This invention pertains to the field of signal detection, and more particularly to a method and apparatus for detecting the presence of a digital television (DTV) signal.

On 25 May 2004, the United States Federal Communication Commission (FCC) issued a Notice of Proposed Rulemaking (NPRM) (FCC 04-113) in ET Docket No. 04-186 to allow unlicensed radio transmitters to operate within the broadcast television spectrum at locations where one or more of the allocated terrestrial television channels are not being used. However, the FCC stressed that such unlicensed transmitters would only be permitted with safeguards that insure no interference with the reception of licensed terrestrial television signals. Therefore, to prevent interference with terrestrial television service, it is important to insure that any such unlicensed transmitters do not operate on any frequencies or channels where a terrestrial television signal could otherwise be received and viewed in the same area.

Accordingly, in order to ensure that no interference is caused to TV stations and their viewers, the Commission proposed to require that these unlicensed transmitters incorporate the capability to identify unused or vacant TV channels and to only transmit on such vacant channels. One idea advanced by the FCC would be to incorporate sensing capabilities in the unlicensed transmitter to detect whether other transmitters (i.e., licensed terrestrial TV broadcast transmitters) are operating on a particular channel in the area before the unlicensed transmitter could be activated. For example, a fixed unlicensed transmitter could be required to incorporate an antenna and a receiver capable of detecting whether a particular terrestrial TV channel is actually in use in the area where the transmitter is to be installed. If the antenna and receiver detect a terrestrial television broadcast signal on a particular channel, then the unlicensed transmitter would be prevented from using that channel. The unlicensed transmitter would only be permitted to operate on a particular channel in a particular location if the antenna and receiver verify that no terrestrial television broadcast signal is present on that channel.

However, such sensing would have to be much more sensitive than the receivers normally used to receive the terrestrial TV signal and decode the audio and video signals. That is because the terrestrial TV signal is subject to fading, structural blockage, etc., and therefore it is possible that the measurements might be made at a location where the terrestrial television signal is lower than in surrounding areas.

Accordingly, it would be desirable to provide a method and apparatus for detecting the presence of a terrestrial television signal, particularly, a terrestrial digital television (DTV) signal. It also would be desirable to provide such a method and apparatus which can detect such signals at low power levels, well below the level necessary for a viewable picture. It would further be desirable to detect a DTV signal without requiring synchronization to the DTV signal. It would still further be desirable to provide such a method and apparatus which can be provided at low cost.

In one aspect of the invention, a method of detecting the presence of a digital television signal comprises sampling a received signal; performing at least one autocorrelation of the sampled received signal over a correlation window comprising N samples to produce autocorrelation results, wherein the autocorrelation is performed with a delay that is an integer multiple of a period between transmissions of a training sequence within a digital television signal; and comparing the autocorrelation results to a threshold and determining that a digital television signal is present when one of the autocorrelation results is greater than the threshold.

In another aspect of the invention, a receiver for detecting the presence of a digital television signal comprises a sampler adapted to sample a received signal at a set rate; an autocorrelator adapted to autocorrelate the sampled received signal to produce autocorrelation results, wherein the autocorrelation is performed with a delay that is an integer multiple of a period between transmissions of a training sequence within a digital television signal; and a comparator adapted to compare the autocorrelation results to a threshold and to indicate detection of a digital television signal when one of the autocorrelation results is greater than the threshold.

FIG. 1 shows the frame structure of a digital television (DTV) signal; FIG. 2 shows a flowchart of a method of detecting the presence of a DTV signal;

FIG. 3 shows a block diagram of a receiver capable of detecting the presence of a DTV signal;

FIG. 4 shows a block diagram of a first embodiment of a correlator that can be used

in the receiver of FIG. 3;

FIG. 5 shows a block diagram of a second embodiment of a correlator that can be used in the receiver of FIG. 3;

FIG. 6 shows a simplified block diagram of one embodiment of a portion of the receiver of FIG. 3.

FIG. 7 shows the output of an embodiment of the correlator of FIG. 4 where the input signal to noise ratio is -5dB;

FIG. 8 shows the output of an embodiment of the correlator of FIG. 4 in the presence of a strong multipath "ghost" signal where the input signal to noise ratio is -5dB.

FIG. 1 shows the frame structure of a digital television (DTV) signal according to the North American DTV standard. A frame consists of 313 segments. Each segment consists of 832 symbols. Within each frame, a training sequence is transmitted. This is also commonly known as a field sync. Accordingly, the training sequence is a periodic signal within the DTV signal, transmitted once every 24.2 ms.

After downconversion and digitization (sampling), the received DTV signal in the presence of noise can be modeled according to equation (1):

(1) r(mT) - ar(rø J)e ^**v - ^{vKi ■} * n{mT)

where T is the sampling rate, f ₀ is the carrier frequency error, and n(mT) is the noise component. Here, the sampling rate T is not necessarily identical to the sampling rate of the DTV signal. In fact, such a feature would lead to simpler implementation since the detector does not have to synchronize with the DTV signal. Meanwhile, a(mT) is the received signal component described by equation (2):

(2) a(mT) ^™ (x§§ h}{mT}

where x is the transmitted DTV information and h is the terrestrial channel impulse response. The channel impulse response can be time-variant. Nevertheless, for the purpose of the following analysis, it is assumed that the time- variation within the detection window will be minimal.

As mentioned earlier, the DTV signal periodically sends a known training sequence every 24.2ms.

In a conventional DTV receiver, detection of this periodic component (the training sequence) is typically accomplished by correlating the received signal with the local replica of the training sequence. However, this method requires almost perfect timing and carrier frequency synchronization with the transmitted signal. In practice, synchronization of the DTV signal has been found to be problematic in the presence of frequency-selective fading. It is also computationally expensive.

Advantageously, FIG. 2 shows a flowchart of a robust method of detecting the presence of a DTV signal by use of this periodic signal, without requiring precise timing and carrier frequency synchronization.

In a first step 210, a receiver tunes to a channel where it is desired to determine whether or not a DTV signal is present.

In a second step 220, the receiver receives whatever signal (including noise) is present in the channel, and downconverts the received signal such that the channel can be digitally sampled. Alternatively, as technology develops, it may be possible to sample the channel directly at RF frequency.

In a third step 230, the signal is digitally sampled at a sampling rate, 1/T. The sampling rate may be the same as, or approximately the same as, the data rate of the bits in the training sequence, but this is not necessary. Beneficially, the sampling rate 1/T > 10.6 Msamples/sec (T < 94.3 nsec).

In a fourth step 240, the sampled received signal is autocorrelated over a window on N samples to produce an autocorrelation result, according to equation (3):

N -J (3) f(m) ~ ]T r(τn ÷ &)*' ^* (« + £ ^} + k)

where r(n) is the received samples, D is the delay (D ~ 24.2ms/T), and N is the correlation window. In one embodiment, N is 500.

Substituting equation (1) into equation (3), and simplifying, produces equation (4):

N _j H

4) f(m) = e ^'u ' ^D Jjtϊ(«ι + k)a (m + £> + £) + Z(m)

where Z(m) is the cross-correlation of the noise term and the term representing any received DTV signal.

As can be seen from equation (4), advantageously the delay used in the autocorrelation is equal, or approximately equal, to the period between transmissions of the training sequence in the DTV signal. In other words, the autocorrelation determines the degree of correlation of the sampled, received signal with a delayed version of itself, where the delay is equal to the period between training sequence transmissions. Under the assumption that the channel is quasi-static within 24.2ms, the periodic components of the DTV signal every 24.2ms are almost the same. Accordingly, if a DTV signal is present, the autocorrelation result of equation (4) will show a large peak every 24.2ms.

Furthermore, it will be shown in simulations below that even if the channel is assumed to be time-varying, it is still possible to robustly detect the presence of the DTV signal. From equation (4), one can see that the carrier frequency error or sampling error does not affect the autocorrelation result. Advantageously, this leads to simplification of the implementation, eliminating the need for precise frequency or timing synchronization. In a fifth step 250, the correlation results of equation (4) are compared at each sample interval T with a threshold, X, for determining whether a digital television signal is present. That is, so long as the threshold X is greater than the autocorrelation results f(m), no DTV signal is detected. However, if and when an autocorrelation result f(m) is greater than X, then it is determined that a DTV signal is present and has been detected. The threshold value, X, is a function of the noise level, the AGC setting and expected signal strength. Usually, the AGC is maximized to catch the weak signals. The main idea is to detect the existence of a correlation peak with respect to the correlation noise floor. Beneficially, during the burst detection step, the AGC is set to a fixed value so that it does not affect the operation. Usually, it is set to the maximum value to catch weak signals.

Beneficially, the threshold X may be determined by calculating a moving average of all of the autocorrelation results f(m) over a moving window, M, larger than the correlation window N. The correlation output is passed through a window averaging unit that can be implemented using a simple low pass filter. This may be achieved by comparing the correlation peak over a certain window to the average correlation output as discussed in further detail with respect to FIG. 6 below.

If valid signal is present, this scheme will normally provide detection within 1 frame symbol period (i.e., 24.2ms).

If a DTV signal is detected, then the process returns to step 210, tunes to a new channel, and repeats the process until all permissable DTV channels have been checked. If a DTV signal is not detected, then the steps 240 and 250 are repeated until a set time period for DTV signal detection has expired without detecting a DTV signal. In that case, in a step 260, it is determined that the channel does not have a DTV signal at that location. In that case, in a step 270 it is determined whether all available DTV channels have been checked. If so, then the process ends. If not, then the process returns to the step 210, tunes to a new channel, and repeats the process until all permissible DTV channels have been checked. FIG. 3 shows a block diagram of a receiver 300 capable of detecting the presence of a DTV signal by performing an autocorrelation to take advance of the periodic transmission of the training sequence in the DTV signal.

The receiver 300 includes a tuner 310, a downconverter 320, a signal sampler 330, an autocorrelator 340, and a comparator 350. The operation of the receiver 300 has been explained above with respect to FIG. 2 and for conciseness will not be repeated.

FIG. 4 shows a block diagram of a first embodiment of an autocorrelator 400 that can be used in the receiver of FIG. 3. The autocorrelator 400 includes a D-length delay 410, a multiplier 420, an N-length delay 430, a summation circuit 440, and a T delay 450. The autocorrelator 400 implements the equation (4) above, and can be implemented at relatively low cost.

The performance of the basic autocorrelation algorithm can be improved by employing multiple levels of hierarchical delayed correlations. These multi- level hierarchical correlations can be described according to equation (5):

(5) f(m) ~ > > ^' r(m -ι- D! + k)r (m + D(I ÷ /> + k)

FIG. 5 shows a block diagram of a second embodiment of an autocorrelator 500 that can be used in the receiver of FIG. 3 and that employs multiple (e.g., two) autocorrelation branches. The autocorrelator 500 includes: correlation branches 505, each including a D-length delay 510 and a multiplier 520; a correlation branch output summer 525; an N-length delay 530; a summation circuit 540; and a T delay 550. The autocorrelator 500 implements the equation (5) above. Of course, more than two correlation branches 505 can be employed.

FIG. 6 shows a simplified block diagram of one embodiment of a portion of the receiver of FIG. 3, including the autocorrelator 340 and comparator 350. In the embodiment of FIG. 6, the comparator 350 includes a window averaging unit 610 adapted to average the autocorrelation results over a moving window, M, and to output the threshold, X, to a decision circuit 620. Beneficially, the size of the moving window M is larger than the size of the autocorrelation window, N. According to the embodiment of FIG. 6, the threshold, X, for determining whether or not a DTV signal is present, is set as an average of all of the autocorrelation results over the moving window, M. Beneficially, the window averaging unit 610 may be a low pass filter. In order to evaluate the performance of the method and system described above, simulations were performed using a model for the DTV signal. The simulation was carried using a 1-bit delay unit for the buffer denoted as D-delay. The total size of this buffer is thus 832*313*2 bits. Modern CMOS technology allows easy integration of such memory into the digital IC. FIG. 7 shows the simulated output of an embodiment of the autocorrelator of FIG. 4 where the input signal-to-noise ratio is -5dB. As FIG. 7 illustrates, the presence of a DTV signal can be detected reliably. The threshold of visibility for DTV signals is about 15dB SNR. This means that the simulated signal SNR is 2OdB worse. This 2OdB SNR difference will provide sufficient margin for practical agile radio systems. For example, other things being equal, the signal power at the receiver making the detection can be lower by 2OdB while the signal power at the nearby DTV can be the minimum that the TV receiver can reliably decode. Thus, this 2OdB difference can be margin that can account for, e.g., deep fading.

As a second example, a simulation was performed by adding a strong (OdB) echo at 1 μs delay. This echo was subjected to 5Hz Doppler frequency to account for channel variation.

FIG. 8 shows the simulated output of an embodiment of the correlator of FIG. 4 in the presence of a strong multipath "ghost" signal where the input signal to noise ratio is again -5dB. FIG. 8 clearly demonstrates that the DTV signal can still be detected even if it is subject to such multipath conditions. A conventional DTV receiver would require 22- 3OdB SNR to decode such a signal. This indicates that the receiver of FIG. 3 is working at least 27dB below the requirement for a DTV receiver, hence a 27dB margin.

Accordingly, the above-described method and apparatus can detect a terrestrial

digital television (DTV) signal at low power levels well below the level necessary for a viewable picture. They do so without requiring synchronization to the DTV signal, and can be implemented at relatively low cost.

While preferred embodiments are disclosed herein, many variations are possible which remain within the concept and scope of the invention. Such variations would become clear to one of ordinary skill in the art after inspection of the specification, drawings and claims herein. The invention therefore is not to be restricted except within the spirit and scope of the appended claims.