JP5660444B2 - Spectrum detection apparatus and spectrum detection method - Google Patents

Description

  The present invention relates to an apparatus and method for detecting a spectrum, and more particularly to an apparatus and method for detecting a spectrum suitable for detecting a spectrum of a television signal.

  Wireless communication standards (for example, IEEE 802, ECMA, IEEE SCC41, etc.) that use the same frequency band as the licensed television broadcasting and that can be operated without a license are being studied. In these standards, a TV user is a primary user, and only when the licensed frequency band of the primary user is not used, the secondary user is allowed to use the frequency band (unused licensed frequency band = white space). Allow to operate wireless communication. Therefore, these license-free wireless communications are premised on first detecting that the licensed frequency band is not used by the primary user.

  In order to detect that it is not used by the primary user, the operation of the wireless communication performed by the secondary user based on the result is such that the primary user does not affect the primary user at all. It is necessary to be able to detect a very low power that cannot be detected (for example, −120 dBm (8 MHz band)). In addition, it is important to be able to detect the detection within a reasonable time and to be highly feasible in terms of cost.

  In 2007, it was applied to the detection of television signals of NTSC (Analog TV standards such as the United States) and ATSC (Digital TV standards such as the United States), for example, a test capable of detecting power of about −114 dBm (6 MHz band). The work is being tested. This method of detecting the spectrum of an NTSC signal uses spectrum cross-correlation and energy detection. There is no report on television signals of PAL (analog TV standards such as Europe except France) and SECAM (analog TV standards such as France).

  Regarding the detection method in DVB-T (digital TV standard such as Europe), there are reports of Non-Patent Documents 1 to 4 below. Non-Patent Documents 1 and 2 use the periodic stationary characteristics of the signal, but when the SNR is worse than -10 dB, the problem is not robust but remains. Non-Patent Document 3 uses a correlation output of a received TV signal with a pilot sequence (a known sequence used in a DVB-T signal) and compares it with a preset threshold (threshold) to make a determination. . In Non-Patent Document 4, since the reliability of detection based on the correlation is improved by performing the frequency offset estimation and the signal power averaging before taking the correlation, a better result can be obtained.

  However, the remaining points are as follows. 1) Based on the maximum value of the correlation output (whether or not there is a TV signal), especially when the TV signal is not broadcast, the correlation output including the maximum power noise is used. As a result, the detection performance can be substantially lowered. 2) Since it is necessary to correlate the pilot signal with respect to the entire period of the received signal, the calculation burden is imposed on the system, and the system becomes complicated and it is difficult to reduce the cost. 3) Since time synchronization is not applied, the detection performance can be substantially reduced. 4) Since the detection performance depends on the averaging of the signal power, it is necessary to buffer a very large number of consecutive symbols, which requires cost and processing time. 5) The phase deviation of the averaged DVB-T signal with respect to the pilot signal may substantially affect the result after cross-correlation, thereby degrading the detection reliability.

US Pat. No. 5,596,422 US Patent Application Publication No. 2009/0143019 US Patent Application Publication No. 2009/0102981 US Patent Application Publication No. 2009/0028256 US Patent Application Publication No. 2010/0035568 US Patent Application Publication No. 2007/0157269 US Patent Application Publication No. 2008/0118006

Chen, Gao, Daut, “Spectrum sensing using cyclostationary properties and application to IEEE 802.22 WRAN”, Proc. IEEE GLOBECOM 2007, November 2007, PP3133-3138 Lin, et al., “A Cyclostationary-Based Spectrum sensing Method Using Stochastic Resonance in Cognitive Radio”, Proc. IEEE ICC 2010, May 2010, PP1-5 Danyo Danev, “On Signal Detection Technique for the DVB-T Standard”, Proc. ISCCSP 2010, March 2010, PP1-5 Vasanth Godam and Monisha Ghosh, “Robust Sensing of DVB-T Signals”, IEEE DySPAN 2010, April 2010, PP1-8 IEEE P802.22 standard

  The present invention has been made in consideration of the above-described circumstances, and can detect a TV signal with a very low power level even at a low SNR, while taking into account the reduction of processing time, low cost, and avoidance of complexity. An object of the present invention is to provide an apparatus and method for detecting a spectrum.

In order to solve the above-described problem, a spectrum detection apparatus according to a first aspect of the present invention receives a first signal as an input and has a pass characteristic in a frequency band where a detection target signal can exist, and receives a second signal. A first filter for output, an AD converter connected to the output side of the first filter for converting to a digital value and outputting a third signal, and connected to the output side of the AD converter; Sliding correlation between the third signal and a reference signal having periodicity similar to that of the detection target signal is calculated, and sliding is performed using a sliding value range in which each obtained correlation value is close to the maximum. The correlation is partially recalculated, and the time value of the third signal with respect to the reference signal is calculated using the range of sliding values where the obtained correlation values are close to the maximum. Set estimates, and time offset compensation unit for outputting a fourth signal correction in the time offset is performed, it is connected to the output side of the time offset correction unit, and outputs a fifth signal of buffered A correlation calculation unit that calculates a correlation between the first buffer and the output side of the first buffer and outputs a sixth signal and is connected to an output side of the correlation calculation unit And a determination unit that determines whether the detection target signal is present based on a comparison with a threshold value and outputs a determination result signal.

  According to this, since the received signal has time synchronization with the reference signal by the time offset correction unit, the calculated correlation value is difficult to be buried in the noise and becomes (almost) the maximum, and the detection performance is remarkably improved. To do.

In addition, the spectrum detection apparatus according to the second aspect (reference aspect) has a pass characteristic in a frequency band in which a detection target signal may exist, and receives a first signal as an input and outputs a second signal; An AD converter connected to the output side of the filter that converts to a digital value and outputs a third signal, and a fourth signal buffered and connected to the output side of the AD converter Multiplying the output buffer by a reference signal connected to the output side of the buffer and having a periodicity similar to that of the detection target signal, and outputting the absolute value of the multiplication result as a fifth signal A multiplier that is connected to an output side of the multiplier, performs an integration operation according to the length of the reference signal, and outputs a sixth signal, and an output of the integrator Connected to, it comprises a determination section for outputting a determination result signal by the determination of Yes No of the detection target signal from a comparison between a threshold value.

  According to this, only one of the received signals having the I signal and the Q signal is used for detection, and even when a phase deviation occurs with respect to the reference signal, the multiplication unit and the integration unit that output the absolute value are provided. By providing, the reliability of detection can be maintained.

Further, the spectrum detection apparatus according to the third aspect (reference aspect) has a pass characteristic in a frequency band in which a detection target signal can exist, and receives a first signal as an input and outputs a second signal; An AD converter connected to the output side of the filter that converts to a digital value and outputs a third signal, and a fourth signal buffered and connected to the output side of the AD converter The sliding correlation between the first buffer to be output and the reference signal connected to the output side of the first buffer and having the same periodicity as the signal to be detected is calculated to obtain each correlation value. A sliding correlation calculation unit for outputting a sixth signal; and a second output connected to the output side of the sliding correlation calculation unit for outputting a buffered seventh signal. A buffer, an averaging unit connected to the output side of the second buffer and outputting an averaged eighth signal, and an output connected to the output side of the averaging unit for conversion to an absolute value The absolute value output calculation unit that outputs the ninth signal by performing the operation, and is connected to the output side of the absolute value output calculation unit, and selects the maximum value among the correlation values that have been averaged and converted into absolute values Then, a determination result is obtained by determining the presence or absence of the detection target signal from a comparison between a peak correlation output detection unit that outputs the tenth signal and a threshold value connected to the output side of the peak correlation output detection unit And a determination unit that outputs a signal.

  According to this, the calculation result of the sliding correlation is less likely to be buried in noise by the second buffer and the averaging unit, and the reliability is high. In addition, since the absolute value output calculation unit is provided, only one of the received signals having the I signal and the Q signal is used for detection, and even if a phase deviation with respect to the reference signal occurs, the reliability of the detection is improved. Can maintain sex.

In addition, the spectrum detection apparatus according to the fourth aspect (reference aspect) has a pass characteristic in a frequency band in which a detection target signal may exist, receives a first signal as an input, and outputs a second signal; An AD converter connected to the output side of the filter that converts to a digital value and outputs a third signal, and a fourth signal buffered and connected to the output side of the AD converter The sliding correlation between the first buffer to be output and the reference signal connected to the output side of the first buffer and having the same periodicity as the signal to be detected is calculated to obtain each correlation value. A sliding correlation calculation unit for outputting a sixth signal; and a second output connected to the output side of the sliding correlation calculation unit for outputting a buffered seventh signal. A buffer, connected to the output side of the second buffer, connected to the output side of the absolute value output calculation unit, an absolute value output calculation unit that converts to an absolute value and outputs an eighth signal, and An averaging unit that outputs the averaged ninth signal; and an output unit connected to the output side of the averaging unit to select the maximum value of the correlation values that have averaged the absolute values. A determination result signal is obtained by determining the presence or absence of the detection target signal from a comparison between a peak correlation output detection unit that outputs a tenth signal and a threshold value connected to an output side of the peak correlation output detection unit. And a determination unit for outputting.

  According to this, by providing an absolute value output calculation unit and an averaging unit that collects and averages the outputs, only one of the received signals having the I signal and the Q signal is used for detection. Even when a phase deviation occurs with respect to the reference signal, the reliability of detection can be maintained.

In addition, the spectrum detection apparatus according to the fifth aspect (reference aspect) receives the first signal as an input and outputs the second signal having pass characteristics in a frequency band in which the detection target signal can exist. And a second filter connected to the output side of the first filter and having a narrower band characteristic than the first filter and outputting a third signal filtered with the narrowband characteristic. An AD converter connected to the output side of the second filter for converting to a digital value and outputting a fourth signal, and a buffered first converter connected to the output side of the AD converter. 5 and a reference signal connected to the output side of the buffer and having a periodicity similar to the periodicity of the signal to be detected is calculated to output a sixth signal. A correlation calculation unit for, connected to said output side of the correlation calculating unit comprises a determination unit for outputting a determination result signal by the determination of Yes No of the detection target signal from a comparison between a threshold value.

  According to this, since the target frequency of the received signal to be correlated is limited by providing the narrow-band second filter, the band that does not affect the detection is removed, thereby reducing the influence of noise. If there is a signal to be detected, the presence is easily reflected in the correlation calculation result. Therefore, the detection performance can be improved.

In addition, the spectrum detection apparatus according to the sixth aspect (reference aspect) has a pass characteristic in a frequency band in which a detection target signal may exist, receives a first signal as an input, and outputs a second signal; An AD converter connected to the output side of the filter that converts to a digital value and outputs a third signal, and a fourth signal buffered and connected to the output side of the AD converter The fifth signal is calculated by calculating the correlation between the output first buffer and the reference signal connected to the output side of the first buffer and having the same periodicity as the detection target signal. A correlation calculation unit for outputting, a second buffer connected to the output side of the correlation calculation unit for outputting a buffered sixth signal, and connected to an output side of the second buffer; A determination result signal obtained by determining whether or not the detection target signal is present from a comparison between an averaging unit that outputs the averaged seventh signal and a threshold value connected to the output side of the averaging unit And a determination unit that outputs.

  According to this, since the buffer and the averaging unit are provided, the calculation result of the correlation is less likely to be buried in noise and has high reliability. Therefore, the detection performance can be improved.

In addition, the spectrum detection apparatus according to the seventh aspect (reference aspect) has a pass characteristic in a frequency band in which a detection target signal can exist, receives the first signal as an input, and outputs the second signal. And a second filter connected to the output side of the first filter and having a narrower band characteristic than the first filter and outputting a third signal filtered with the narrowband characteristic. An AD converter connected to the output side of the second filter for converting to a digital value and outputting a fourth signal; and a square addition operation connected to the output side of the AD converter A square addition unit that outputs a fifth signal, and a determination unit that is connected to the output side of the square addition unit and determines whether or not the detection target signal is present from comparison with a threshold value and outputs a determination result signal Comprising a.

  According to this, since the target frequency of the reception signal to be square-added is limited by providing the narrow-band second filter, it is easily reflected in the square addition result depending on the presence of the detection target signal. Become. Therefore, the detection performance can be improved.

In addition, the spectrum detection apparatus according to the eighth aspect (reference aspect) has a pass characteristic in a frequency band in which a detection target signal may exist, and receives a first signal as an input and outputs a second signal; An AD converter connected to the output side of the filter that converts to a digital value and outputs a third signal; and an averaged fourth signal connected to the output side of the AD converter An output averaging unit; a square addition unit connected to the output side of the averaging unit for calculating a square addition and outputting a fifth signal; and a threshold connected to the output side of the square addition unit A determination unit that determines whether or not the detection target signal is present based on a comparison with a value and outputs a determination result signal.

  According to this, since the averaging unit is provided, the signal is less likely to be buried in noise and becomes highly reliable. Therefore, the detection performance can be improved.

In addition, the spectrum detection apparatus according to the ninth aspect (reference aspect) has a pass characteristic in a frequency band in which a detection target signal can exist, receives the first signal as an input, and outputs the second signal. And a second filter connected to the output side of the first filter and having a narrower band characteristic than the first filter and outputting a third signal filtered with the narrowband characteristic. An AD converter connected to the output side of the second filter for converting to a digital value and outputting a fourth signal; and an absolute value addition operation connected to the output side of the AD converter. The absolute value adding unit that outputs the fifth signal and the threshold value connected to the output side of the absolute value adding unit is used to determine the presence / absence of the detection target signal and output a determination result signal Do ; And a tough.

  According to this, since the target frequency of the reception signal to be subjected to the absolute value addition is limited by providing the narrow-band second filter, it is reflected in the absolute value addition result according to the presence of the detection target signal. It becomes easy to be done. Therefore, the detection performance can be improved.

Further, the spectrum detection apparatus according to the tenth aspect (reference aspect) has a pass characteristic in a frequency band in which a detection target signal can exist, and receives a first signal as an input and outputs a second signal; An AD converter connected to the output side of the filter that converts to a digital value and outputs a third signal; and an averaged fourth signal connected to the output side of the AD converter An output averaging unit connected to the output side of the averaging unit, an absolute value addition unit that performs an absolute value addition operation and outputs a fifth signal, and an output side of the absolute value addition unit And a determination unit that determines the presence / absence of the detection target signal based on a comparison with a threshold value and outputs a determination result signal.

  According to this, since the averaging unit is provided, the signal is less likely to be buried in noise and becomes highly reliable. Therefore, the detection performance can be improved.

  The spectrum detection method according to the eleventh aspect obtains a second signal by filtering the first signal, which is a received signal, with a filter having a pass characteristic in a frequency band in which a detection target signal can exist. An AD converter provided on the output side performs conversion to a digital value to obtain a third signal, and a time offset correction unit provided on the output side of the AD converter performs the third signal and the detection target signal. Calculate the sliding correlation with a reference signal that has the same periodicity as that of the periodicity, and partially recalculate the sliding correlation using the range of sliding values where the obtained correlation values are close to the maximum. A time offset between the third signal and the reference signal is estimated using a sliding value range in which each correlation value is close to the maximum. A time offset correction is performed using the time offset to obtain a fourth signal, a buffer provided on the output side of the time offset correction unit obtains a buffered fifth signal, and the buffer output The correlation calculation unit provided on the side calculates the correlation with the reference signal to obtain a sixth signal, and the determination unit provided on the output side of the correlation calculation unit compares with the threshold value. The presence / absence of the detection target signal is determined.

  According to this, since the received signal has time synchronization with the reference signal by the time offset correction unit, the calculated correlation value is hardly buried in the noise and maximized, and the detection performance is substantially improved.

  According to the present invention, there is provided a spectrum detecting apparatus and method capable of detecting a TV signal having a very low power level even at a low SNR, while taking into consideration reduction of processing time, low cost, and avoidance of complexity. Can be provided.

Regardless of the description below, FIG. 5, FIG. 6, FIG. 8, FIG. 9, FIG. 10, FIG. 12, FIG. is there. However, items to be referred to as embodiments of the invention are included.
The block diagram which shows the structure of the spectrum detection apparatus which is one Embodiment (1st Embodiment). The table | surface which shows the subcarrier position of the continuous pilot (CP) in 2K mode of DVB-T. The wave form diagram which shows an example (in the case of PAL etc.) of the reference sequence which can be used by each embodiment. The block diagram which shows the structure of the spectrum detection apparatus which is another (2nd) embodiment. The block diagram which shows the structure of the spectrum detection apparatus which is another (3rd) embodiment. The block diagram which shows the structure of the spectrum detection apparatus which is another (4th) embodiment. The block diagram which shows the structure of the spectrum detection apparatus which is another (5th) embodiment. The block diagram which shows the structure of the spectrum detection apparatus which is another (6th) embodiment. The block diagram which shows the structure of the spectrum detection apparatus which is another (seventh) embodiment. The block diagram which shows the structure of the spectrum detection apparatus which is another (8th) embodiment. The block diagram which shows the structure of the spectrum detection apparatus which is another (9th) embodiment. The block diagram which shows the structure of the spectrum detection apparatus which is another (10th) embodiment. The block diagram which shows the structure of the spectrum detection apparatus which is another (11th) embodiment. The block diagram which shows the structure of the spectrum detection apparatus which is another (12th) embodiment. The block diagram which shows the structure of the spectrum detection apparatus which is another (13th) embodiment. The block diagram which shows the structural example of the frequency offset correction | amendment part which can be used by 1st thru | or 11th Embodiment.

  Based on the above, embodiments of the present invention will be described below with reference to the drawings. First, the principle of the main components that can be used in each embodiment will be comprehensively described. After that, individual examples will be described.

(Time offset estimation and correction)
The sliding correlation between the received signal and the reference sequence is calculated, and the time offset is estimated and corrected according to the following steps.
1. Find the sliding correlation output position m (sequence order) in the peak and record it (no decision). This is repeated M times (M> 1) at an interval divided by the repetition time of the detection target signal (DVB-T signal). Based on the obtained distribution of M positions m, a time offset range is estimated as a range of m having a high occurrence frequency.
2. Among the outputs of some of the sliding correlations, a peak position m is found and recorded (not determined). Some sliding correlations here are: Is determined based on the result of estimating the time offset range in (i.e., the range of m having a high occurrence frequency in 1.) is used. This is repeated K times (K> 1) at an interval divided by the repetition time of the detection target signal (DVB-T signal). Then, based on the obtained distribution of K positions m, the time offset range is estimated again as a range of m having a high occurrence frequency.
3. If accuracy is required for the accuracy of the time offset, further some sliding correlations are performed. Do the same. Here, a part of the sliding correlation is determined based on the previous time offset range estimation result (that is, the range of m having a high occurrence frequency in the previous time is used).
4). The received signal is time-corrected (time-shifted) based on the estimated time offset value.
Estimating and correcting the time offset is one feature point in the embodiment.

(Decision based on the calculation result of selective (= partial) sliding correlation)
The sliding correlation is calculated selectively (partially), and using the calculation result, it is determined whether or not the detection target signal exists in the received signal.

Pure TV signal x _0, the result of taking the correlation between this and time without offset reference sequence ref is larger than when the time of the offset, the maximum (peak value). Assume that a correlation result equivalent to (similar to) the same is obtained, and it is determined whether the threshold has been reached. Received TV signal x is composed of a pure TV signal _{x 0} and noise n. The sliding correlation between the received signal and the reference sequence ref can be expressed as follows.
When the TV signal _{x 0} does not exist:
R _{n · ref} (m) = Σ _{i = 0} ^N−1 ref (i) n (i + m),
i = 0, 1,..., N-1; m = 0, 1,.
... (1)
When the TV signal _{x 0} is present:
R _{x · ref} (m)
= R _{x0 · ref} (m) + R _{n · ref} (m)
= Σ _{i = 0} ^N−1 ref (i) x ₀ (i + m) + Σ _{i = 0} ^N−1 ref (i) n (i + m)
... (2)
Here, N is the number of samples of the reference sequence that is a repetitive pattern, m is a sliding value represented by an integer of 0 or more and less than N, and i is an addition parameter for Σ.

A time offset with respect to the reference sequence ref of the actual received TV signal x (that is, equivalently, a position m at which the correlation R _{x · ref} is maximized. I want to estimate the position m where R _{x0 · ref} is maximum. (The position m where R _{x0 · ref} is the maximum value R _{x0 · ref} ^max is represented as L [R _{x0 · ref} ^max ] in the N output sequences.) The position m of R _{x0 · ref} ^max is estimated from the information of the position m of _{x · ref} ^max . By choosing the correct position m, a threshold is used to ensure that the result of equation (1) determines that no TV signal is present, and for the result of equation (2), the threshold is reached based on the output of the highest SNR. It can be determined whether or not. This clearly improves the detection performance. If the position is known as being within a certain range rather than the exact position due to a limitation arising from the estimation accuracy of the position m, whether or not the threshold has been reached is determined by selecting a peak in the output within that range.

And here a more advanced method can be used. First, at the position m of the one having the maximum value (peak value) of R _{x · ref} which is an output of a perfect sliding correlation (that is, calculated for all cases of m = 0, 1,..., N−1). A time offset (that is, a certain m range) is roughly estimated in a relatively large range based on the distribution (a distribution of repetitive calculations at different times). Then, based on the distribution of the position m of the peak of R _{x · ref} obtained from the sliding correlation (that is, for m in the range) (the distribution of the one repeatedly calculated at different times) The offset (that is, a certain m range) is accurately estimated gradually over a small range. In this way, a part of the sliding correlation calculation can be repeated many times to narrow a part of the range where the sliding correlation is obtained. If the position m (range) obtained by this is applied, the peak of R _{x0 · ref} is also obtained. The possibility of becoming very high.

  Compared to taking full sliding correlation, taking selective (= partial) sliding correlation has the following two advantages. One is to reduce calculation load and avoid system complexity. The other is improved detection performance, for the reasons already described above. A part of the range where the sliding correlation is taken depends on the estimation result of the time offset. The estimation result is difficult to obtain as an accurate value, and is usually shown in a certain range. The determination based on the result of the selective (partial) sliding correlation is one feature point in the embodiment.

(Averaging part of the received signal sequence)
(Correlation or averaging of calculation results corresponding to correlation)
A part of the sequence of received signals is averaged. Part of this should be correlated with the reference sequence. Further, the correlation or the calculation result corresponding to the correlation is averaged and used for determination. As a result, the SNR is improved. In some cases, the calculation load can be reduced and the complexity of the system can be avoided to some extent. Averaging can ideally be performed until the result of the averaging does not differ from the result of recalculation with the addition of the next data, but it is not necessary to wait until that time and use the averaging of an appropriate number of data. You can also. The use of averaging appropriately is one feature on the embodiment.

(Measures for the case where the correlation is calculated for only one of the I and Q channels)
The following describes how to deal with the case where the correlation is performed on only one of the I channel and the Q channel.

When the received TV signal is a DVB-T signal, the phase deviation of the detection target signal (DVB-T signal) is substantial if the correlation (or sliding correlation) is made only with one channel, not both the I channel and Q channel. The detection performance is degraded. That is,
R _{x · ref} = R _{x0 · ref} + R _{n · ref} = | R _{x0 · ref} | cos (φ) + R _{n · ref}
R _{x · ref} : Correlation of noise-containing DVB-T with reference sequence R _{x0 · ref} : Correlation of pure DVB-T signal with reference sequence (x ₀ is pure DVB-T signal)
R _{n · ref} : Correlation between noise and reference sequence φ: Phase deviation of pure DVB-T signal from reference sequence φ = 2kπ ± (1/2) π (k = 0, 1,...) R _{x0 -Ref} is zero. Such influence can be reduced by, for example, averaging the absolute value of the correlation output. That is,
_{/ R x · ref = Σ i} | R i x · ref | = Σ i | R i x0 · ref || cos (φ i) | + Σ i | R i n · ref |
Although noise cannot be canceled by this method, it is possible to avoid falling into the worst case in which R _{x0 · ref} = 0. Dealing with the occurrence of the phase deviation is one feature point in the embodiment.

(Specific band extraction filtering)
For a signal whose power distribution changes in the frequency domain as in the case of, for example, a PAL signal, the SNR can be improved by performing narrowband filtering on the high-power band of the received signal. Further, for example, in the case of a DVB-T signal, the SNR can be improved by performing narrow band filtering particularly on the signal band among the bands allocated to the TV signal. Alternatively, in the case of a DVB-T signal, it may be possible to improve the SNR by performing filtering using a comb filter that can extract a subcarrier frequency band in which pilot power is increased. By these, detection performance can be improved. Such filtering of a specific band is one feature point in the embodiment.

  Each embodiment is used to detect very low level TV signals, especially in a communication environment in the Television White Space (TVWS) frequency band, and can be applied to cognitive radio communications and networks thereof.

  FIG. 1 is a block diagram showing a configuration of a spectrum detection apparatus according to one embodiment (first embodiment). As shown in FIG. 1, this spectrum detection apparatus includes a TV channel BPF, a specific band extraction filter 12 (optional configuration), a baseband conversion unit 13 (optional configuration), an AD converter 14, a time offset correction unit 15, a frequency An offset correction unit 16 (optional configuration), a buffer 17, an averaging unit 18 (optional configuration), a selective sliding correlation calculation unit 19, a peak correlation output detection unit 20, and a determination unit 21 are included. What was drawn with a broken line among the configurations shown in the figure can be optionally provided. This is the same in other figures.

  The TV channel BPF 11 is a band-pass filter that accepts an RF signal as input and has a filtered signal corresponding to the RF signal having a pass characteristic in a frequency band in which a detection target signal can exist. The bandwidth can be set to, for example, 6 MHz, 7 MHz, 8 MHz, or the like according to the standard of each TV signal.

  The specific band extraction filter 12 is a filter having a pass band characteristic narrower than that of the TV channel BPF 11. In view of the periodicity of the detection target signal, a filter having the same pass bandwidth can be obtained. For example, when the detection target signal is PAL, a BPF having a passband of 15 kHz (horizontal synchronization frequency) can be used. Further, for example, in the case of a DVB-T signal, it can be a narrow band filter particularly for a signal band among bands allocated to a TV signal. Alternatively, in the case of a DVB-T signal, it may be possible to use a comb filter that can extract a subcarrier frequency band in which pilot power is increased. By performing such filtering for extracting a specific band, a band not used for signal detection is substantially cut, and as a result, spectrum detection performance can be improved.

  The baseband conversion unit 13 converts the RF signal into a baseband signal. This is in accordance with the fact that the RF signal input to the TV channel BPF 11 is a signal modulated for transmission, but it is not essential to provide it, and the subsequent signal processing is performed in the RF band. It is also possible.

  The AD converter 14 samples an input analog signal at a predetermined frequency and further converts it into a digital signal. By converting to a digital signal, subsequent signal processing (especially processing by the time offset correction unit 15) can be easily performed by software.

  Prior to the correlation calculation operation with the reference sequence in the selective sliding correlation calculation unit 19, the time offset correction unit 15 corrects (time shifts) the input digital signal so as to remove the time difference (time offset) from the reference sequence. Is. Therefore, the reference sequence will be described first. The reference sequence is a signal that has a periodicity similar to that of the detection target signal and is generated locally by this apparatus.

  The local generation of the reference sequence is as follows.

1) When detection target signal is DVB-T and similar TV signal:
In DVB-T, roughly 10% of the total subcarriers are allocated for the boosted pilot. The boosted pilot includes a continuous pilot (CP) and a scattered pilot (SP), and their positions and sizes are as follows. That is,
C _{p, l, k} = 4 (1-2ω _k ) / 3
Here, the sequence ω _k is obtained by a pseudo random binary sequence (PRBS) generator using a polynomial X ¹¹ + X ² +1 and setting its initial state to (11111111111). k is as follows at the position of the subcarrier where CP and SP are placed. That is, the subcarrier position of the CP in 2K mode is shown in FIG. The SP is placed on the remaining subcarriers as many as the calculated value of k-3 (l mod 4) divided by 12 (where “l mod 4” is the remainder modulo 4 with l as the dividend). .

  In DVB-T, the pilot signal arranged in this way becomes a periodic signal with 4 symbols. The reference sequence is generated as a repetitive signal following such 4 symbols.

2) For PAL, NTSC, SECAM, and other similar TV signals:
An example of the generated signal is shown in FIG. It has a horizontal sync pulse and a luminance signal (Y signal) set to a medium value. That is, the reference sequence in this case is generated as a repetitive signal of the horizontal synchronization frequency. Note that the luminance signal level is variable.

  The function of the time offset correction unit 15 is to correct (time shift) the digital signal from the AD converter 14 so that there is no time difference with respect to the local reference sequence generated as described above.

The time offset correction unit 15 can be configured using software as its function. Although not shown, the time offset correction unit 15 is provided with a reference sequence generation unit, and further, a sliding correlation for calculating a sliding correlation between the digital signal input to the time offset correction unit 15 and the reference sequence. There is a calculator. The process in the time offset correction unit 15 is as follows.
1. Find the sliding correlation output position m (sequence order) in the peak and record it (no decision). This is repeated M times (M> 1) at an interval divided by the repetition time of the detection target signal (DVB-T signal). Based on the obtained distribution of M positions m, a time offset range is estimated as a range of m having a high occurrence frequency.

The “sliding correlation” is as follows.
The sliding correlation between the digital signal obtained from the AD converter 14 and the reference sequence ref can be expressed as follows.
When the TV signal _{x 0} does not exist:
R _{n · ref} (m) = Σ _{i = 0} ^N−1 ref (i) n (i + m),
i = 0, 1,..., N-1; m = 0, 1,.
... (1)
When the TV signal _{x 0} is present:
R _{x · ref} (m)
= R _{x0 · ref} (m) + R _{n · ref} (m)
= Σ _{i = 0} ^N−1 ref (i) x ₀ (i + m) + Σ _{i = 0} ^N−1 ref (i) n (i + m)
... (2)
Here, N is the number of samples of the reference sequence that is a repetitive pattern, m is a sliding value represented by an integer of 0 or more and less than N, and i is an addition parameter for Σ.

Above 1. Following is the following.
2. Among the outputs of some of the sliding correlations, a peak position m is found and recorded (not determined). Some sliding correlations here are: Is determined based on the result of estimating the time offset range in (i.e., the range of m having a high occurrence frequency in 1.) is used. This is repeated K times (K> 1) at an interval divided by the repetition time of the detection target signal (DVB-T signal). Then, based on the obtained distribution of K positions m, the time offset range is estimated again as a range of m having a high occurrence frequency.

2. Following is the following.
3. If accuracy is required for the accuracy of the time offset, further some sliding correlations are performed. Do the same. Here, a part of the sliding correlation is determined based on the previous time offset range estimation result (that is, the range of m having a high occurrence frequency in the previous time is used).

3. above. Following is the following.
4). Based on the estimated time offset value (if the time offset is obtained as a range, using a representative value (for example, average value, mode, median, etc.)), the digital signal obtained from the AD converter 14 is converted to the time. Correct (time shift) and output.

  The time offset correction unit 15 corrects the digital signal from the AD converter 14 so that there is no time difference with respect to the local reference sequence (m actually used at that time is a correct value due to the influence of noise). It may not be at least close to that). As a result, the correlation calculation result in the selective sliding correlation calculation unit 19 to be described later becomes a large value as compared with the case where the time offset is maintained, and is not buried in noise and is suitable for determination.

  The frequency offset correction unit 16 corrects a frequency offset included in a signal input thereto. The frequency offset may be generated because a certain transmission medium passes through until the TV signal reaches the TV channel BPF11. If such a frequency offset is corrected rather than left, the subsequent processing result such as the selective sliding correlation 19 is improved. A specific configuration example of the frequency offset correction unit 16 will be described later (FIG. 16).

  The buffer 17 stores a signal input thereto for a predetermined time. The averaging unit 18 is configured to improve the SNR by averaging signals input thereto.

  The selective sliding correlation calculator 19 calculates a correlation between a signal input thereto and a reference sequence. Here, the signal input to this and the reference sequence are signals having a relationship in which the time offset between them has been (almost) corrected by the operation of the time offset correction unit 15. Therefore, even if the sliding correlation is referred to, it is not necessary to calculate each of the sliding correlation values from m = 0 to m = N−1 (that is, complete), and some sliding correlation values around m = 0 are calculated. If this is the case, this will likely include a peak value in each value calculated for all m.

  Here, the calculation range of a part of the sliding correlation values can be determined based on the range of m that has already been finally estimated by the time offset correction unit 15.

  Compared with taking full sliding correlation, taking some selective (= partial) sliding correlation has the following two advantages. One is to reduce calculation load and avoid system complexity. The other is improvement of detection performance by obtaining a peak value of correlation calculation with high probability.

  The peak correlation output detection unit 20 selects a peak value from the calculation results of the selective sliding correlation calculation unit 19. The determination unit 21 is located on the output side of the peak correlation output detection unit 20, determines whether or not the detection target signal is present from comparison with a predetermined threshold value, and outputs a determination result signal.

  According to this embodiment, the following advantages are essential. First, since the detection target signal has time synchronization with the reference sequence by the time offset correction unit 15, the calculated correlation value is hardly buried in the noise and becomes (almost) the maximum, and the detection performance is remarkably improved. . In addition, since the selective sliding correlation calculation unit 19 calculates some selective (= partial) sliding correlations, the calculation load can be reduced and the complexity of the system can be reduced. The detection performance is improved by obtaining a value with high probability.

  Next, FIG. 4 is a block diagram showing a configuration of a spectrum detection apparatus which is another embodiment (second embodiment). In the figure, the same reference numerals are given to the same components as those shown in the already described drawings, and the description thereof is omitted unless there is a matter to be added.

  In this spectrum detection apparatus, a simple correlation calculation unit 19A is used instead of the selective sliding correlation calculation unit 19 in FIG. Further, since the calculation result of the correlation calculation unit 19A is an output of one value, the peak correlation output detection unit 20 in FIG. 1 is not necessary.

  This spectrum detector can be said to be a simplified version of that shown in FIG. That is, the signal input to the correlation calculation unit 19A and the reference sequence are signals having a relationship in which the time offset between them has already been (almost) corrected by the operation of the time offset correction unit 15. This is the point described with reference to FIG. Therefore, if the correlation value is simply calculated using the signal input to the correlation calculation unit 19A and the reference sequence, it is very likely that a peak in the sliding correlation has been obtained. Therefore, the correlation value obtained in this way is output to the determination unit 21 for determination.

  Also in this spectrum detection device, since the signal to be detected has time synchronization with the reference sequence by the time offset correction unit 15, the calculated correlation value is difficult to be buried in the noise and becomes (almost) the maximum, and the detection is performed. The performance is significantly improved.

  Next, FIG. 5 is a block diagram showing a configuration of a spectrum detection apparatus which is still another embodiment (third embodiment). In the figure, the same reference numerals are given to the same components as those shown in the already described drawings, and the description thereof is omitted unless there is a matter to be added.

  In this spectrum detection apparatus, the specific band extraction filter 12 is essential, while the time offset correction unit 15 is optionally configured. The other points are the same as those of the spectrum detection apparatus shown in FIG.

  According to this, by providing the specific band extraction filter 12, the target frequency of the detection target signal to be correlated is limited. Therefore, a band that does not significantly affect detection is removed, it is difficult to be affected by noise, and if there is a detection target signal, its presence is easily reflected in the correlation calculation result. Thereby, the performance of spectrum detection can be improved. When the time offset correction unit 15 is not provided, the correlation calculation by the correlation calculation unit 19A is generally a correlation calculation between signals having a time offset, and thus a peak value is not necessarily obtained. However, since a constant correlation value can be calculated, it is possible to make a determination.

  Next, FIG. 6 is a block diagram showing a configuration of a spectrum detection apparatus which is still another embodiment (fourth embodiment). In the figure, the same reference numerals are given to the same components as those shown in the already described drawings, and the description thereof is omitted unless there is a matter to be added.

  Compared with the one shown in FIG. 5, this spectrum detection device returns the specific band extraction filter 12 to an optional configuration, while providing a multiplication unit 19B and an integration unit 22 instead of the correlation calculation unit 19A. It has become.

  The multiplier 19B performs multiplication with a reference sequence having the same periodicity as that of the detection target signal, and further outputs an absolute value of the multiplication result. The integration unit 19B performs an integration operation according to the length of the reference sequence and outputs the result. That is, the multiplication unit 19B and the integration unit 22 calculate the sum after converting the product into an absolute value in the process of calculating the product and subsequently calculating the sum in comparison with the correlation calculation unit 19A in FIG. It can be said that it was changed to.

  This spectrum detection apparatus is useful when, for example, the signal to be detected is a signal such as a DVB-T signal, and the correlation calculation is performed by only one of the channels, not the I channel and the Q channel. It is. In such a case, due to the phase deviation of the detection target signal from the reference sequence, in the worst case, the value becomes zero in normal correlation calculation. Therefore, if the product is calculated in the process of calculating the correlation, it is converted into an absolute value, and then the sum is calculated. According to this, since the worst case where the correlation calculation value becomes zero can be avoided, the detection performance can be maintained.

  Next, FIG. 7 is a block diagram showing a configuration of a spectrum detection apparatus which is still another embodiment (fifth embodiment). In the figure, the same reference numerals are given to the same components as those shown in the already described drawings, and the description thereof is omitted unless there is a matter to be added.

  This spectrum detector is a further improvement of that shown in FIG. That is, a buffer 23 and an averaging unit 24 are newly provided on the output side of the correlation calculation unit 19 </ b> A and on the preceding stage of the determination unit 21. In FIG. 4, the averaging unit 18 provided on the output side of the buffer 17 and on the upstream side of the correlation calculation unit 19A is not provided. It can be said that the averaging unit 24 is provided on the output side of the correlation calculating unit 19A instead of providing the averaging unit 18.

  The buffer 23 stores the correlation value calculated by the correlation calculation unit 19A every time the calculation is performed (that is, a plurality of times). The averaging unit 24 calculates and outputs an average of a plurality of correlation values stored in the buffer 23. According to such a configuration, in the output of the averaging unit 24, the difference between the output values in the case where the detection target signal does not exist and the case where the detection target signal exists is much greater than in the case where only one correlation value is used. Becomes prominent. Therefore, highly reliable determination is possible.

  As shown in FIG. 4, the averaging unit 18 may be provided on the upstream side of the correlation calculating unit 19A, but generally, if the averaging unit is provided, cost and processing time are required due to a buffer required in the previous stage. . Therefore, in the spectrum detection apparatus shown in FIG. 7, the averaging unit 24 is provided only on the output side of the correlation calculation unit 19A.

  Next, FIG. 8 is a block diagram showing a configuration of a spectrum detection apparatus which is still another embodiment (sixth embodiment). In the figure, the same reference numerals are given to the same components as those shown in the already described drawings, and the description thereof is omitted unless there is a matter to be added.

  In this spectrum detection apparatus, the time offset correction unit 15 is optionally configured as compared with the spectrum detection apparatus shown in FIG. The other points are the same as those shown in FIG. When the time offset correction unit 15 is not provided, the correlation calculation by the correlation calculation unit 19A is generally a correlation calculation between signals having a time offset, and thus a peak value is not necessarily obtained. However, since a constant correlation value can be calculated, it is possible to make a determination.

  Since the buffer 23 and the averaging unit 24 are provided as already described with reference to FIG. 7, the difference between the output values when the detection target signal does not exist and when it exists is 1 This is much more noticeable than when only one-time correlation values are used. Therefore, highly reliable determination is possible.

  Next, FIG. 9 is a block diagram showing a configuration of a spectrum detection apparatus which is still another embodiment (seventh embodiment). In the figure, the same reference numerals are given to the same components as those shown in the already described drawings, and the description thereof is omitted unless there is a matter to be added.

  This spectrum detection apparatus is configured by selectively adopting some elements in each of the embodiments described above. That is, the multiplication unit 19B and the integration unit 22 are the same as those in the spectrum detection apparatus shown in FIG. The operation and effect are the same. The buffer 25 and the averaging unit 26 are the same as the buffer 23 and the averaging unit 24 shown in FIG. The operation and effect are the same.

  Next, FIG. 10 is a block diagram showing a configuration of a spectrum detection apparatus which is still another embodiment (eighth embodiment). In the figure, the same reference numerals are given to the same components as those shown in the already described drawings, and the description thereof is omitted unless there is a matter to be added.

  This spectrum detection apparatus will be described in comparison with that shown in FIG. 10 replaces the multiplication unit 19B and the integration unit 22 shown in FIG. 6 with a sliding correlation calculation unit 19C, a buffer 27, an averaging unit 28, an absolute value output calculation unit 29, and a peak output detection unit 30. Is provided.

  The sliding correlation calculation unit 19C calculates a sliding correlation with the reference signal and outputs each correlation value (m = 0 to N−1). The buffer 27 is connected to the output side of the sliding correlation calculation unit 19C, and outputs a buffered signal. The averaging unit 28 is connected to the output side of the buffer 27 and outputs an averaged signal. The absolute value output calculation unit 29 is connected to the output side of the averaging unit 28, performs conversion to an absolute value, and outputs a signal. The peak output detection unit 30 is connected to the output side of the absolute value output calculation unit 29, selects the maximum value among the correlation values averaged and converted into absolute values, and outputs this to the determination unit 21 To do.

  According to this, each calculation result of the sliding correlation is less likely to be buried in noise by the buffer 27 and the averaging unit 28, and the reliability becomes high. Further, by providing the absolute value output calculation unit 29, only one of the received signals having the I signal and the Q signal is used for detection, and a phase deviation with respect to the reference signal is generated, so that the calculation result of the correlation is negative. Even if it becomes a value of, it is treated as meaningful without immediately rejecting it, so the reliability of detection can be improved.

  Next, FIG. 11 is a block diagram showing a configuration of a spectrum detection apparatus which is still another embodiment (the ninth embodiment). In the figure, the same reference numerals are given to the same components as those shown in the already described drawings, and the description thereof is omitted unless there is a matter to be added.

  This spectrum detection apparatus will be described in comparison with that shown in FIG. The spectrum detection apparatus shown in FIG. 11 is that the time offset correction unit 15 is changed from an optional configuration to an essential configuration, and in addition, a selective sliding correlation calculation unit 19 is provided instead of the sliding correlation calculation unit 19C. This is a difference from the one shown in.

  Therefore, the advantage of having the selective sliding correlation calculation unit 19 is as already described in the description of FIG. Since the time offset correction unit 15 is essential, a selective sliding correlation calculation unit 19 can be provided instead of the sliding correlation calculation unit 19C.

  Next, FIG. 12 is a block diagram showing a configuration of a spectrum detection apparatus which is still another embodiment (tenth embodiment). In the figure, the same reference numerals are given to the same components as those shown in the already described drawings, and the description thereof is omitted unless there is a matter to be added.

  The sliding correlation calculation unit 19C is connected to the output side of the buffer 17, calculates a sliding correlation with the reference signal, and outputs a signal as each correlation value. The buffer 27 is connected to the output side of the sliding correlation calculation unit 19C, and outputs a buffered signal. The absolute value output calculation unit 29 is connected to the output side of the buffer 27 and performs conversion into an absolute value. The averaging unit 31 is connected to the output side of the absolute value output calculation unit 29 and outputs an averaged signal. The peak output detection unit 30 is connected to the output side of the averaging unit 31, selects the maximum value among the correlation values whose absolute values are averaged, and outputs this to the determination unit 21.

According to this, only one of the received signals having the I signal and the Q signal is detected by providing the absolute value output calculating unit 29 and the averaging unit 31 that collects and averages a plurality of the outputs. Even when a phase deviation occurs with respect to the reference signal, the reliability of detection can be maintained. That is, the output of the averaging unit 31 is
_{/ R x · ref = Σ i} | R i x · ref | = Σ i | R i x0 · ref || cos (φ i) | + Σ i | R i n · ref |
Thus, the case where R _{x0 · ref} becomes zero with the phase deviation φ as the average value / R _{x · ref} can be avoided.

  Next, FIG. 13 is a block diagram showing a configuration of a spectrum detection apparatus which is still another embodiment (eleventh embodiment). In the figure, the same reference numerals are given to the same components as those shown in the already described drawings, and the description thereof is omitted unless there is a matter to be added.

  This spectrum detection apparatus will be described in comparison with that shown in FIG. The spectrum detection apparatus shown in FIG. 13 is provided with a selective sliding correlation calculation unit 19 in place of the sliding correlation calculation unit 19C in addition to changing the time offset correction unit 15 from an optional configuration to an essential configuration. This is a difference from the one shown in.

  Therefore, the advantage of having the selective sliding correlation calculation unit 19 is as already described in the description of FIG. Since the time offset correction unit 15 is essential, a selective sliding correlation calculation unit 19 can be provided instead of the sliding correlation calculation unit 19C.

Next, FIG. 14 is a block diagram showing a configuration of a spectrum detection apparatus which is still another embodiment (a twelfth embodiment). As shown in the figure, the spectrum detection apparatus includes an input amplification unit and BPF 51, a narrowband BPF 55, a sampling unit and AD converter 52, a signal averaging unit 56, a square addition unit 53, and a determination unit 54. The narrow band BPF 55 and the signal averaging unit 56 are configured to include at least one of them. As the detection target signal, PAL / NTSC / SECAM is considered as a target.

  The input amplification unit and the BPF 51 have functions similar to those of the TV channel BPF 11 in the spectrum detection apparatus shown in FIG. The narrow band BPF 55 is a BPF having a pass band of 15 kHz (horizontal synchronization frequency), for example. Similarly, the sampling unit and the AD converter 52 have the same function as the AD converter 14 in FIG. The signal averaging unit 56 is connected to the sampling unit and the output side of the AD converter 52, and outputs an averaged signal.

  The square addition unit 53 is connected to the sampling unit and the output side of the AD converter 52 and performs an operation of square addition. The determination unit 54 has the same function as the determination unit 21 in the spectrum detection apparatus shown in FIG. That is, the determination unit 54 is located on the output side of the square addition unit 53, determines whether or not there is a detection target signal based on a comparison with a predetermined threshold value, and outputs a determination result signal.

  The spectrum detection apparatus shown in FIG. 14 does not perform the process of calculating the correlation value with the reference sequence, but instead performs the square addition of the signal by the square addition unit 53. That is, the presence or absence of the detection target signal is determined by power detection.

  Thus, in the spectrum detection apparatus that determines whether or not the detection target signal is present by power detection, the advantage of including at least one of the narrowband BPF 55 and the signal averaging unit 56 is as follows. By providing the narrow band BPF 55, the target frequency of the received signal to be square-added is limited, so that it is easily reflected in the square addition result depending on the presence of the detection target signal. Therefore, the detection performance can be improved. In addition, when the signal averaging unit 56 is provided, the signal is less likely to be buried in noise and has high reliability. Therefore, the detection performance can be improved.

  Next, FIG. 15 is a block diagram showing a configuration of a spectrum detection apparatus which is still another embodiment (a thirteenth embodiment). In the figure, the same reference numerals are given to the same components as those shown in the already described drawings, and the description thereof is omitted unless there is a matter to be added.

  This spectrum detection apparatus is different from that shown in FIG. 14 in that an absolute value addition unit 57 is provided instead of the square addition unit 53. Other configurations are the same. The absolute value adding unit 57 is connected to the sampling unit and the output side of the AD converter 52 and performs an operation of adding absolute values. It is considered that performing absolute value addition is substantially equivalent to calculating power by performing square addition.

  Next, a configuration example of the frequency offset correction unit 16 that can be used in the first to eleventh embodiments will be described with reference to FIG. FIG. 16 is a block diagram illustrating a configuration example of the frequency offset correction unit 16 that can be used in the first to eleventh embodiments.

  The frequency correction unit 16 includes an FFT unit 61, a buffer 62, a frequency correction unit 63, an inverse FFT unit 64, and a frequency offset estimation unit 65. A signal on the output side of the AD converter 14 is guided to the FFT unit 61, and a signal on the output side of the inverse FFT unit 64 is output to the buffer 17.

  The FFT unit 61 performs FFT processing on the input signal (time domain signal) and converts it to spectrum information on the frequency axis. The buffer 62 is provided on the output side of the TTF unit 61 and outputs a buffered signal. The frequency correction unit 63 is provided on the output side of the buffer 62 and corrects the frequency offset on the frequency axis based on the estimated value of the frequency offset supplied from the frequency offset estimation unit 65. The inverse FFT unit 64 is provided on the output side of the frequency correction unit 63 and converts a given signal into a time domain signal.

  The frequency offset estimation unit 65 is provided on the output side of the FFT unit 61, compares the spectrum information given from the FFT unit 61 with the standard frequency information (matches), and the spectrum information given from the FFT unit 61. Estimate the frequency offset. Information on the frequency offset obtained by the estimation is supplied to the frequency correction unit 63. In the above, the standard frequency information is frequency information about each subcarrier of an OFDM (orthogonal frequency-division multiplexing) method defined in a signal such as DVB-T that is a detection target signal.

  The advantages of providing the frequency offset correction unit 16 in the spectrum detection apparatus of the first to eleventh embodiments are as already described in the spectrum detection apparatus shown in FIG.

  DESCRIPTION OF SYMBOLS 11 ... TV channel BPF, 12 ... Specific band extraction filter, 13 ... Baseband conversion part, 14 ... AD converter, 15 ... Time offset correction part, 16 ... Frequency offset correction part, 17 ... Buffer, 18 ... Averaging part, 19 ... selective sliding correlation calculation unit, 19A ... correlation calculation unit, 19B ... multiplication unit, 19C ... sliding correlation calculation unit, 20 ... peak correlation output detection unit, 21 ... determination unit, 22 ... accumulation unit, 23 ... buffer, 24 ... Averaging unit, 25 ... buffer, 26 ... averaging unit, 27 ... buffer, 28 ... averaging unit, 29 ... absolute value output calculating unit, 30 ... peak output detecting unit, 31 ... averaging unit, 51 ... input amplifying unit And BPF, 52... Sampling unit and AD converter, 53... Square addition unit, 54. Determination unit, 55... Narrowband BPF, 56. Absolute value adding unit, 61 ... FFT unit, 62 ... buffer, 63 ... frequency correction unit, 64 ... inverse FFT unit 65 ... frequency offset estimation unit.

Claims (10)

A first filter having a pass characteristic in a frequency band in which a detection target signal may exist and receiving a first signal as an input and outputting a second signal;
An AD converter connected to the output side of the first filter for converting to a digital value and outputting a third signal;
The sliding correlation between the third signal and a reference signal having the same periodicity as that of the detection target signal is connected to the output side of the AD converter, and the obtained correlation values are close to the maximum. The sliding correlation is partially recalculated using the range of sliding values to be, and the time offset of the third signal from the reference signal using the range of sliding values in which each obtained correlation value is close to the maximum. And a time offset correction unit that outputs a fourth signal in which the time offset is corrected,
A first buffer connected to the output side of the time offset correction unit and outputting a buffered fifth signal;
A correlation calculating unit connected to the output side of the first buffer and calculating a correlation with the reference signal and outputting a sixth signal;
A spectrum detection apparatus, comprising: a determination unit connected to an output side of the correlation calculation unit and configured to determine whether or not the detection target signal is present based on a comparison with a threshold value and to output a determination result signal. . The correlation calculation unit partially calculates a sliding correlation with the reference signal using the range of the sliding value when the time offset is estimated in the time offset correction unit, and obtains each correlation obtained Output the value as the sixth signal,
A peak correlation output detector that outputs a seventh signal by selecting the maximum value among the respective correlation values as the sixth signal, connected to the output side of the correlation calculator and the preceding stage of the determination unit The spectrum detection device according to claim 1, further comprising: A second buffer connected to the output side of the correlation calculation unit and connected to the preceding stage of the determination unit, and outputs a buffered eighth signal;
2. An averaging unit connected to the output side of the second buffer and to the preceding stage of the determination unit, and further, outputs an averaged ninth signal. Spectrum detector. A second buffer connected to the output side of the correlation calculation unit and connected to the upstream side of the peak correlation output detection unit to output the buffered eighth signal;
An averaging unit connected to the output side of the second buffer and to the previous stage side of the peak correlation output detection unit, and outputs an averaged ninth signal;
An absolute value output calculation unit that is connected to the output side of the averaging unit and the previous stage of the peak correlation output detection unit, converts the absolute value, and outputs a tenth signal. The spectrum detection apparatus according to claim 2, wherein A second buffer connected to the output side of the correlation calculation unit and connected to the upstream side of the peak correlation output detection unit to output the buffered eighth signal;
An absolute value output calculation unit connected to the output side of the second buffer and the previous stage side of the peak correlation output detection unit to convert to an absolute value and output a ninth signal; and the absolute value output calculation An averaging unit connected to the output side of the unit and connected to the upstream side of the peak correlation output detection unit, and outputs an averaged tenth signal;
The spectrum detection apparatus according to claim 2, further comprising: The second filter connected to the output side of the first filter and the upstream side of the AD converter has a narrow band characteristic than the first filter and outputs a filtered signal filtered with the narrow band characteristic. claim 1 Symbol placement of the spectrum detecting device characterized by comprising a filter further. The apparatus further comprises a baseband converter connected to an output side of the first filter and a pre-stage side of the AD converter for converting into a baseband signal and outputting a converted signal. 1 Symbol placement of the spectrum sensing device. Connected to the output side of the time offset correction unit and the previous stage of the first buffer, performs time frequency conversion on the fourth signal, and uses the matching result between the frequency of the obtained signal and the standard frequency estimate a frequency offset of said fourth signal by performing correction of the frequency offset to said fourth signal and further comprising a frequency offset compensation unit for outputting a frequency corrected signal according to claim 1 Symbol placement Spectrum detector. Wherein and an output side of the first buffer is connected to the front side of the correlation calculation unit, the spectrum of Claim 1 Symbol mounting characterized by comprising further an averaging unit which outputs the average signal Detection device. Filtering the first signal, which is a received signal, with a filter having pass characteristics in a frequency band where the detection target signal may exist, to obtain a second signal;
The AD converter provided on the output side of the filter performs conversion to a digital value to obtain a third signal,
Obtained by calculating a sliding correlation between the third signal and a reference signal having the same periodicity as that of the detection target signal in the time offset correction unit provided on the output side of the AD converter. The sliding correlation is partially recalculated using the range of sliding values where each correlation value is close to the maximum, and the sliding signal range where each obtained correlation value is close to the maximum is used as the value of the third signal. Estimating the time offset with the reference signal, and performing the time offset correction using the time offset to obtain a fourth signal,
The buffer provided on the output side of the time offset correction unit obtains a buffered fifth signal,
A correlation calculation unit provided on the output side of the buffer calculates a correlation with the reference signal to obtain a sixth signal,
A spectrum detection method, wherein a determination unit provided on an output side of the correlation calculation unit compares the detection target signal with a threshold value to determine whether or not the detection target signal is present.

Images