following three class channel estimation circuits:

1) channel estimation circuit of frequency domain decision feedback usefulness;

2) channel estimation circuit used of frequency domain PN code conversion;

3) channel estimation circuit of the relevant usefulness of the time domain of PN sign indicating number;

Ternary selector switch: its three inputs link to each other with the output of above-mentioned three kinds of channel estimation circuits respectively, and its output links to each other with another input of above-mentioned adaptive equalizer;

The channel estimation circuit of described frequency domain decision feedback usefulness, its two inputs link to each other with DFT circuit, the output of level decision circuit in the above-mentioned Adpative equalizing circuit respectively;

The input of the channel estimation circuit of the relevant usefulness of that described frequency domain PN code conversion is used and PN sign indicating number time domain links to each other with the PN sign indicating number output of above-mentioned separator respectively;

The channel estimation circuit that local PN code generation circuit, its two outputs are used with the frequency domain PN code conversion respectively and the input of the channel estimation circuit of the relevant usefulness of PN sign indicating number time domain link to each other.

Above-mentioned secondary digital frequency conversion module comprises that the AD converter carries out the conversion of analog to digital signal, be decomposed into 1 tunnel (in-phase component) and Q road (quadrature component) data-signal by hilbert filter then, the IQ signal is divided into two-way: the one tunnel gives control module, the IQ signal and the frame synchronization of catching later produce an AGC control voltage together in AGC, remove to control the gain amplifier of high-frequency tuner; Another road IQ signal is given low-converter, multiplies each other with the local carrier that recovers, then through obtaining digital baseband signal behind sample value interpolation and the SRRC low pass filter.

Above-mentioned phase correction module is fed back, and the data behind the phasing therefrom obtain phase noise and estimate through the DFT conversion, then this estimation is fed back to phase correction module the data before the DFT are carried out phase compensation.

Above-mentioned channel estimating and balance module are based on the PN signal section and carry out channel estimating, obtain the channel response of each OFDM piece, then the data behind the phasing are carried out channel equalization, so that proofread and correct the data sampling that each receives, balanced data feeds back to channel estimating then, so that the next frame channel estimating is more accurate.

The present invention is taking all factors into consideration the influence to the receiver overall performance of ideal basis band model that the TDS-OFDM system faced and various non-ideal factor, the characteristic that has in conjunction with the TDS-OFDM system, characteristics at the received terrestrial digital broadcasting channel, the whole implementation structure of the TDS-OFDM receiver that proposes is through the checking of each link, main performance all meets and exceeds system requirements, be better than existing other digital television transmission system, the implementation algorithm complexity reduces greatly, has novelty aspect signal modulation and the net synchronization capability.

Theory analysis, Computer Simulation have been carried out at a kind of TDS-OFDM receiver general structure that the present invention proposes, main performance all meets and exceeds requirement of system design, and FPGA model machine and asic chip have been finished, in the actual tests in national terrestrial DTV test and many places, obtained sufficient checking, proved the novelty of a kind of TDS-OFDM receiver general structure that the present invention proposes, performance is better than existing other digital television transmission systems.

Description of drawings

Fig. 1 is the DTTB system configuration.

Fig. 2 is the terrestrial DTV transport layer.

Fig. 3 is the ATSC frame structure.

Fig. 4 is the DVB-T frame structure.

Fig. 5 is the DVB-T pilot signal.

Fig. 6 is the DMB-T hierarchical frame structure.

Fig. 7 is the structure chart of DMB-T signal frame.

Fig. 8 is general digital television receiver model.

Fig. 9 is the classification of ML synchronized algorithm.

Figure 10 is a TDS-OFDM system sending and receiving end block diagram.

Figure 11 is a kind of overall construction drawing of TDS-OFDM receiver.

Figure 12 TDS-OFDM receiver carrier wave recovers theory diagram.

Figure 13 TDS-OFDM receiver frame synchronization principles block diagram.

Figure 14 TDS-OFDM receiver STR clock recovery theory diagram.

Figure 15 TDS-OFDM receiver adaptive channel is estimated and balanced theory diagram.

Figure 16 TDS-OFDM receiver is removed the phase noise theory diagram.

Figure 17 TDS-OFDM receiver automatic gain control principle block diagram.

Embodiment

Below in conjunction with accompanying drawing theory analysis of the present invention and specific embodiment are described in detail.

Figure 10 has shown the TDS-OFDM system base band block diagram of realizing with DFT.Suppose and comprise N=3780 subcarrier in the ofdm system.Wherein, we use 36 subcarriers in the frequency spectrum to transmit TPS, other subcarrier transmission data.Each subcarrier is all by a complex symbol X _{I, k}Modulation, wherein subscript i is the sequence number of TDS-OFDM signal frame, k is the sequence number of sub-carrier number.Can use QPSK or qam mode, X according to the transmission plan design on each subcarrier _{I, k}Promptly represent the point on the respective constellation figure.We can modulate with the IDFT that N=3780 is ordered at transmitting terminal.

Make [X ₀, X ₁..., X _N-1] the expression data symbol, owing to be the data of a signal frame here, for the ease of mark, ignored subscript i, and made 0≤k≤N-1.Use linear transformation I DFT with complex data symbol [X ₀, X ₁..., X _N-1] be mapped as OFDM symbol [x ₀, x ₁..., x _N-1]:

$x_n=\sqrt{\frac{1}{N}}\underset{k=0}{\overset{N-1}{\mathrm{\Σ}}}{X}_k{e}^{j2\mathrm{\πn}\frac{k}{N}}---\left(3\right)$

The discrete-time signal sampling interval of IDFT output is the same with the PN sequence of insertion, is system symbol period T=T _u/ N=1/7.56 microsecond, wherein T _uBe the length of frame in the TDS-OFDM signal frame, i.e. T _uBe 500 microseconds, 1/T _uBe the frequency interval between the OFDM subcarrier, be 2KHz.In order to overcome intersymbol interference, before the OFDM frame, also to add PN frame synchronization head, we use T _Frame=T _u+ Δ represents to comprise the length of the whole TDS-OFDM signal frame of PN sequence, and Δ is the length of frame synchronization head.Like this, whole signal frame comprises N _Frame=N+ Δ/T=N+N _gIndividual sampled point, wherein N _gThe number of an expression frame synchronization sampled point, this concrete use N that implements _g=512.Discrete signal frame passes through forming filter (SRRC low pass filter) afterwards, and the continuous complex signal of base band of transmitting terminal output just can be expressed as:

s(t)＝P _t(t)*[PN(i)+x(i)] (4)

Wherein, p _t(t) be the impulse response of SRRC filter, * represents linear convolution, PN (i) be the i frame by the PN sequence c (i, k), 0≤k≤N _gThe frame head that-1} forms, x (i) be the i frame by sequence x (i, k), the frame that 0≤k≤N-1} forms.PN (i) takies the different time with x (i).

Suppose that (t, τ), at receiving terminal, the signal that receives at first passes through matched filter p to above signal by a multidiameter fading channel h _r(t), at this moment signal becomes:

r _f(t)＝(s(t)+n(t))*h _c(t，τ) (5)

H wherein _c(t, τ)=p _t(t) * h (t, τ) * p _r(t) being the impulse response that comprises the compound channel of forming filter, channel and matched filter, is the impulse waveform that receiver end is seen.The additive white Gaussian noise of n (t) expression channel makes the variance of its real part and imaginary part be:

$E\left[\mathrm{Re}{\left(n\left(n\right)\right)}^2\right]=E\left[\mathrm{Im}{\left(n\left(n\right)\right)}^2\right]={\mathrm{\σ}}_n^2---\left(6\right)$

And the average power of putative signal s (n) is 2 σ _s ², that is:

$E\left[\mathrm{Re}{\left(s\left(n\right)\right)}^2\right]=E\left[\mathrm{Im}{\left(s\left(n\right)\right)}^2\right]={\mathrm{\σ}}_n^2---\left(7\right)$

Signal-to-Noise $\mathrm{SNR}={\mathrm{\σ}}_s^2/{\mathrm{\σ}}_n^2.$

We suppose the receiving terminal ideal synchronisation, sample at moment t=nT to received signal.Because code element is synchronous, after frame head PN sequence is removed in i the signal frame, N remaining frame sampled point { r _k, 0≤k≤N-1} will be input to and carry out demodulation among the DFT.Suppose to send and receiving filter satisfies Nyquist (Nyquist) law that receiver is output as in that DFT is later:

$Y_k=\sqrt{\frac{1}{N}}\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}{r}_n{e}^{-j2\mathrm{\πk}\frac{n}{N}}---\left(8\right)$

Wherein

$r_k=\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}{H}_{n,k}{X}_n{e}^{j2\mathrm{\πk}\frac{n}{N}}+n^{\′}\left(k\right)---\left(9\right)$

And H _{N, k}Expression compound channel h (t, τ) in the frequency response complex values at frequency n/T place:

$H\left(k\right)=\frac{1}{\sqrt{N}}\underset{l=0}{\overset{L-1}{\mathrm{\Σ}}}h\left(l\right)e^{-j2\mathrm{\πk}\frac{l}{N}}---\left(10\right)$

Wherein L is the length of channel echoes.

n′(k)＝p _r(t)*n(t)| _t＝kT (11)

Therefore have

Y _k＝H _n，kX _k+N′(k) (12)

$N^{\′}\left(k\right)=\frac{1}{\sqrt{N}}\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}{n}^{\′}\left(n\right)e^{-j2\mathrm{\πk}\frac{n}{N}}---\left(13\right)$

N ' is a white Gaussian noise (k), and n ' DFT conversion (k) can prove that the noise component(s) on the different subchannels is incoherent.

In real system, various all can not be desirable synchronously, so when analyzing DMB-T system realistic model, consider the influence of various non-ideal factors, these non-ideal factors mainly comprise:

1) problem of timing error.This comprises two aspects of frame synchronization sum of errors frequency errors of sampling clock.After the frame synchronization error was the lock-on signal frame head, can there be skew ε T in intercepting DFT piece position with respect to ideal position.The sampling clock error be receiving end sampling clock T ' can not with clock originator T complete matching.

2) carrier frequency problem.In the ofdm system link, only when transmitter and receiver uses identical carrier frequency, each subcarrier is only quadrature.And in practice, oscillator can not provide the carrier frequency with complete single determined value, and always has the random phase shake, so its frequency is not constant.

So, more than these asynchronous factors can comprise in system's equivalent channel model.Consider their combined influence, we obtain m received signal r (t constantly _m) be:

$r\left(t_m\right)=e^{j2\mathrm{\π}{\mathrm{\θ}}_0\left(m{T}^{\′\′}\right)}\underset{k=-\∞}{\overset{\∞}{\mathrm{\Σ}}}s\left(k\right)h_s^{\′}({\mathrm{mT}}^{\′}-\mathrm{kT},{\mathrm{mT}}^{\′})+n^{\′}\left({\mathrm{mT}}^{\′}\right)---\left(14\right)$

Wherein, h _ε' be the channel impulse response that has skew ε T,

Expression carrier frequency shake T " and initially differs θ ₀The phase place rotation of introducing.

As seen, be in the system of transmission means with OFDM, be the key that can whole system operate as normal synchronously.The main task of DMB-T inner receiver is estimated above various synchronization parameter exactly, and is adjusted accordingly to received signal.These synchronous workings comprise carrier synchronization, and frame synchronization and sampling clock are synchronous, and inner receiver also wants outside receiver that channel condition information (Channel Status Information, estimation CSI) are provided We know, outstanding chnnel coding and signal modulation system be generally all at the white noise modelling, so the design criterion of inner receiver is for making outer receiver performance approach the performance that various parameter Estimation and adjustment all are desirable outer receivers as much as possible.For this criterion of quantitative description, the SNR loss that definition is actual: SNR _LossSNR _LossBe defined as and obtain identical systems performance (as the error rate), real system is equivalent to the extra SNR of idealized system needs.That is:

SNR _loss(dB)＝SNR _real(dB)-SNR _ideal(dB) (15)

SNR wherein _Real(dB) and SNR _Ideal(dB) be respectively the SNR of real system and idealized system under the identical systems performance.More than each the amount all represent with dB.

In addition, except suppose that desirable we also think: channel is constant (this is that the DFT conversion is needed) synchronously in an OFDM frame, the impulse response of channel is finite length (just can ignore the influence of ISI to system like this).

In the real system, more than supposition all might be false.The influence of various imperfect transmission conditions to systematic function is discussed below, so that the SNR that imperfect transmission conditions are caused _LossA quantitative understanding is arranged, the index that clear and definite DMB-T receiver general structure of the present invention later on will reach.

Suppose because frame is asynchronous and cause that the DFT window position has been offset ε T with respect to ideal position in the receiver.For the CP-OFDM system, owing to adopt CP at interval as protection, and also the length Δ of CP is greater than the maximum additional delay τ of channel _Max, timing offset does not destroy the orthogonality between the subcarrier within the specific limits, and it is equivalent to produce phase place rotation with the sub-carrier frequencies linear change to the influence of restituted signal, at the edge of frequency band, the rotation maximum of phase place:

$Y_{i,n}=e^{j2\mathrm{\&pi;}\frac{n}{N}\mathrm{\&epsiv;}}{X}_{i,n}\&CenterDot;H_{i,n}+n_{i,n},(-\mathrm{\&Delta;}+{\mathrm{\&tau;}}_{\max })/T<\mathrm{\&epsiv;}<0---\left(16\right)$

But the DMB-T of Tsing-Hua University has adopted the PN sequence as protection at interval, though reduced overhead, timing offset is had higher requirement.The ISI that the timing offset that exists produces will destroy the orthogonality between each subcarrier, cause inter-carrier interference (Inter-carrier Interference, ICI).For example, when ε＞0, the output of inner receiver can be expressed as:

$Y_{i,n}=e^{j2\mathrm{\&pi;}\frac{n}{N}\mathrm{\&epsiv;}}\frac{N-\mathrm{\&epsiv;}}{N}{X}_{i,n}\&CenterDot;H_{i,n}+n_{i,n}+n_{\mathrm{\&epsiv;}}(i,n)---\left(17\right)$

Wherein because the ICI that ISI causes can be summed up as an additional noisiness n _ε(i, n).As seen, the timing of code element synchronously must be very accurate, makes n _ε(i is n) than Gaussian noise n _{I, n}Little a lot.The protection of adopting than general CP-OFDM based on the relevant frame synchronization performance of PN that the present invention proposes below is related algorithm accuracy height at interval, can satisfy system requirements.

Many document analysises because the asynchronous problem that causes of OFDM receiver sampling clock frequency.We define normalized sampling clock frequency offset be β=(T '-T)/T, then inner receiver is output as:

$Y_{i,n}=e^{j2\mathrm{\&pi;in\&beta;}\frac{\mathrm{Ts}}{T_u}}\sin c\left(\mathrm{n\&beta;}\right)X_{i,n}\&CenterDot;H_{i,n}+n_{i,n}+n_{\mathrm{\&beta;}}(i,n)---\left(18\right)$

As seen from formula (18), sampling frequency offset mainly contains the influence of two aspects: the first, and it makes useful signal by sinc (n β) decay, and phase change exp (the j2 π in β T that becomes when causing _s/ T _u), the amount of phase rotation and sub-carrier number n are directly proportional, and along with frame number i linear growth.Sample clock frequency deviation must cause the linear growth of frame timing error simultaneously, worsens the frame synchronization problem, and receiver must be followed the tracks of it.The second, the sampling clock error causes ICI.Because the error of sample frequency means the DFT error of blanking time, therefore destroyed the orthogonality between each subcarrier of ofdm system.N in the formula _β(i n) represents the extra noise that is caused by ICI, and its variance can be approximated to be formula (19), has only when n β＜＜1 n _β(i n) just can ignore.

${\mathrm{\&sigma;}}_{\mathrm{\&beta;}}^2\&ap;\frac{{\mathrm{\&pi;}}^2}{3}{\left(\mathrm{n\&beta;}\right)}^2---\left(19\right)$

Ofdm system is high more a lot of to the sensitivity of carrier wave than carrier wave communication system to the sensitivity of carrier wave frequency deviation, so frequency shift (FS) is a major issue that influences performance in the ofdm system design.Suppose that there is carrier wave frequency deviation Δ f in system, then the output of inner receiver can be expressed as:

$Y_{i,n}=e^{j2\mathrm{\&pi;i\&Delta;f}{T}_s}\sin c\left(\mathrm{\&pi;\&Delta;f}{T}_u\right)X_{i,n}\&CenterDot;H_{i,n}+n_{i,n}+n_{\mathrm{\&Delta;f}}(i,n)---\left(19\right)$

From (19) as can be known, carrier wave frequency deviation is similar to the influence and the sampling frequency deviation of systematic function, mainly is divided into two kinds: the one, and make useful signal amplitude fading, phase place rotate, next has produced serious ICI, and ICI is become noise n by equivalence _{Δ f}(i, n).Because the system SNR that carrier wave frequency deviation causes loss, wherein SNR _{Loss, Δ f}With dB is that unit can be expressed as:

${\mathrm{SNR}}_{\mathrm{loss},\mathrm{\&Delta;f}}\&ap;\frac{10}{31n10}{\left(\mathrm{\&pi;}\frac{\mathrm{N\&Delta;f}}{B}\right)}^2\frac{E_s}{N_0}---\left(20\right)$

Wherein, B is a system bandwidth.As seen, SNR _{Loss, Δ f}With Δ f square and square being directly proportional of sub-carrier number N.So, when N is big more, also be that frequency interval is more little between the subcarrier, be subject to the influence of carrier wave frequency deviation more.In order to prevent ICI, ofdm system must remain on 10 with the variance of carrier wave frequency deviation ^-6Below.

If carry out coherent detection at the OFDM receiver, then receiver must carry out channel estimating.According to general type, the estimated value of channel frequency response value in the ofdm system But approximate representation becomes actual value H _{I, n}With an additive white Gaussian noise component n _{I, n} ^HSum.That is:

${\hat{H}}_{i,n}=H_{i,n}+n_{i,n}^H---\left(21\right)$

N wherein _{I, n} ^HRepresented channel estimation error, its variances sigma _H ²Be channel estimating mean square deviation (Mean SquareError, MSE).Obtain channel estimating After carry out channel equalization, the balanced data symbolic representation is:

${\hat{X}}_{i,n}=\frac{Y_{i,n}}{{\hat{H}}_{i,n}}\&ap;X_{i,n}+\frac{n_{i,n}}{H_{i,n}}+\frac{X_{i,n}}{H_{i,n}}{n}_{i,n}^H---\left(22\right)$

(X wherein _{I, n}/ H _{I, n}) n _{I, n} ^HRepresented the extra noise component of introducing owing to channel estimating is imperfect, it causes the signal constellation which dislocation, the dislocation of outside constellation point than inner constellation point more seriously, influenced by channel estimation errors bigger so comprised the amplitude modulation(PAM) transmission plan of numerous constellation point.We suppose With n _{I, n} ^HWith n _{I, n}It is independent,

The whole system noise variance is like this:

${\mathrm{\&sigma;}}_{\mathrm{sys}}^2={\mathrm{\&sigma;}}_n^2+{\mathrm{\&sigma;}}_H^2---\left(23\right)$

We lose with regard to the system SNR that has obtained causing owing to channel estimation errors like this, are expressed as with dB:

${\mathrm{SNR}}_{\mathrm{loss},H}={10\log }_{10}\frac{{\mathrm{\&sigma;}}_{\mathrm{sys}}^2}{{\mathrm{\&sigma;}}_n^2}={10\log }_{10}(1+\frac{{\mathrm{\&sigma;}}_H^2}{{\mathrm{\&sigma;}}_n^2})={10\log }_{10}(1+G)---\left(24\right)$

As seen, SNR _{Loss, H}Depend on G fully.From formula (24) SNR as can be seen _{Loss, H}It is non-linear influenced by channel estimation errors, when G is big, and SNR _{Loss, H}Loss is serious, but along with G reduces and reduces rapidly.After G is less than some, SNR _{Loss, H}Almost no longer change.This point is pointed out us, and channel estimating does not also require entirely accurate, as long as reach a certain requirement, because the raising of general precision of channel estimation is a cost to improve complexity.

The influence of phase noise can be modeled as two parts: the one, and (Common Phase Error, CPE), it causes the phase place rotation of the data symbol that all receive in the present frame to the common phase part, causes the integral body rotation of signal constellation (in digital modulation).Generally, oscillator three dB bandwidth B _φThan little many of OFDM symbol rate, the influence of the phase error that can reduce to have with tracking technique.Second portion is to disperse part, and it is similar to white Gaussian noise, will cause defocusing of corresponding received signal constellation point, and this is that the SNR loss amount that the ICI of introducing causes is shown in (25) because the spacing between the frequency domain sub-carriers no longer is 1/T accurately:

${\mathrm{SNR}}_{\mathrm{loss},P}\&ap;\frac{11}{61n10}4\mathrm{\&pi;}\frac{{\mathrm{NB}}_{\mathrm{\&phi;}}}{B}\frac{E_s}{N_0}--\left(25\right)$

By formula (25), the deterioration that disperses phase noise to cause is the dull increasing function of sub-carrier number N, as N=1 when (being equivalent to single-carrier system), reaches minimum.Deterioration amount also with B _φ/ B is proportional, guarantee little SNR loss, and 3dB phase noise bandwidth must be far smaller than subcarrier spacing.

The present invention mainly be at the proposition of above-mentioned various factors a kind of DMB-T receiver general structure.Certainly, except the non-ideal factor of introducing above, also have a lot of other non-ideal factors, as the non-linear influence that brings of transmitter amplifier, additional noise that time varying channel brings or the like.

With reference to the method for designing of general Digital Television inner receiver, the present invention proposes the inner receiver general structure that is fit to DMB-T.

The signal frame that DMB-T sends is made up of PN frame synchronization head and OFDM data, in order to carry out Domain Synchronous, can use known PN sequence to carry out.We consider one by the PN sequence c (i, k), 0≤k≤N _gThe frame head of i signal frame forming.Suppose the transmission phase theta _TBe θ _T=Ω t+ θ.Wherein θ is fixing skew, and Ω is a frequency departure, and Ω=2 π Δ f, Δ f are carrier wave frequency deviation.We require the frequency response of the frequency response of channel and prefilter in frequency range | ω | and≤2 π B+| Ω _Max| in be smooth, wherein, B is the monolateral band bandwidth of signal, Ω _MaxIt is maximum frequency error.Under this condition, the corresponding received signal r of PN sequence _f(t) can further obtain by (5) formula:

$r_f(t,\mathrm{\&Omega;})=\underset{n=0}{\overset{N_g-1}{\mathrm{\&Sigma;}}}{c}_n{h}_c(t-\mathrm{nT}-\mathrm{\&epsiv;T})e^{j(\mathrm{\&Omega;t}+\mathrm{\&theta;})}+n^{\&prime;}\left(t\right)---\left(26\right)$

We suppose r _f(t) sampling { r _f(kT _S) be sufficient statistic, h _c(t-nT-ε T) satisfies the Nyquist sampling thheorem, promptly

$h_c\left(n\right)=\left\{\begin{array}{cc}{h}_0& n=0\\ 0& n\&NotEqual;0\end{array}\right.---\left(27\right)$

Regard again noise process as white Gaussian process, in this case, likelihood function can be reduced to:

$\mathrm{\&rho;}\left(r_f\right|\mathrm{\&epsiv;},\mathrm{\&Omega;})\&Proportional;\exp \{-\frac{1}{{\mathrm{\&sigma;}}_n^2}[2\mathrm{Re}[\underset{n=0}{\overset{N_g-1}{\mathrm{\&Sigma;}}}{\left|{\mathrm{<figure-callout\; id="0"\; label="h"\; filenames="A20041000348400321.png"\; state="\{\{state\}\}">h</figure-callout>}}_0\right|}^2{\left|c_n\right|}^2-2c_n^{*}{r}_n\left(\mathrm{\&epsiv;}\right)e^{-j(\mathrm{\&Omega;T}+\mathrm{\&theta;})}]\left]\right\}---\left(27\right)$

Dummy suffix notation r wherein _n(ε)=r _f(nT+ ε T).

Thereby obtain target function:

$\mathrm{L\&rho;}\left(r_f\right|\mathrm{\&epsiv;},\mathrm{\&theta;})=\exp \{-\frac{2}{{\mathrm{\&sigma;}}_n^2}\mathrm{Re}[\underset{n=0}{\overset{N_g-1}{\mathrm{\&Sigma;}}}{c}_n^{*}{r}_n\left(\mathrm{\&epsiv;}\right)e^{-j(\mathrm{\&Omega;T}+\mathrm{\&theta;})}\left]\right\}---\left(28\right)$

The ML target function that formula (28) obtains has been supposed nyquist requirements, and supposes that channel model is idealized, and the awgn channel of use satisfies this condition.Be channel model and the time-varying parameter that to consider more approaching reality in principle, but the mathematical operation that needs like this is too complicated.So adopt under awgn channel by the rough approximate algorithm that obtains lock unit in the present embodiment.

A lot of important conclusions all obtain from formula (28) target function.In the DMB-T digital receiver as our design, regularly recovered before the phase bit recovery.By its reason of analysis mode (28) is conspicuous.In order to minimize the computation complexity in the receiver, carrier phase is estimated and recovers to adopt as far as possible little sampling rate, just symbol rate 1/T.If regularly known, the corresponding sampling of the symbol of each matched filter output enough is used for carrier phase to be estimated and symbol detection, and the digital algorithm that is used for phase estimation so just can use with the timing relevant type (D ε type) of character rate as 1/T.

In order to find the solution the maximum target function, we adopt the Error Feedback structure.Because the video broadcasting data are streamed as continuous data, only when beginning, need the not too strict search procedure of a time requirement.And when being used to follow the tracks of purpose, the Error Feedback structure is very suitable, and can reasonably realize under the complexity.By to the target function differentiate, the nearest valuation of substitution then Obtain error signal:

$\frac{\&PartialD;L}{\&PartialD;\mathrm{\&epsiv;}}(\mathrm{\&Omega;}={\widehat{\mathrm{\&Omega;}}}_n,\mathrm{\&epsiv;}={\widehat{\mathrm{\&epsiv;}}}_n),\frac{\&PartialD;L}{\&PartialD;\mathrm{\&Omega;}}(\mathrm{\&Omega;}={\widehat{\mathrm{\&Omega;}}}_n,\mathrm{\&epsiv;}={\widehat{\mathrm{\&epsiv;}}}_n)---\left(29\right)$

Because the extraction of our error is based on the PN synchronous head, so every frame obtains a subsynchronous estimation, our algorithm just approaches the bursty data type like this.The supposition synchronization parameter is constant when receiving the OFDM data, and estimated value is used as actual value.If useful data and these two random quantitys of synchronization parameter are not separated, just must unite estimation, the computational complexity of uniting estimation is to be difficult to realize.After using suitable frame structure that these two stochastic sources are separated, data flow transmitted is made up of frame information and useful data information.Utilize known frame information can estimate the synchronization parameter of data segment, during data symbol sent, we were supposing to carry out the information bit decoding under the known condition of parameter, so the complexity of algorithm reduces greatly.

Based on top analysis, we have provided the general structure of receiver, as shown in figure 11.Below by each module is introduced:

1) AFE (analog front end)

AFE (analog front end) is a high-frequency tuner, and it amplifies the RF signal that receives, and finishes channel selection, and the signal of selecting is transformed to a fixing medium-frequency IF 1 (36.25Mhz) from RF band.The voltage of control tuner automatic gain control AGC is provided by intermediate-frequency section.Channel selection realizes by the divide ratio that changes PLL.The IF1 signal is through the filtering of 8MHz bandwidth filter.A local oscillator in intermediate frequency unit transforms to little medium-frequency IF 2 (4.5MHz) with IF1, and the signal of this moment is a bandpass signal near base band.

2) AD conversion and hilbert filter

Through after the AFE (analog front end), the filtering of analog signal process is through four sampling T _s(30.40MHz) become digital little intermediate-freuqncy signal, the sampling clock of AD converter does not pass through phase-locked, is free-running.Therefore, digital little intermediate-freuqncy signal will be recovered through the carrier wave of back just can obtain accurate carrier frequency after the processing module.Become complex signal behind the little intermediate-freuqncy signal process hilbert filter of numeral (Hilbert Filter), promptly be demultiplexed back into two paths of data: I road (in-phase component) and Q road (quadrature component).

3) carrier wave recovers and down-conversion

The frequency of receiver oscillator can not be very stable, so always there is a quantitative frequency shift (FS), regularly digital and other synchronized algorithms only could be normal under the situation of little frequency difference, is necessary so carry out the carrier wave recovery.A coarse frequency estimation that scope is bigger to be arranged when receiver powers up, need more high-precision Frequency Estimation AFC that frequency error is reduced to below the 1Hz afterwards.It is that sampled data by with time domain is multiplied by that carrier frequency shift is proofreaied and correct (one has the complex exponential that increases progressively phase place, and wherein the phase place recruitment is Variable k is a data sequence number, For with T _sNormalized frequency offset estimating value).

After the carrier signal that is restored, digital little intermediate-freuqncy signal realizes down-conversion by multiplier, obtains digital baseband signal.Carrier wave recovers as shown in figure 12.

4) clock recovery

The DMB-T clock recovery comprise the PN sign indicating number catch (Code Acquisition, CA) and symbol timing recovery (SymbolTiming Recovery, STR) two parts.During beginning, receiver is not known the phase place of PN sign indicating number in the received signal frame, catch by sign indicating number and obtain this phase place, thereby the PN sequence becomes known signal, can be used for other synchronization modules.The sign indicating number acquisition algorithm is that received signal is relevant with the local PN sequence slip that produces, because the local PN sequence that produces and the PN sequence of reception only are that a time migration is arranged, so very strong peak value will occur in the correlated results.As shown in figure 13.After sign indicating number was caught, timing error was only at ± T _sIn/2 scopes, we need more accurate timing synchronous, and STR is to remaining timing error ε ₀Estimate, obtain estimated value By linear interpolation, with sampled signal { r _f(kT _S) convert the synchrodata consistent to sending symbol rate 1/T.Because sampling clock has drift, STR adopts the second order feedback control loop to come the departure signal, finishes the tracking to sampling clock simultaneously.As shown in figure 14.

Simultaneously, after the PN sign indicating number is caught, just can from digital baseband stream, separate PN sign indicating number (frame head) part and data division (frame), give different processing modules then.

5)DFT

At receiver end, suppose that correct timing is synchronous, after the PN sequence is eliminated the influence of OFDM data, carry out the demodulation that DFT finishes OFDM by the time domain second mining sampling point after N is proofreaied and correct.

6) channel estimating and equilibrium

Receiver channel estimates that (Channel Estimation) part is mainly each OFDM piece the estimation of channel response is provided, so that proofread and correct each data sampling that receives (coherent detection).After finishing Frequency Estimation, stationary phase error that signal is also residual has also comprised this error in the channel estimating.Channel estimating also provides subcarrier reliability information for soft-decision channel decoder (outer receiver).After obtaining channel estimating, the channel equalization part realizes channel equalization (Channel Equalization) at frequency domain by simple division arithmetic.As shown in figure 15.

7) phase noise is removed

The present invention uses based on transmission parameter signaling (Transmission Parameter Signals, TPS) removal phase noise (Phase Noise Correction) method, its TPS signal behind the DFT obtains the frequency domain baseband signal of phase noise, through IDFT the frequency-region signal that obtains is converted into the time domain phase noise then and estimates, then use to obtain the time domain phase noise and estimate the data before the DFT are carried out phase compensation.As shown in figure 16.

Therefore, as can be seen from Figure 11, adopt the signal processing sequence of embodiments of the invention as follows:

Becoming digital signal after a high frequency analog signals is passed through tuner and AD conversion, is 1 tunnel (in-phase component) and Q road (quadrature component) data-signal by Variational Solution Used behind the hilbert filter.

The IQ signal is divided into two-way: the one tunnel gives control module, and the IQ signal and the frame synchronization of catching later produce an AGC control voltage together in AGC, removes to control the gain amplifier of high-frequency tuner, and the generation of AGC control voltage as shown in figure 17; Another road IQ signal is given low-converter, multiplies each other with the local carrier that recovers, then through obtaining digital baseband signal behind sample value interpolation and the SRRC low pass filter.

One way word baseband signal is through being used to control carrier wave to the AFC signal and recovering after the Frequency Estimation, obtain one relatively accurately local carrier be used for above-mentioned low-converter; Another way word baseband signal makes PN sign indicating number in the signal frame that receiver obtains to receive after through PN sign indicating number trapping module, after the PN sign indicating number is caught, it is synchronous to obtain more accurate timing through clock recovery module, being used for above-mentioned sample value interpolation handles, after the PN sign indicating number is caught simultaneously, just can be decomposed into PN sign indicating number (frame head) and DFT data-signal (frame) two parts to the signal frame that receives.

Data division is through the phase correction module of band feedback, and the data behind the phasing therefrom obtain phase noise and estimate through the DFT conversion, then this estimation is fed back to phase correction module the data before the DFT are carried out phase compensation.

The PN signal section is given channel estimation module, the channel response that obtains each OFDM piece is estimated, then the data behind the phasing are carried out channel equalization, so that proofread and correct each data sampling that receives (coherent detection), balanced data feeds back to channel estimating more then, so that the next frame channel estimating is more accurate.

Give forward error correction coding FEC module through the digital signal after the above-mentioned inner receiver processing, finally finish the best error correction decoding that sends sequence, recover to send sequence.

Theory analysis, Computer Simulation have been carried out at a kind of TDS-OFDM receiver general structure that the present invention proposes, main performance all meets and exceeds requirement of system design, and FPGA model machine and asic chip have been finished, in the actual tests in national terrestrial DTV test and many places, obtained sufficient checking, proved the novelty of a kind of TDS-OFDM receiver general structure that the present invention proposes, performance is better than existing other digital television transmission systems.

In conjunction with the accompanying drawings specific embodiments of the invention are had been described in detail above, but the present invention is not restricted to the foregoing description, under the spirit and scope situation of the claim that does not break away from the application, those skilled in the art can make various modifications or remodeling.

Claims (7)

1, the time-domain synchronization OFDM receiver general structure is characterized in that, it realizes that in digital circuit it contains:

Once simulate frequency changer circuit, it has a reception time-domain synchronization OFDM is the antenna of TDS-OFDM signal;

The secondary digital frequency converter, A/D translation circuit, Hilbert filter circuit, lower frequency changer circuit, sample value interpolating circuit and square root raised cosine that it contains mutual serial connection successively are the SRRC low-pass filter circuit; Wherein, A/D translation circuit input links to each other with the output of free-running clock circuit, and its another input links to each other with above-mentioned output of once simulating frequency changer circuit;

Frame synchronization is PN sign indicating number capture circuit, and its input links to each other with the output of above-mentioned SRRC low pass filter;

Symbol timing recovery is the STR circuit, also claims the STR clock recovery circuitry, and its input links to each other with the PN sign indicating number output of vertical sync circuit;

Carrier recovery circuit, its input links to each other with the output of above-mentioned SRRC low-pass filter circuit, and its output links to each other with another input of low-converter in the above-mentioned secondary digital frequency converter;

Automatic gain control is agc circuit, its two inputs link to each other with the TDS-OFDM signal output part of above-mentioned Hilbert circuit and PN sequence, the frame data output end of following separator respectively, and its output links to each other with the input of once simulating frequency changer circuit;

PN synchronizing signal and data extractor, its input links to each other with the output of above-mentioned SRRC low-pass filter circuit;

Phase noise is removed circuit, and it is made of phase noise estimation, phasing and compensating circuit; Wherein another input of phase-correcting circuit links to each other with the frame data-signal output of above-mentioned separator; The input of phasing and compensating circuit output and DFT translation circuit links to each other;

Channel estimating and equalizing circuit, it contains DFT circuit, channel estimation circuit and the channel equalization circuit that is in series successively; Wherein the input of DFT translation circuit links to each other with above-mentioned phasing and compensating circuit output, and the input of channel estimation circuit is that PN sign indicating number, frame data-signal output link to each other with the frame synchronization of above-mentioned separator respectively; The output of channel equalization circuit links to each other with channel estimation circuit input in the channel estimation circuit frequency domain decision-feedback circuit;

The decoding FEC circuit is the fec decoder circuit, and its input links to each other with the output of above-mentioned channel equalization circuit, and its output is data-signal.

2, time-domain synchronization OFDM receiver general structure according to claim 1, it is characterized in that: described vertical sync circuit is formed by interpolation filter, 4/T times multiplier, correlator, peak detector, sequence number detection switch, phase matcher and local PN code generation circuit serial connection successively, and the output of local PN code generation circuit links to each other with the input of above-mentioned correlator; The input that wherein inserts filter links to each other with above-mentioned SRRC low-pass filter circuit output.

3, time-domain synchronization OFDM receiver general structure according to claim 1, it is characterized in that: described STR clock recovery circuitry is made of timing error extractor, error normalization circuit, loop filter, one times of sampling interval delayer, digital controlled oscillator of being connected in series successively, and wherein the input of timing error extractor links to each other with the output of correlator in the vertical sync circuit; The output of digital controlled oscillator links to each other with the input of insertion filter in the vertical sync circuit.

4, time-domain synchronization OFDM receiver general structure according to claim 1, it is characterized in that: described carrier recovery circuit is connected in series successively by deaccentuator and restore circuit and forms, wherein the output of restore circuit links to each other with another input of low-converter in the secondary digital frequency conversion circuit, and the output of the phase matched circuit in the input of deaccentuator and the above-mentioned vertical sync circuit links to each other.

5, time-domain synchronization OFDM receiver general structure according to claim 1, it is characterized in that: described agc circuit contains: the time domain agc circuit, it is formed by accumulator and low pass filter serial connection, and the input of accumulator links to each other with the PN sequence output of above-mentioned separator; The frequency domain agc circuit, it is connected in series successively by DFT circuit, smoothing circuit, accumulator, low pass filter and forms; The input of DFT circuit links to each other with the frame data output end of above-mentioned separator; Adder, its two inputs link to each other with the output of two low pass filters among above-mentioned time domain AGC, the frequency domain AGC respectively, and its output links to each other with the above-mentioned frequency conversion input of once simulating.

6, time-domain synchronization OFDM receiver general structure according to claim 1, it is characterized in that: described phase noise is removed circuit, wherein noise estimating circuit contains: Chuan Jie compare selection circuit, band pass filter, second multiplier, IDFT circuit and the second complex conjugate computing circuit, Chuan Jie channel estimation circuit, first multiplier and the first complex conjugate computing circuit successively successively; Wherein another input of first multiplier links to each other with the output of compare selection circuit, the output of the first complex conjugate computing circuit links to each other with another input of second multiplier, and the input of channel estimation circuit links to each other with the PN sign indicating number output of above-mentioned separator; Described phase compensating circuit input links to each other with the output of the above-mentioned second complex conjugate computing circuit, and its another input links to each other with the frame data output end of above-mentioned separator; The input of described DFT circuit also links to each other with the frame data output end of above-mentioned separator, and the output of DFT circuit links to each other with the input of compare selection circuit, and its input also links to each other with the frame data output end of above-mentioned separator.

7. time-domain synchronization OFDM according to claim 1 receives general structure, it is characterized in that, and described channel estimating and equalizing circuit, it contains:

Adpative equalizing circuit, it is to be connected in series successively by a DFT circuit, adaptive equalizer and level decision circuit to form, wherein, the DFT circuit is that the DFT data output end links to each other with the frame data of above-mentioned separator;

Channel estimation circuit, it contains following three class channel estimation circuits:

The channel estimation circuit of frequency domain decision feedback usefulness;

The channel estimation circuit that the frequency domain PN code conversion is used;

The channel estimation circuit of the relevant usefulness of the time domain of PN sign indicating number;

Ternary selector switch: its three inputs link to each other with the output of above-mentioned three kinds of channel estimation circuits respectively, and its output links to each other with another input of above-mentioned adaptive equalizer;

The channel estimation circuit of described frequency domain decision feedback usefulness, its two inputs link to each other with DFT circuit, the output of level decision circuit in the above-mentioned Adpative equalizing circuit respectively;

The input of the channel estimation circuit of the relevant usefulness of that described frequency domain PN code conversion is used and PN sign indicating number time domain links to each other with the PN sign indicating number output of above-mentioned separator respectively;

The channel estimation circuit that local PN code generation circuit, its two outputs are used with the frequency domain PN code conversion respectively and the input of the channel estimation circuit of the relevant usefulness of PN sign indicating number time domain link to each other.

Images