JP7817092B2 - Receiver and impulse noise removal method - Google Patents

Description

The present invention relates to a receiving device and an impulse noise removal method.

In-vehicle receiving devices are known that receive broadcast signals, such as digital television broadcast waves, modulated using a modulation method known as OFDM (Orthogonal Frequency Division Multiplexing).

Broadcast signals received by such in-vehicle receiving devices may contain impulse noise, which occurs when the vehicle's electrical system starts or stops. Note that "impulse noise" here typically refers to noise that appears intermittently as an impulse waveform (a waveform that repeatedly rises and falls sharply in a very short period of time).

In response to this, a technology has been known in the past that uses a band pass filter (BPF) to extract portions of broadcast waves that correspond to guard bands outside the signal region, and then detects impulse noise in these extracted portions (see, for example, Patent Document 1).

JP 2012-147276 A

However, conventional technology has the problem that it may not be possible to accurately detect impulse noise.

For example, vehicles may generate not only impulse noise, but also steady noise that is unrelated to impulse noise. Note that "steady noise" here typically refers to noise that is smaller in amplitude than impulse noise but appears continuously over time.

Impulse noise has a large amplitude but only appears for an extremely short period of time, so its energy is small when viewed across the entire guard band and it may be buried in steady noise. As a result, the detection sensitivity of impulse noise decreases.

In this case, simply lowering the threshold to increase the detection sensitivity of impulse noise is undesirable, as it could result in steady noise being detected as impulse noise, resulting in an increase in false positives.

The present invention has been made in consideration of the above, and aims to provide a receiving device and an impulse noise removal method that can accurately detect impulse noise.

To solve the above-mentioned problems and achieve the objectives, a receiving device according to the present invention, a receiving device 10 of a first embodiment, has a control unit that performs processing to remove impulse noise from broadcast waves. The control unit adjusts the bandwidth of a bandpass filter so as to extract some frequency components, including peak frequencies, from the guard band of the broadcast waves. The control unit detects a first noise position, which is a position in the time domain where the level of the frequency components extracted by the bandpass filter exceeds a first threshold, and a second noise position, which is a position in the time domain where the level of the broadcast waves exceeds a second threshold. The control unit performs processing to remove noise from the broadcast waves at positions where the first noise position and second noise position overlap.

According to the present invention, impulse noise can be detected with high accuracy.

FIG. 1 is a diagram showing an outline of an impulse noise removal technique. FIG. 2 is a block diagram showing the configuration of the receiving device. FIG. 3 is a block diagram showing the configuration of the noise canceller. FIG. 4 is a diagram illustrating the adjustment of the bandpass filter. FIG. 5 is a diagram illustrating periodic impulse noise. FIG. 6 is a diagram for explaining the time averaging process in the time averaging unit. FIG. 7 is a diagram illustrating the threshold value determination process in the threshold value determination unit. FIG. 8 is a diagram illustrating the estimated position extension process in the estimated position extension unit. FIG. 9 is a diagram illustrating the noise removal process in the noise removal unit. FIG. 10 is a flowchart showing the processing procedure executed by the noise canceller.

[First embodiment]
First, an outline of the impulse noise removal technique used by the receiving device of the first embodiment will be described with reference to Fig. 1. Fig. 1 is a diagram showing an outline of the impulse noise removal technique.

Figure 1(A) shows an overview of the process of estimating potential noise contamination intervals based on the level of the guard band portion of the broadcast wave. Figure 1(B) shows an overview of the process of estimating noise locations based on the level of the band portion of the broadcast wave that includes the broadcast signal. Figure 1(C) shows an overview of the process of detecting impulse noise based on the relationship between the results of both estimations.

The receiving device of the first embodiment estimates the noise position based on some frequency components of the guard band of the broadcast wave, specifically, frequency components including the peak frequency. Furthermore, the receiving device estimates the noise position based on the band portion of the broadcast wave that includes the broadcast signal, and then detects impulse noise based on both estimation results.

Specifically, as shown in Figure 1 (A), the receiving device extracts a portion of the guard band from the broadcast wave using a bandpass filter (see Figure 1 (A-1)). This is because the guard band, which is a gap where there is no signal, is suitable for detecting impulse noise components that may be mixed in. Bandpass filters will be described later using Figure 4.

Here, (A-2) in Figure 1 shows an example of the detection results when the level of the guard band portion extracted by a bandpass filter is viewed in the time domain (i.e., expanded on the time axis t). As shown in (A-2) in Figure 1, when a filter circuit such as a bandpass filter is used, impulse noise is detected with a delay from position t1, where the impulse noise actually enters, and with a wider width. This is due to the group delay of the filter, and is the same when using digital filters such as FIR (Finite Impulse Response) filters or IIR (Infinite Impulse Response) filters.

Based on this detection result, the receiving device estimates section w1, which shows a signal level exceeding threshold b1, as a potential contamination section (see (A-2) in Figure 1).

Then, as shown in (A-3) of Figure 1, the receiving device determines the estimation result, which includes the extended section w1+ obtained by extending the estimated contamination candidate section, as the "first estimation result P1" based on the frequency components. Details regarding this extension will be described later using Figure 8.

On the other hand, as shown in Figure 1 (B), the receiving device determines a "second estimation result P2" based on the waveform level of the broadcast wave in the band portion that includes the broadcast signal. Note that Figure 1 (B-1) shows an example in which the only signals exceeding threshold b2 are impulse noise at position t1 and a strong signal at position t2.

Note that the "band portion of the broadcast wave that includes the broadcast signal" may be the entire broadcast wave input to the receiving device, or a band portion of the entire broadcast wave that includes the broadcast signal of one or more channels, or may include guard bands.

Here, as shown in (B-2) of Figure 1, when determining the "second estimation result P2," the receiving device uses delayed data obtained by delaying the broadcast signal. This is to ensure consistency when comparing the "first estimation result P1," which assumes a delay due to a bandpass filter, with the "second estimation result P2."

Note that this type of matching can also be achieved by delaying a processing result such as the "second estimation result P2" rather than the broadcast signal itself. The following describes the case where the broadcast signal is delayed.

Then, as shown in (B-3) of Figure 1, the receiving device determines the estimation result including positions t1 and t2 that exceed threshold b2 as the "second estimation result P2" based on the time domain.

Next, as shown in Figure 1(C), the receiving device detects impulse noise by taking the logical product of the "first estimation result P1" and the "second estimation result P2" that takes into account the delay difference.

That is, as shown in the area surrounded by the dashed closed curve 1 in Figure 1(C), the signal at position t1 of the "second estimation result P2" included in the extended section w1+, which is a potential contamination section of the "first estimation result P1," is ultimately detected as impulse noise. Furthermore, the strong signal at position t2, which is not included in the extended section w1+, is ultimately not detected as impulse noise.

The receiving device then removes the impulse noise at the noise position t1 detected in this way. The method for removing impulse noise by the receiving device will be described later using Figure 9.

Furthermore, the receiving device can adjust the bandwidth and cutoff frequency (hereinafter referred to as "bandwidth, etc.") of the bandpass filter based on the peak frequency of the guard band portion obtained by performing FFT (Fast Fourier Transform) processing on the broadcast wave. This allows the receiving device to suppress the effects of stationary noise, etc., and further improve the accuracy of impulse noise removal.

Note that (A) in Figure 1 corresponds to the first estimation unit 2, which will be described later using Figure 3. Also, (B) in Figure 1 corresponds to the second estimation unit 3. Also, (C) in Figure 1 corresponds to the noise detection unit 4.

Furthermore, the thresholds b1 and b2 mentioned above can be varied based on the time average of the signal level. The thresholds will be described later using Figures 6 and 7.

Figure 2 is a block diagram showing the configuration of the receiving device 10. Note that Figure 2 shows only the components necessary to explain the features of the receiving device 10, and omits descriptions of general components.

As shown in FIG. 2, the receiving device 10 is included in a digital television (DTV) receiver installed in the vehicle V. In addition, impulse noise generated from electronic devices in the vehicle V may be superimposed on the DTV broadcast signal.

The receiving device 10 includes an RF (Radio Frequency)/IF (Intermediate Frequency) unit 11, a demodulation LSI (Large Scale Integration) 10a, and a decoder 17.

The RF/IF unit 11 extracts and amplifies the signal of the broadcast wave to be received from the broadcast signal received by the antenna. The RF/IF unit 11 includes a high frequency signal (RF signal) amplifier circuit and an intermediate frequency signal (IF signal) amplifier circuit. The RF/IF unit 11 outputs the processed signal (IF signal) to the demodulation LSI 10a. The demodulation LSI 10a demodulates the broadcast signal and outputs it as video and audio.

The demodulation LSI 10a includes a synchronization unit 12, a noise canceller 13, an FFT unit 14, an equalization unit 15, and an error correction unit 16.

The synchronization unit 12 synchronizes the frame timing/symbol timing of the broadcast signal contained in the broadcast wave. The synchronization unit 12 outputs the processed IF signal to the noise canceller 13.

The noise canceller 13 detects and removes impulse noise contained in the input IF signal. The noise canceller 13 outputs the processed IF signal to the FFT unit 14. Details of the noise canceller 13 will be described later using Figure 3.

The FFT unit 16 performs FFT processing on the IF signal and outputs it to the equalization unit 15. At this time, the FFT unit 14 also notifies the noise canceller 13 of the results of the FFT processing (e.g., peak frequency).

The equalization unit 15 performs equalization processing to correct distortion in the input IF signal. The error correction unit 16 detects and corrects errors in the data signal.

The decoder 17 outputs the error-corrected data signal to an external device. The external device may be, for example, an in-car device that displays digital television broadcast images.

In addition, in the receiving device 10, the demodulation LSI 10a functions as an OFDM processing unit.

The configuration of the noise canceller 13 will now be explained in more detail using Figure 3. Figure 3 is a block diagram showing the configuration of the noise canceller.

As shown in FIG. 3, the noise canceller 13 includes a data delay unit 13a, a BPF 13b, time averaging units 13c, 13d, 13h, and 13i, threshold determination units 13e and 13j, comparators 13f and 13k, an estimated position expansion unit 13g, an AND circuit 13l, and a noise removal unit 13m.

In the noise canceller 13, the BPF 13b, time averaging units 13c and 13d, threshold determination unit 13e, comparator 13f, and estimated position extension unit 13g function as the first estimation unit 2.

Furthermore, the time averaging units 13h and 13i, the threshold determination unit 13j, and the comparator 13k function as the second estimation unit 3. Furthermore, the AND circuit 13l functions as the noise detection unit 4.

The data delay unit 13a stores the input signal (hereinafter referred to as "actual data") in a memory or the like and delays it. The data delay unit 13a outputs the delayed input signal (hereinafter referred to as "delayed data") to the time averaging unit 13h, time averaging unit 13i, and noise removal unit 13m that make up the second estimation unit 3.

BPF 13b is a bandpass filter that passes the portion corresponding to the band set based on the actual data and outputs it to the time averaging unit 13c.

Here, the noise canceller 13 acquires the peak frequency contained in the guard band of the broadcast wave and adjusts (sets) the bandwidth of the BPF 13b so as to extract a portion of the frequency components, including the peak frequency, from the guard band.

Then, the noise canceller 13 detects a first noise position where the level of the frequency component extracted by the BPF 13b exceeds a first threshold, and a second noise position where the level of the broadcast wave exceeds a second threshold. The first noise position and the second noise position are positions in the time domain (e.g., a time or a time interval).

The process of outputting the first noise position is performed by the first estimation unit 2. The first threshold corresponds to threshold b1. The first noise position corresponds to the first estimation result P1 (e.g., the extended section w1+).

The process of outputting the second noise position is performed by the second estimation unit 3. The second threshold corresponds to threshold b2. The second noise position corresponds to the second estimation result P2 (e.g., position t1 or position t2).

Furthermore, the noise canceller 13 performs a process (corresponding to the noise removal unit 13m described below) to remove noise from the broadcast wave at the position where the first noise position and the second noise position overlap (for example, the area surrounded by closed curve 1 in Figure 1).

The process of detecting position and removing noise performed by the noise canceller 13 will be explained later using Figures 5 to 9. Here, the process of adjusting the bandwidth of the BPF 13b performed by the noise canceller 13 will be explained using Figures 4 and 5. As mentioned above, by adjusting the bandwidth of the BPF 13b, the receiving device 10 can suppress the effects of stationary noise and further improve the accuracy of removing impulse noise.

As shown in Figure 3, the noise canceller 13 obtains the peak frequency from the FFT unit in order to adjust the bandwidth of the BPF 13b, etc.

Then, as shown in Figure 4, the noise canceller 13 changes the band of the BPF 13b from R1 to R2. Figure 4 is a diagram explaining the adjustment of the bandwidth of the bandpass filter. At this time, the peak frequencies are assumed to be f1 and f2 (where f1 < f2).

Band R1 is the band set as the initial value. For example, band R1 is the band corresponding to the entire guard band.

Band R2 is narrower than band R1 and includes peak frequencies f1 and f2. It may be a band with f1 as the lower limit and f2 as the upper limit.

Specifically, the noise canceller 13 can be adjusted by changing the coefficients that determine the bandwidth of the BPF 13b, etc.

The number of peak frequencies is not limited to that shown in Figure 4, and may be one or three or more. The noise canceller 13 adjusts the bandwidth of the BPF 13b so that a predetermined number of peak frequencies are included in the guard band. This makes it possible to increase the proportion of impulse noise in the portion extracted by the BPF 13b while minimizing the influence of stationary noise, etc.

In addition, the noise canceller 13 further adjusts the bandwidth of the BPF 13b, etc., based on the peak frequency obtained as a result of performing an FFT on the broadcast signal from which noise has been removed by the noise removal process, i.e., the peak frequency of the remaining noise. This makes it possible to recursively improve the accuracy of impulse noise detection.

The noise canceller 13 may acquire, as peak frequencies, information indicating frequencies whose amplitude (power) is equal to or greater than a threshold as a result of the FFT performed by the FFT unit 14, or information indicating a certain number of frequencies with the highest amplitude (power) as peak frequencies. The noise canceller 13 may acquire the spectrum of the broadcast wave from the FFT unit 14 and acquire, as peak frequencies, frequencies whose amplitude (power) exceeds a predetermined threshold on the noise canceller 13 side.

In the above explanation, the FFT unit 14 used for demodulation processing is used to extract peak frequencies within the guard band for impulse noise detection. This eliminates the need for a separate FFT unit, reducing costs. However, an FFT for peak frequency extraction may also be provided in addition to the FFT unit 14 used for demodulation processing.

Here, impulse noise may exhibit periodicity in both the frequency domain and the time domain. Adjusting the bandwidth of BPF 13b, as described in Figure 4, is effective in detecting such impulse noise.

Figure 5 is a diagram explaining periodic impulse noise. As shown in Figure 5, impulse noise superimposed on digital television broadcast signals occurs with a period of Δt in the time domain and may appear as a frequency of 1/Δt in the frequency domain.

In particular, the spectrum of an impulse wave is expressed as the sum of sine waves of multiple frequencies, so multiple peaks appear with a fundamental frequency of 1/Δt. On the other hand, steady noise tends to have a uniform waveform across a wide frequency band.

By adjusting the bandwidth of BPF 13b, etc., it is possible to reduce the energy of stationary noise contained in the extracted portion and relatively increase the proportion of impulse noise energy. This prevents impulse noise from being buried in stationary noise and allows for accurate detection of impulse noise.

Here, digital television broadcast channel 14 is assumed to be 479.142857MHz to 476.357857MHz to 481.927857MHz, and channel 15 is assumed to be 482.357857MHz to 487.927857MHz. In this case, 481.927857MHz to 482.357857MHz is the guard band.

Let's assume that the peak frequencies identified by the FFT unit 14 are f1 = 482.100 MHz and f2 = 482.200 MHz. In this case, the noise canceller 13 adjusts the band extracted by the BPF 13b to 482.100 MHz to 482.200 MHz.

In the example of FIG. 4, the noise canceller 13 performs an adjustment to narrow band R1 to band R2. On the other hand, the noise canceller 13 may also move the band of the BPF 13b in the frequency direction depending on the position in the frequency direction of the peak frequency in the guard band. In this way, the noise canceller 13 can flexibly adjust the bandwidth, etc. of the BPF 13b depending on the peak frequency.

For example, the noise canceller 13 may move the band R1', which is between 482.100 MHz and 482.200 MHz, to the band R2', which is between 482.250 MHz and 482.350 MHz.

The method of band adjustment by the noise canceller 13 is not limited to the above. The noise canceller 13 can expand, contract, or move the band in either the positive or negative frequency direction, or both, so as to include the peak frequency.

Note that even if a channel is in a specified band, there may be cases where a broadcast signal is not present depending on the region and time of year. For this reason, when adjusting the band of BPF 13b, noise canceller 13 may treat channels where no broadcast signal is present as guard bands. In other words, part of the adjusted band may overlap with a broadcast signal channel.

Furthermore, the noise canceller 13 may add bands corresponding to each of the multiple peak frequencies in the guard band to the BPF 13b. This allows for even more accurate detection of impulse noise.

For example, if the peak frequencies are 481.950 MHz, 482.050 MHz, 482.150 MHz, and 482.250 MHz, the noise canceller 13 may add band R3, which is 481.950 MHz to 482.050 MHz, and band R4, which is 482.150 MHz to 482.250 MHz.

The following describes the processing that occurs after the bandwidth of BPF 13b has been adjusted. The time averaging unit 13c takes the time average of the portion that has passed through BPF 13b. The time averaging unit 13c outputs the time-averaged data to the time averaging unit 13d and the comparator 13f.

The time averaging unit 13d takes a relatively longer time average of the time-averaged data input from the time averaging unit 13c than the time averaging unit 13c. The time averaging unit 13d outputs the time-averaged data to the threshold determination unit 13e.

Here, the time-averaged data includes an offset value that offsets noise other than impulse noise (hereinafter referred to as "general noise"). Details of the time averaging units 13c and 13d will be described later using Figure 6.

The threshold determination unit 13e determines a threshold corresponding to threshold b1 (see Figure 1) based on the offset value included in the time-averaged data input from the time averaging unit 13d. The threshold determination unit 13e outputs the determined threshold to the comparator 13f. Details of the threshold determination unit 13e will be described later using Figure 7.

Based on the time-averaged data input from the time averaging unit 13c and the threshold value input from the threshold determination unit 13e, the comparator 13f outputs the section showing a signal level exceeding the threshold value (see section w1 in Figure 1) as a potential contamination section to the estimated position expansion unit 13g.

The estimated position expansion unit 13g expands the potential contamination section input from the comparator 13f and outputs the "first estimation result P1" (see Figure 1) including the expanded section (see expanded section w1+ in Figure 1) to the AND circuit 13l. Details of the estimated position expansion unit 13g will be described later using Figure 8.

The time averaging unit 13h takes the time average of the delayed data input from the data delay unit 13a. The time averaging unit 13h outputs the time-averaged data to the comparator 13k.

The time averaging unit 13i takes a relatively longer time average of the delayed data input from the data delay unit 13a than the time averaging unit 13h. The time averaging unit 13i outputs the time-averaged data to the threshold determination unit 13j.

Note that the time-averaged data includes an offset value that offsets general noise, as in the case of the time averaging unit 13d already described. Details of the time averaging units 13h and 13i will be described later using Figure 6.

The threshold determination unit 13j determines a threshold corresponding to threshold b2 (see Figure 1) based on the offset value included in the time-averaged data input from the time averaging unit 13i. The threshold determination unit 13j outputs the determined threshold to the comparator 13k. Details of the threshold determination unit 13j will be described later using Figure 7.

Comparator 13k outputs to AND circuit 13l a "second estimation result P2" (see Figure 1) including positions indicating signal levels exceeding the threshold based on the time-averaged data input from time averaging unit 13h and the threshold input from threshold determination unit 13j. Positions indicating signal levels exceeding the threshold are, for example, positions t1 and t2 in Figure 1.

The AND circuit 13l is a logic circuit that takes the logical product of the "first estimation result P1" input from the estimated position expansion unit 13g and the "second estimation result P2" input from the comparator 13k, and outputs the noise position that is true as a result of the logical product calculation to the noise removal unit 13m.

The noise removal unit 13m removes impulse noise based on the delayed data input from the data delay unit 13a and the noise position input from the AND circuit 13l. Specific removal techniques will be described later using Figure 9.

Here, the time averaging processing in the time averaging units 13c, 13d, 13h, and 13i will be explained using Figure 6. Figure 6 is a diagram explaining the time averaging processing in the time averaging units 13c, 13d, 13h, and 13i.

Note that Figure 6(A) shows an example configuration of the time averaging units 13c, 13d, 13h, and 13i, and Figure 6(B) shows an example operation of the time averaging units 13c and 13d in the first estimation unit 2. Furthermore, the operation examples of the time averaging units 13h and 13i in the second estimation unit 3 overlap with the operation examples of the time averaging units 13c and 13d shown in Figure 6(B), and therefore will not be described here.

As shown in Figure 6(A), a general IIR (Infinite Impulse Response) filter can be used for the time averaging units 13c, 13d, 13h, and 13i.

In this case, the time averaging units 13c, 13d, 13h, and 13i are composed of a multiplier 131, an adder 132, a delay element 133, and a multiplier 134. Here, the multiplier 131 multiplies the input from the previous stage by a coefficient "K (0<K<1)" related to the time averaging, and outputs the result to the adder 132. The delay element 133 delays the output of the adder 132 by "Z ^-1 " (the so-called transfer function of the Z transform), and then outputs the result to the multiplier 134.

Multiplier 134 multiplies the input from delay element 133 by coefficient "1-K" and outputs the result to adder 132. Adder 132 then adds the output of multiplier 131 and the output of multiplier 134 and outputs the result to the subsequent stage.

Here, we will explain an example of the operation of the time averaging units 13c and 13d, using the first estimation unit 2 as an example. As shown in Figure 6(B), the time averaging unit 13c smoothes the waveform output by the BPF 13b, which has sharp fluctuations (see the area surrounded by rectangle 5), by taking a time average (see the area surrounded by rectangle 6).

In other words, by gradually changing the output value indicated by the waveform, it is possible to prevent erroneous detection of impulse noise. Furthermore, the time averaging unit 13d further smooths the waveform by taking a relatively longer time average than the time averaging unit 13c.

This allows the time averaging unit 13d to obtain an offset value that offsets the general noise component (see the area surrounded by rectangle 7). In other words, it is possible to prevent erroneous detection based on general noise components other than impulse noise.

Next, the threshold determination process in the threshold determination units 13e and 13j will be explained using Figure 7. Figure 7 is a diagram explaining the threshold determination process in the threshold determination units 13e and 13j.

As shown in Figure 7, the threshold determination units 13e and 13j are composed of a multiplier 145 and an adder 146. Here, the multiplier 145 multiplies the input (time-averaged value) from the previous stage (time averaging unit 13d or 13i) by a coefficient and outputs the result to the adder 146.

Adder 146 adds the output of multiplier 145 to a fixed threshold set in advance by the user and outputs the result to the subsequent stage (comparator 13f or 13k). The fixed threshold may also be stored in a storage device such as a hard disk drive or non-volatile memory.

This allows the threshold determination units 13e and 13j to determine the threshold based on the time averaging processing performed by the time averaging units 13d and 13i. In other words, adaptive control can be performed to change the criteria for determining impulse noise depending on the reception conditions, etc.

Up until now, we have explained the time averaging units 13c, 13d, 13h, and 13i using Figure 6, and the threshold determination units 13e and 13j using Figure 7, but each of these processing units can be configured using simple, small-scale arithmetic circuits. Therefore, it is possible to configure the receiving device 10 using small components that are low-cost and have excellent processing speeds.

Next, the estimated position expansion process in the estimated position expansion unit 13g will be described using Figure 8. Figure 8 is a diagram illustrating the estimated position expansion process in the estimated position expansion unit 13g. Note that in Figure 8 (1), the waveform corresponding to the detection result (the portion extracted by BPF 13b) smoothed by taking the time average is shown as a sine wave, and fine waveform fluctuations are omitted.

Here, as shown in (1) of Figure 8, the estimated position expansion unit 13g considers the section w1 in the detection result showing a signal level exceeding the threshold b1 as input from the comparator 13f as a potential contamination section.

In this case, as shown in Figure 8 (2), the estimated position extension unit 13g extends the section w1 in the positive direction of the time axis to create an extended section w1+. This is done to prevent impulse noise from being missed due to the section w1 being too narrow.

As shown in Figure 8 (2), the amount of expansion is preferably several times the length of the interval w1 to ensure that detection misses are prevented. The upper limit of the amount of expansion can be set up until the next impulse noise detection period.

Then, as shown in (2) of Figure 8, the estimated position extension unit 13g outputs the "first estimated result P1" (see Figure 1) including the extension section w1+ to the AND circuit 13l. Note that, although the "first estimated result P1" is illustrated here as a pulse wave including the extension section w1+, this does not limit the output format of the estimated position extension unit 13g.

Next, the noise removal process in the noise removal unit 13m will be described using Figure 9. Figure 9 is a diagram explaining the noise removal process in the noise removal unit 13m. Note that Figure 9 (A) shows noise removal method 1, Figure 9 (B) shows noise removal method 2, and Figure 9 (C) shows noise removal method 3.

As shown in Figure 9, the noise removal unit 13m receives the noise position t1 of the detected impulse noise from the AND circuit 13l included in the noise detection unit 4. The area enclosed by the closed curve in Figure 9 (C) is an enlarged view of the vicinity of the noise position t1.

As shown in Figure 9 (A), the noise removal unit 13m can also be configured to remove impulse noise by replacing the output value at noise position t1 with "0".

Also, as shown in Figure 9 (B), the noise removal unit 13m can be configured to attenuate impulse noise components by replacing the output value at noise position t1 with the time average value at that timing.

Furthermore, as shown in Figure 9 (C), the noise removal unit 13m pre-stores a waveform of a typical impulse noise (hereinafter referred to as "noise replica 18a") in the storage unit 18, which is a storage device such as a hard disk drive or non-volatile memory.

As shown in Figure 9 (C), the noise removal unit 13m can also be configured to attenuate the impulse noise component by superimposing a noise replica 18a on the waveform at noise position t1 and then subtracting the noise replica 18a.

The processing procedure executed by the noise canceller 13 will be explained using Figure 10. Figure 10 is a flowchart showing the processing procedure executed by the noise canceller 13.

For example, the noise canceller 13 of the receiving device 10 is configured with a processing unit (such as a microcomputer (DSP) capable of high-speed calculation processing of demodulated signals), and noise is removed through this calculation processing. The processing unit of the noise canceller 13 may also be referred to as a control unit.

First, as shown in FIG. 10, the noise canceller 13 acquires data (IF signal) from the synchronization unit 12 (step S101).

Next, the noise canceller 13 adjusts the bandpass filter based on the peak frequency and extracts a portion of the guard band using the adjusted bandpass filter (step S102).

Then, a threshold is determined based on the time average of the extraction results (step S103). Note that this threshold corresponds to the threshold b1 described above (see Figure 1).

Then, the noise canceller 13 determines whether there is a section on the time axis that shows a signal level that exceeds the determined threshold (step S104). If it is determined that such a section exists (step S104, Yes), the noise canceller 13 expands the corresponding section (step S105) and estimates the expanded section as a candidate section for noise contamination (hereinafter referred to as a "noise section") (step S106).

Next, the noise canceller 13 inputs delayed data obtained by delaying the actual data (step S107) and determines a threshold value based on the time average of the delayed data (step S108). Note that this threshold value corresponds to the threshold value b2 described above (see Figure 1).

Then, the noise canceller 13 determines whether there is a position on the time axis that indicates a signal level that exceeds the determined threshold (step S109). If it is determined that such a position exists (step S109, Yes), the noise canceller 13 determines whether the corresponding position is included in a noise interval based on the actual data (step S110).

If it is determined that the corresponding position is included in the noise section (step S110, Yes), the noise canceller 13 removes the impulse noise at the corresponding position (step S111) and then repeats the processing from step S101.

If the judgment conditions of step S104 are not met (step S104, No), if the judgment conditions of step S109 are not met (step S109, No), or if the judgment conditions of step S110 are not met (step S110, No), the noise canceller 13 repeats the processing from step S101 without removing the impulse noise.

As described above, the receiving device 10 of the first embodiment has a noise canceller 13, which is a control unit that performs processing to remove impulse noise from broadcast waves. The noise canceller 13 adjusts the bandwidth of the BPF 13b so as to extract some frequency components, including peak frequencies, from the guard band of the broadcast waves. The noise canceller 13 detects a first noise position, which is a position in the time domain where the level of the frequency components extracted by the BPF 13b exceeds a first threshold, and a second noise position, which is a position in the time domain where the level of the broadcast waves exceeds a second threshold. The noise canceller 13 performs processing to remove noise from the broadcast waves at positions where the first noise position and second noise position overlap.

In this way, by adjusting the bandwidth of BPF 13b based on the peak frequency thought to be caused by impulse noise, the impulse noise components are relatively increased in the extracted portion. Therefore, according to this embodiment, impulse noise can be removed with high accuracy.

In the first embodiment, the second estimation unit 3 has been described as mainly pinpointing the "position" of a potential noise intrusion, but the "position" is included in the "section." Therefore, in other words, the second estimation unit 3, like the first estimation unit 2, estimates the "section" of a potential impulse noise intrusion.

Furthermore, in the first embodiment, we have described a case where impulse noise is detected by taking the logical product of the estimation results of the first estimation unit 2 and the second estimation unit 3 executed in parallel, but the present invention is not limited to this method.

For example, the estimation results of the first estimation unit 2 may be obtained first, and then the second estimation unit 3 may perform estimation processing only on the candidate impulse noise sections included in the estimation results. In this case, threshold determinations, etc., do not need to be performed on sections other than the candidate impulse noise sections, thereby reducing the processing load on the receiving device.

As described above, the receiving device and impulse noise removal method of the first embodiment are useful when you want to remove impulse noise with high accuracy, and are particularly suitable for application to in-vehicle receiving devices, where impulse noise is likely to occur due to constantly changing reception environments, such as while driving.

10 Receiver 11 RF/IF section 12 Synchronizer 13 Noise canceller 13a Data delay section 13b BPF
13c, 13d Time averaging unit 13e Threshold value determining unit 13f Comparator 13g Estimated position extending unit 13h, 13i Time averaging unit 13j Threshold value determining unit 13k Comparator 13l AND circuit 131, 134, 135 Multipliers 132, 136 Adder 133 Delay element 14 FFT unit 15 Equalization unit 16 Error correction unit 17 Decoder 18 Storage unit 18a Noise replica

Claims (6)

a control unit that performs processing to remove impulse noise from broadcast waves;
The control unit
Acquire a peak frequency included in a guard band of the broadcast wave,
adjusting a bandpass filter so as to extract a portion of frequency components including the peak frequency from the guard band;
Detecting a first noise position, which is a position in a time domain where the level of the frequency component extracted by the band-pass filter exceeds a first threshold, and a second noise position, which is a position in a time domain where the level of the broadcast wave exceeds a second threshold;
a receiving device that performs a process of removing noise from the broadcast wave at a position where the first noise position and the second noise position overlap. The control unit
The receiving device according to claim 1 , wherein the bandpass filter is adjusted so as to include a predetermined number of peak frequencies in the guard band. The control unit
The receiving device according to claim 1 , wherein the band of the band-pass filter is shifted in the frequency direction in accordance with the position of a peak frequency in the guard band in the frequency direction. The control unit
The receiving device according to claim 1 , wherein bands corresponding to a plurality of peak frequencies in the guard band are added to the band-pass filter. The control unit
The receiving device according to claim 1 , further acquiring a peak frequency from the broadcast signal from which noise has been removed by the noise removal process, and adjusting the band of the band-pass filter based on the acquired peak frequency. An impulse noise removal method executed by a control unit that performs processing to remove impulse noise from a broadcast wave,
Acquire a peak frequency included in a guard band of the broadcast wave,
adjusting a bandpass filter so as to extract a portion of frequency components including the peak frequency from the guard band;
Detecting a first noise position, which is a position in a time domain where the level of the frequency component extracted by the band-pass filter exceeds a first threshold, and a second noise position, which is a position in a time domain where the level of the broadcast wave exceeds a second threshold;
An impulse noise removal method for removing noise from the broadcast wave at a position that is a logical product of the first noise position and the second noise position.