WO2009069082A2

WO2009069082A2 - Detection method for signal frame configuration and signal frame header for broadcast signal

Info

Publication number: WO2009069082A2
Application number: PCT/IB2008/054959
Authority: WO
Inventors: Ming GONG
Original assignee: Nxp B.V.
Priority date: 2007-11-29
Filing date: 2008-11-26
Publication date: 2009-06-04
Also published as: CN101453582A; EP2220835A2; WO2009069082A3; CN101453582B

Abstract

Low-complexity detection methods and apparatus are provided for detecting signal frame configuration in a DTV receiver or the like. In one embodiment, the detection takes advantage of a difference in average power between the signal frame header and the signal frame body in certain signal frame configurations. The specific frame header length of the signal frame configurations may also be exploited to identify the configuration and simplify the detection algorithm. Advantages of the method include that it is not sensitive to carrier frequency error or sampling frequency error and is robust and reliable in various wireless channels. Moreover, it entails only a small amount of computation and minimal additional buffer space for intermediate data storage. The power consumption can be made very low, a particular advantage for potable and mobile DTV receivers when they scan DTV programs. In one embodiment, the method only uses amplitude information of the received signals and can in large part share resources with other modules, such as an AGC module.

Description

DETECTION METHOD FOR SIGNAL FRAME CONFIGURATION AND SIGNAL FRAME HEADER FOR BROADCAST SIGNAL

The release of the national specification for terrestrial digital television broadcasting in China in August 2006 is expected to lead to a substantial DTV market in China. The signal for terrestrial digital television broadcasting is physically organized with a signal frame as shown in FIG. 1. A signal frame includes a frame header and a frame body. The frame header and frame body have the same baseband symbol data rate (7.56Msym/sec). Under this frame structure, more than one choice is provided for practical deployment of the terrestrial broadcasting system in diverse environments for various segments of the DTV market.

In the frame body portion, two frame body options are provided in the Chinese DTV terrestrial broadcasting system. One is a multi-carrier option, in which encoded data is transmitted in the frequency domain and the frame body is filled with an OFDM symbol. The multi-carrier option is derived from Chinese DMB-T. The other option is based on single carrier technology, in which encoded data is transmitted in the time domain directly. A particular broadcasting operator can select either one of the options.

In the frame header portion, three possible configurations of the signal frames with different signal frame headers are specified. In the signal frame header, a PN sequence is sent for the purpose of synchronization and channel estimation. At the same time, the PN header also serves as a guard time interval for the following frame body. The possibility of different configurations of the signal frame offer some degree of scalability and flexibility for the broadcasting operator. Three different PN sequences are used in the signal frame header portion for the three configurations. For configurations 1 and 3, a length L_PN-chip (e.g., 420) PN sequence is used that is a cyclic extension of a shorter basic m-sequence with length of N (e.g., 255). For configuration 2, a length LpN-chip (e.g., 595) PN sequence is used that is derived from a longer basic m- sequence with length of N (e.g., 1023). The operator will choose one of the three signal frame configurations according to its specific system deployment plan.

When the receiver begins to scan DTV signals, it does not have any information on what kind of signal frame configuration is used in a specific DTV channel; therefore, it is necessary for a DTV receiver to detect the signal frame configuration before it can start a time and frequency synchronization phase, a channel estimation phase and a data demodulation phase. For a DTV receiver, the capability of scanning DTV programs is required. Detection of the signal frame configuration is one of the most important steps in the procedure of scanning a DTV program.

Since the signal contains a fragment of a known PN sequence, the receiver may correlate the received signal with a local PN sequence and detect the embedded PN sequence. With such traditional correlation methods, in ideal reception conditions and with a reliable channel, the receiver can determine whether a certain PN sequence is present or not. With three parallel or successive PN correlators, with each corresponding to a different PN used in the DTV system, the receiver can determine the configuration of the signal frame. Some problems exist for this method. Firstly, when the receiver starts to receive, some carrier frequency error may exist and may influence the correlation result considerably, especially when the carrier frequency error is large. Secondly, in a multipath channel with long time dispersion, the correlation result will suffer severe interference from other data as well as channel noise. The correlation result may therefore be unreliable.

Low-complexity detection methods and apparatus are provided for detecting signal frame configuration in a DTV receiver or the like. In one embodiment, the detection takes advantage of a difference in average power between the signal frame header and the signal frame body in certain signal frame configurations. The specific frame header length of the signal frame configurations may also be exploited to identify the configuration and simplify the detection algorithm.

More particularly, in accordance with one embodiment of the invention, a method of detecting a signal frame configuration of a broadcast signal includes measuring signal power; calculating a short term average of signal power within each of a plurality of selected portions of the signal to produce a sequence of average power values; performing spectral analysis of the average power values to produce a sequence of transformed values; and using the transformed values, determining that the signal frame configuration is or is not one of a plurality of known possible signal frame configurations.

Advantages of the method include that it is not sensitive to carrier frequency error or sampling frequency error and is robust and reliable in various wireless channels. Moreover, it entails only a small amount of computation and minimal additional buffer space for intermediate data storage. The power consumption can be made very low, a particular advantage for portable and mobile DTV receivers when they scan DTV programs. In one embodiment, the method only uses amplitude information of the received signals and can in large part share resources with other modules, such as an AGC module.

Other features and advantages will be understood upon reading and understanding the detailed description of exemplary embodiments, found herein below, in conjunction with reference to the drawings, a brief description of which is provided below.

FIG. 1 is a diagram of the structure of a DTV signal frame;

FIG. 2 is a block diagram of a DTV receiver to which the present invention may be applied;

FIG. 3 shows the short term average power for a signal with signal frame configuration 1 in an AWGN channel with SNR of OdB;

FIG. 4 shows the spectrum analysis result |Z(m)| (except for the DC value Z(O)) for signal frame configuration 1;

FIG. 5 shows in greater detail the low frequency part of FIG. 4;

FIG. 6 shows the spectrum analysis result |Z(m)| (except for the DC value Z(O)) for signal frame configuration 3;

FIG. 7 shows in greater detail the low frequency part of FIG. 6; - A -

FIG. 8 shows the spectrum analysis result |Z(m)| (except for the DC value Z(O)) for signal frame configuration 2;

FIG. 9 shows a block diagram of a detector for detecting signal frame configuration;

FIG. 10 shows the plot of an additional phase angle of a decision input as a function of a frame header offset at the start of sampling.

There follows a more detailed description of the present invention. Those skilled in the art will realize that the following detailed description is illustrative only and is not intended to be in any way limiting. Other embodiments of the present invention will readily suggest themselves to such skilled persons having the benefit of this disclosure. Reference will now be made in detail to embodiments of the present invention as illustrated in the accompanying drawings. The same reference indicators will be used throughout the drawings and the following detailed description to refer to the same or like parts.

The signal for terrestrial digital television broadcasting in the Chinese DTV terrestrial broadcasting system is physically organized with a signal frame as shown in FIG. 1. In the frame header portion, three possible configurations of the signal frames with different signal frame headers are specified. In the signal frame header, a PN sequence is sent for the purpose of synchronization and channel estimation. At the same time, the PN header also serves as a guard time interval for the following frame body.

The baseband signal sent at the transmitter is

s(t) = ∑x(Vβ (t -iT_s) ®SRRC(t) ( 1 )

where SRRC(t) is a square-root raised cosine filter, T_s is 1/7.56M second and

J PN_f (i-fL_F) when fL_F ≤ i ≤ fL_F +L_PN -1 X^{{l) ~} \d_f {i-fL_F -L_PN) when fL_F +L_PN < i < fL_F + L_F ( 2 ) PN _f (j), j = 0,...,L_PN - 1 is the PN sequence used in the/-th signal frame, with average power of P_PN = E{PN_f ²(i)} . The quantity d_f (j),j = 0,...,L_BODY -1 is the modulated data in the/-th signal frame, with average power of P_B0DY = E{d_f ²(ϊ)} . The PN sequence is derived from an m- sequence with length of N.

The possibility of different configurations of the signal frame offers some degree of scalability and flexibility for the broadcasting operator. The three configurations are listed in Table 1.

Table 1: The three configurations of the signal frame

Three different PN sequences PN_f (j),j = 0,...,L_PN -1 are used in the signal frame header portion for the three configurations. For configuration 1 and 3, the LpN-chip PN sequence is a cyclic extension of the shorter basic m-sequence with length of N. For configuration 2, the L_PN- chip PN sequence is derived from the longer basic m-sequence with length of N.

The operator will choose one of the three signal frame configurations according to its specific system deployment plan.

As described above, when the receiver begins to scan DTV signals, it does not have any information on what kind of signal frame configuration is used in a specific DTV channel; therefore, it is necessary for a DTV receiver to detect the signal frame configuration before it can start a time and frequency synchronization phase, a channel estimation phase and a data demodulation phase. For a DTV receiver, the capability of scanning DTV programs is required. Detection of the signal frame configuration is one of the most important steps in the procedure of scanning a DTV program. The general structure of a DTV receiver is shown in FIG. 2. A signal 203 is received by an RF module 201 and sampled in an analog-to-digital converter ADC 205. The sampled signal 207 is applied to an automatic gain control (AGC) loop 209 and to a signal frame configuration detector 211, which produces configuration information 213. The sampled signal 207 and the configuration information 213 are applied to subsequent modules, such as time/frequency synchronization 215, channel estimation 217, OFDM demodulation (221) and equalization (219) for the multi-carrier option, and time equalization (219) for the single carrier option. In such a DTV receiver, when the desired radio frequency is set and the output of the RF module becomes stable following an AGC procedure, the digital processing module 210 obtains samples of the baseband signal from the ADC. At this moment, one of the most important tasks is to detect the signal frame configuration. Only when the signal frame configuration is known can other modules be started.

Since the signal contains a fragment of a known PN sequence, the receiver may correlate the received signal with a local PN sequence and detect the embedded PN sequence. With such traditional correlation methods, in ideal reception conditions and with a reliable channel, the receiver can determine whether a certain PN sequence is present or not. With three parallel or successive PN correlators, with each corresponding to a different PN used in the DTV system, the receiver can determine the configuration of the signal frame. Some problems exist with this method. Firstly, when the receiver starts to receive, some carrier frequency error may exist and may influence the correlation result considerably, especially when the carrier frequency error is large. Secondly, in a multipath channel with long time dispersion, the correlation result will suffer severe interference from other data as well as channel noise. The correlation result may therefore be unreliable.

The received baseband signal is

r(t) = s(t) ®h(t) + n(t) ( 3 )

L-I where s(t) is the transmit signal given by (1) and h(t) = V (X₂δ (t -T₂) is the channel impulse

2=0 response function. The term n{t) represents AWGN noise. The signal is sampled with a sampling rate of M₁Ar₈ (M₁ can be 1, 2, 4, 8, etc.; often 4 or 8 is chosen). First, the short term average power of the input signal is calculated. To simplify the calculation, the signal is down sampled with sampling rate of 1/(M₂Ts) , where M₂ = 1. The down sampled signal is

r(i) = r(iM₂T_s +τ) ( 4 )

where τ is some random time delay. The short term average power of the input signal is calculated as follows in a window with length of Lw-

L_w can be chosen around 420. This sliding average plays the same role as a low pass filter with length of Lw, which will average out a large part of the channel noise. Since only signal power is taken into account, the method is insensitive to the carrier frequency error.

For signal frame configurations 1 and 3 above, the average power of the signal frame header is 3dB higher than that of the frame body; therefore, a difference in short term average power can be observed around each frame header. For illustration, P(ϊ) is given in FIG. 3, which is simulated in an AWGN channel with OdB SNR using signal frame configuration 1. In a multipath channel, a similar waveform can be observed.

For configurations 1 and 3, periodic peaks can be observed in the short term average power. For configuration 2, since the average power for the frame header portion equals that of the frame body, no periodic pattern is obtained. Additionally, for configurations 1 and 3, the period is somewhat different due to the different lengths of the signal frame header.

By using spectrum analysis, robust detection of the signal frame configuration may be achieved as follows.

P(i) is calculated as follows for C time instants.

Y(k) = P(kD), k = 0,1,..., C -I ( 6 ) where Z⁾= 105 and C =360. D is chosen as Z⁾=I 05 because it is the common factor for the frame header length (in units of Ts) of the signal frame configurations 1 and 3. It is also one of the factors of the frame body length. Since P(J) is already a low pass signal, the down sampling of P{\) with sampling rate 1/(DTs) will not cause much alias overlapping. Due to the large sampling interval, the method is insensitive to the sampling frequency error. A DFT is then performed using the C values:

C-I

Z{m) = γ Y{k)e -j lπ mkIC

( 7 ) i=0

For configuration 1, the signal period is (420+3780) Ts=105*40Ts=40Z>Ts. In the above DFT analysis, a signal with frequency of 1/[(420+378O)Ts] corresponds to the signal Z(m) with the index of

1/[40DTs] _ C 1/[C- DTs] ^" 40 ^~

If a signal with configuration 1 is sent, |Z(9)| and other |Z(9k)| will be conspicuously large compared with other values in the frequency domain except for the DC value. Similarly, for configuration 3, the signal period is (945+3780) Ts=105*45Ts=45Z>Ts, corresponding to the signal Z(m) with the index of

1/[45DTs] _ C _ 1/[C- DTs] ^" 45 ^"

That is to say, if a signal with configuration 3 is sent, |Z(8)| and other |Z(8k)| will be conspicuously large compared with other values in the frequency domain except for the DC value.

FIG. 4 shows the spectrum analysis result |Z(m)| (except for the DC value Z(O)) for signal frame configuration 1. FIG. 5 shows in greater detail the low frequency part of FIG. 4. A peak can be observed in the frequency spectrum having an index of 9. FIG. 6 shows the spectrum analysis result |Z(m)| (except for the DC value Z(O)) for signal frame configuration 3. FIG. 7 shows in greater detail the low frequency part of FIG. 6. A peak can be observed in the frequency spectrum having an index of 8.

If a signal with configuration 2 is sent, there is no obvious peak in the spectrum, as shown in FIG. 8.

According to the above analysis, detection of the signal frame configuration can be made using the amplitude value in frequency indices 8 and 9. One embodiment of a signal frame configuration detector is shown in FIG. 9. The received signal is squared (901), then a short-term averaging operation is performed (903). A number C of short term average value power values is collected (905). In an exemplary embodiment, each average is taken over D samples (DTs). A DFT is performed using the C short term average power values (907). Based on the results of the DFT, a decision is performed (909).

In the decision module, many different algorithms can be used to distinguish the different Z(m) patterns of the different signal frame configurations, for example Algorithm A below:

If |Z(8)|>TH₁₁*ZAVER8 and |Z(9)|< TH₁₂*Z_AVER9

The signal frame configuration is 3. Else if |Z(9)|> TH_n*Z_AvER9 and |Z(8)|<TH₁₂*Z_AVER8

The signal frame configuration is 1. Else

The signal frame configuration may be 2. End

In the foregoing algorithm, Z_AVERS is the average amplitude of Z(m), m=l, 2, ..., 179, except m = 8, 16; Z_AVER9 is the average amplitude of Z(m), m = 1, 2, ..., 179, except m = 9, 18:

1 ^c~1 ZAVER₈ = TT^ ∑ I^ZH ( 8 ) m≠8,16 C L _m=ϊ m≠9,18

THn is a threshold for detection, which may be 6, for example; TH12 is a threshold for detection, which may be 1 , for example.

Of course, other algorithms can be adopted, for example the following Algorithm B:

If |Z(8)| + |Z(16)|> TH₂₁*ZAVER8 and |Z(9)| + |Z(18)|< TH₂₂*Z_AVER9

The signal frame configuration is 3.

Else if |Z(9)| + |Z(18)|> TH₂I*Z_AVER9 and |Z(8)| + |Z(16)|<TH₂₂*Z AVERS

The signal frame configuration is 1.

Else

The signal frame configuration may be 2.

End

TH₂I is a threshold for detection, which may be 8, for example; TH₂₂ is a threshold for detection, which may be 1 , for example.

Various measures may be taken to simplify implementation. In the diagram of the detector in FIG. 9, the square operator and the short term power averaging module are also essential for and may be shared with the AGC module, shown in FIG. 2. It is notable that in the short term power averaging module, no multiplier is needed. For each incoming sampling time instant, the module only needs to update the accumulated result by adding the incoming sample and subtracting the tail sample, previously stored in a loop buffer. Then the tail sample is replaced by the incoming sample and the tail pointer is adjusted at the same time. No extra calculation is needed at this stage.

In the foregoing description, the detector analyzes the C short term power averaging values. An FFT of size of C can be used for this purpose. However, not all the frequency elements are needed in the decision algorithm. Only the elements Z(8) and Z(9) are needed. Of course, to ensure a more reliable decision, elements Z(16), Z(18), etc., can be involved. To calculate the Z(m) separately, the following simple algorithm can be used: (l)S = Y(0),W₀ = W = e ^N (2)For k = \ to C-\,

S ^ S + Y(k) - W ( ⁹ )

W ^ W- W₀ (3)Z(m) = S

The computational cost of this calculation is relatively modest, involving only several hundred complex multiplication and accumulation operations.

Instead of computing Z_AVERS/9 (the average floor power of all the frequency elements except the DC element), another value ZP can be calculated in time domain:

C-I

ZP² = ∑(Y(k)-Y) k=0

( 10 )

— 1 ^c"1 Where Y = - ∑Y(k)

C- i=0

The algorithm in the decision module changes accordingly, as exemplified by the following Algorithm C:

If |Z(8)|² |>TH₃i*(ZP² - |Z(8)|²) and |Z(9)|² |< TH₃₂*(ZP² - |Z(9)|²)

The signal frame configuration is 3. Else if |Z(9)|² |>TH₃*(ZP² - |Z(9)|²) and |Z(8)|² |<TH₃₂*(ZP² - |Z(8)|²)

The signal frame configuration is 1. Else

The signal frame configuration may be 2. End

TH₃] is a threshold for detection, which may be 36, for example; TH₃₂ is a threshold for detection, which may be 6, for example. Alternatively, the following Algorithm D may be used: If |Z(8)|²+|Z(16)|² >TH₄i*(ZP² - |Z(8)|²- |Z(16)|²) and

|<TH₄₂*(ZP -j2 -

The signal frame configuration is 3. Else if |Z(9)|²+|Z(9)|² >TH₄*(ZP²-|Z(9)|²- |Z(18)|²) and |Z(8)|²+|Z(16)|² |<TH₄₂*(ZP² - |Z(8)|²- |Z(16)|²)

The signal frame configuration is 1. Else

The signal frame configuration may be 2. End

TH41 is a threshold for detection, which may be 64, for example; TH42 is a threshold for detection, which may be 8, for example.

The algorithms described may be used to realize an additional beneficial function, namely the determination of approximate frame header positioning for signal frame configurations 1 and 3. If a signal with signal frame configuration 1 or 3 is sent, the short term average power displays periodic characteristics. For configuration 1, the basic period is 4200Ts, whose value is shown in the 9-th frequency element in the spectrum of formula (7). In an AWGN channel, in the C sampled short term average power values, if the first value Y(O) is aligned with the peak of the sampled short term average power (that is to say, Y(O) is aligned with the frame header) values Y(40k) will tend to align with the frame header and have local peak positive values, while other values tend to be small. The value Z(9) can be rewritten as, I ,2π ⁹*^40k - ;2π —

Z(9) = Y(40k)e ³⁶⁰ + Y(i)e ^c _{( n )}

When Y(O) is aligned with the frame header, the first term in (11) will dominate the value of Z(9). Due to all the Y(40k) values being large positive real numbers, the phase angle of Z(9) is almost 0.

If there is a time delay τ from the frame header at the start of sampling (that is to say, Y(O) is taken with a frame header offsetτ , meaning that sampling starts from a point τ after a certain frame header), an additional angle phase will added to Z(9). The factor is:

C 105T, _ ^ ₄₂OO t- — t-

FIG. 10 shows the additional phase due to the delayτ .

Thus, if signal frame configuration 1 is confirmed, Z(9) is calculated by (7) or (9). The approximate header position τ can be estimated as follows: _τ « ^ZZ^ * 42007; ( 12 )

2π

Similarly, if signal frame configuration 3 is confirmed, Z(8) is calculated by (7) or (9). The approximate header position τ can be estimated as follows: τ_;= -^T-I^* 47257 ( 13 )

2π

In the case of an AWGN channel with high SNR, the above two formulas present accurate estimation. However, in multipath channel or low SNR cases, they only can provide an approximate estimation of the signal frame header position. In spite of this, knowledge of the approximate position of the signal frame header will be helpful for the following processing in the digital receiver. For example, when receiver begins to perform time synchronization, it only needs to search a limited time range to find the accurate frame header. Formulas (12) and (13) will reduce considerably the computation burden for the following modules.

Although embodiments of the present invention have been described in detail, it should be understood that various changes, substitutions and alternations can be made without departing from the spirit and scope of the inventions as defined by the appended claims.

Claims

C L A I M SWhat is claimed is:

1. A method of detecting a signal frame configuration of a broadcast signal, comprising: measuring signal power; calculating a short term average of signal power within each of a plurality of selected portions of the signal to produce a sequence of average power values; performing spectral analysis of the average power values to produce a sequence of transformed values; and using the transformed values, determining that the signal frame configuration is or is not one of a plurality of known possible signal frame configurations.

2. The method of Claim 1 , comprising determining that the signal frame configuration is one of a plurality of known possible signal frame configurations.

3. The method of Claim 1 , wherein determining that the signal frame configuration is or is not one of a plurality of known possible signal frame configurations comprises examining periodicity properties of the sequence of transformed values.

4. The method of Claim 3, wherein examining periodicity properties comprises calculating an average of the transformed values excluding selected values of a particular periodicity.

5. The method of Claim 4, wherein examining periodicity properties comprises comparing a value produced using an average of the transformed values excluding selected values of a particular periodicity to a value produced using one or more excluded values.

6. The method of Claim 5, wherein, if for a first particular periodicity the value produced using one or more excluded values exceeds the value produced using an average of the transformed values excluding selected values of the first particular periodicity, the signal frame configuration is determined to be a first one of the plurality of known possible signal frame configurations.

7. The method of Claim 6, wherein, if for a second particular periodicity the value produced using one or more excluded values exceeds the value produced using an average of the transformed values excluding selected values of the second particular periodicity, the signal frame configuration is determined to be a second one of the plurality of known possible signal frame configurations.

8. The method of Claim 7, wherein, if the signal frame configuration is determined not to be either a first or a second one of the plurality of known possible signal frame configurations, it is determined to likely be a third one of the plurality of known possible signal frame configurations.

9. The method of Claim 1, wherein performing spectral analysis comprises calculating a discrete fourier transform of the sequence of average power values.

10. An apparatus for detecting a signal frame configuration of a broadcast signal, comprising: means for measuring signal power; means for calculating a short term average of signal power within each of a plurality of selected portions of the signal to produce a sequence of average power values; means for performing spectral analysis of the average power values to produce a sequence of transformed values; and means for, using the transformed values, determining that the signal frame configuration is or is not one of a plurality of known possible signal frame configurations.

11. The apparatus of Claim 10, wherein the means for determining determines that the signal frame configuration is one of a plurality of known possible signal frame configurations.

12. The apparatus of Claim 10, wherein the means for determining that the signal frame configuration is or is not one of a plurality of known possible signal frame configurations examines periodicity properties of the sequence of transformed values.

13. The apparatus of Claim 12, wherein examining periodicity properties comprises calculating an average of the transformed values excluding selected values of a particular periodicity.

14. The apparatus of Claim 13, wherein examining periodicity properties comprises comparing a value produced using an average of the transformed values excluding selected values of a particular periodicity to a value produced using one or more excluded values.

15. The apparatus of Claim 14, wherein, if for a first particular periodicity the value produced using one or more excluded values exceeds the value produced using an average of the transformed values excluding selected values of the first particular periodicity, the signal frame configuration is determined to be a first one of the plurality of known possible signal frame configurations.

16. The apparatus of Claim 15, wherein, if for a second particular periodicity the value produced using one or more excluded values exceeds the value produced using an average of the transformed values excluding selected values of the second particular periodicity, the signal frame configuration is determined to be a second one of the plurality of known possible signal frame configurations.

17. The apparatus of Claim 16, wherein, if the signal frame configuration is determined not to be either a first or a second one of the plurality of known possible signal frame confϊgurations, it is deteπnined to likely be a third one of the plurality of known possible signal frame configurations.

18. The apparatus of Claim 9, wherein the means for performing spectral analysis calculates a discrete fourier transform of the sequence of average power values.