JP6915525B2 - Channel estimation device and channel estimation method for broadband wireless communication systems - Google Patents

Description

The present invention relates to a channel estimation device and a channel estimation method for a broadband wireless communication system having a bandwidth of several hundred MHz or more in the quasi-millimeter wave and millimeter wave bands.

In recent years, a technique relating to a broadband wireless communication system using a quasi-millimeter wave such as 28 GHz, 60 GHz, 73 GHz and the like and a millimeter wave band as a carrier frequency has been reported (Non-Patent Document 1). These bandwidths are several hundred MHz or more, which are wider than those of conventional wireless communication systems that use microwaves (6 GHz or less) such as wireless LANs, and are expected to increase the transmission capacity. However, since the frequency is higher than that of microwaves, RF (radio frequency) components such as local oscillators (hereinafter referred to as LO (local oscillator)) and bandpass filters (hereinafter referred to as BPF (bandpass filter)) and analog digital Wireless communication using the conventional microwave band is required for the performance of circuit components such as converters (hereinafter referred to as ADCs (analog to digital converters)) and digital-to-digital converters (hereinafter referred to as DACs (digital to digital converters)). It is higher than the system, and its imperfections are a factor that hinders the increase in transmission capacity as a wireless communication system (Non-Patent Document 2).

T.S. RAPPAPORT et al., “Millimeter Wave Mobile Communications for 5G Cellular: It Will Work!” IEEE Access, pp. 335-349, vol. 1, May, 2013. Y. Zou et al, "Impact of Major RF Impairments on mm-Wave Communications Using OFDM Waveforms," Proc. IEEE Globecom, Feb. 2017.

In a wideband wireless communication system, due to the influence of the imperfections of the circuit components, noise over the entire band, which is not generated in a conventional wireless communication system using a microwave band, is generated. Hereinafter, noise generated by the influence of imperfections of such circuit components, noise generated by imperfections of band limiting processing, and the like are referred to as "wideband noise". An example of wideband noise is shown in FIG.

Due to the influence of this wideband noise, the channel estimation deteriorates. In the conventional method, it is conceivable to mitigate the influence of wideband noise by increasing the number of known signals for channel estimation and averaging the results of channel estimation. However, this method increases the number of known signals for channel estimation and reduces the data transmission capacity. Further, when the time variation of the channel is large, the channel changes time among many known signals, which makes accurate channel estimation difficult.

The present invention overcomes the performance deterioration of channel estimation over a wide band of several hundred MHz or more due to imperfections of circuit components in quasi-millimeter wave band and millimeter wave band wideband wireless communication systems, and improves the channel estimation performance. It is an object of the present invention to provide a channel estimation device and a channel estimation method for a wideband wireless communication system capable of performing the above.

The first invention is a quasi-millimeter wave and a millimeter wave band, which is arranged in a receiving device of a broadband wireless communication system having a bandwidth of several hundred MHz or more, and channel estimation is performed using a known signal transmitted from the transmitting device. In a channel estimation device used for data signal demodulation, a channel estimation processing unit that estimates a channel from a known signal using a predetermined method and a channel estimation processing unit input the channel estimation result and use a Savitzky Golay filter to bandwidth. The received SNR is calculated from the channel estimation result of the noise smoothing processing unit that uses the channel estimation result that mitigates the influence of wideband noise over the entire width to demolish the data signal, and the channel estimation processing unit, and the calculated received SNR is calculated. The Savitzky Golay filter is equipped with a control unit that sets the parameters of the Savitzky Golay filter using the Savitzky Golay filter. , The window size L is set based on the following formula .
L = Lmin + {(Lmax-Lmin) / (SNRmax-SNRmin)} * SNRrx
However, SNRrx is the received SNR, SNRmin is the minimum value of the received SNR, SNRmax is the maximum value of the received SNR, Lmax is the maximum value of the window size, and Lmin is the minimum value of the window size.

The second invention is a quasi-millimeter wave and a millimeter wave band, which is arranged in a receiving device of a broadband wireless communication system having a bandwidth of several hundred MHz or more, and channel estimation is performed using a known signal transmitted from the transmitting device. In the channel estimation device used for demolishing the data signal, the channel estimation processing unit that estimates the channel from the known signal using a predetermined method and the channel estimation processing unit input the channel estimation result and use the Bilateral filter to input the bandwidth. The reception SNR is calculated from the channel estimation result of the noise removal processing unit that uses the channel estimation result of removing the wideband noise over the entire band for demolishing the data signal and the channel estimation processing unit of the channel estimation processing unit, and the calculated reception SNR is used for Bilateral. The Bilateral filter is equipped with a control unit that sets the parameters of the filter. As parameters , the Bilateral filter has a region parameter in which the noise removal function becomes more effective as the value becomes smaller, and a space in which the noise removal effect becomes larger as the value becomes larger. The parameters are set, and the control unit sets the area parameter to a predetermined value or less and sets the spatial parameter S based on the following formula .
S = Smin + {(Smax-Smin) / (SNRmax-SNRmin)} * SNRrx
However, SNRrx is the received SNR, SNRmin is the minimum value of the received SNR, SNRmax is the maximum value of the received SNR, Smax is the maximum value of the spatial parameter, and Smin is the minimum value of the spatial parameter.

A third invention is a quasi-millimeter wave and a millimeter wave band, which is arranged in a receiving device of a broadband wireless communication system having a bandwidth of several hundred MHz or more, and channel estimation is performed using a known signal transmitted from the transmitting device. In a channel estimation device used for data signal demodulation, a channel estimation processing unit that estimates a channel from a known signal using a predetermined method and a channel estimation processing unit input the channel estimation result and use a Bilateral filter to input the bandwidth. A noise removal processing unit that uses the channel estimation result that removes wideband noise over the entire band to demolish the data signal, and a channel estimation result that removes wideband noise from the noise removal processing unit are input, and a Savitzky Golay filter is used. The reception SNR is calculated from the channel estimation result of the noise smoothing processing unit that uses the channel estimation result that mitigates the influence of wideband noise to demolish the data signal, and the channel estimation processing unit, and the calculated reception SNR is used as the parameter of the Bilateral filter. The Bilateral filter is equipped with a control unit that sets the parameters of the Savitzky Golay filter and the Bilateral filter. As parameters, the smaller the value, the more the noise removal function works. Spatial parameters that increase are set, the Savitzky Golay filter sets the window size, which is the number of data used when approximating the input signal with a polynomial, as parameters, and the control unit sets the area parameter to a predetermined value or less. , Spatial parameter S and window size L are set based on the following formulas .
S = Smin + {(Smax-Smin) / (SNRmax-SNRmin)} * SNRrx
L = Lmin + {(Lmax-Lmin) / (SNRmax-SNRmin)} * SNRrx
However, SNRrx is the received SNR, SNRmin is the minimum value of the received SNR, SNRmax is the maximum value of the received SNR, Smax is the maximum value of the spatial parameter, Smin is the minimum value of the spatial parameter, Lmax is the maximum value of the window size, and Lmin is the maximum value. The minimum window size.

A fourth invention is a quasi-millimeter wave and a millimeter wave band, which is arranged in a receiving device of a broadband wireless communication system having a bandwidth of several hundred MHz or more, and channel estimation is performed using a known signal transmitted from the transmitting device. In the channel estimation method used for data signal demodulation, a channel estimation processing step that estimates a channel from a known signal using a predetermined method and a channel estimation result of the channel estimation processing step are input, and a band is used using a Savitzky Golay filter. The received SNR is calculated from the noise smoothing processing step in which the channel estimation result in which the influence of wideband noise over the entire width is mitigated is used for data signal demodulation, and the channel estimation result in the channel estimation processing step, and the calculated received SNR is calculated. have a control step of setting the parameters of the Savitzky Golay filter with, Savitzky Golay filter, as a parameter, the window size is the number of data used when approximating an input signal by a polynomial is set, control steps Then, the window size L is set based on the following formula .
L = Lmin + {(Lmax-Lmin) / (SNRmax-SNRmin)} * SNRrx
However, SNRrx is the received SNR, SNRmin is the minimum value of the received SNR, SNRmax is the maximum value of the received SNR, Lmax is the maximum value of the window size, and Lmin is the minimum value of the window size.

A fifth invention is a quasi-millimeter wave and a millimeter wave band, which is arranged in a receiving device of a broadband wireless communication system having a bandwidth of several hundred MHz or more, and channel estimation is performed using a known signal transmitted from the transmitting device. In the channel estimation method used for demolishing a data signal, a channel estimation processing step that estimates a channel from a known signal using a predetermined method and a channel estimation result of the channel estimation processing step are input, and a bandwidth is used using a Bilateral filter. The reception SNR is calculated from the noise removal processing step in which the channel estimation result after removing the wide band noise over the entire band is used for demolishing the data signal and the channel estimation result in the channel estimation processing step, and the calculated reception SNR is used for Bilateral. have a control step for setting a filter parameter, Bilateral filter, as a parameter, and the area parameter functions as noise removal value is reduced is to work more, the effect of noise removal as the value becomes larger increases Spatial parameters are set, and in the control step, the area parameter is set to a predetermined value or less, and the spatial parameter S is set based on the following formula .
S = Smin + {(Smax-Smin) / (SNRmax-SNRmin)} * SNRrx
However, SNRrx is the received SNR, SNRmin is the minimum value of the received SNR, SNRmax is the maximum value of the received SNR, Smax is the maximum value of the spatial parameter, and Smin is the minimum value of the spatial parameter.

A sixth invention is a quasi-millimeter wave and a millimeter wave band, which is arranged in a receiving device of a broadband wireless communication system having a bandwidth of several hundred MHz or more, and channel estimation is performed using a known signal transmitted from the transmitting device. In the channel estimation method used for data signal demodulation, the channel estimation processing step that estimates the channel from a known signal using a predetermined method and the channel estimation result of the channel estimation processing step are input, and the bandwidth is used by using a Bilateral filter. using a noise removing step of subjecting the channel estimation result of removing the broadband noise in the demodulated data signals over the entire band, the broadband noise of the noise removal processing steps to enter the channel estimation result removed, a further Savitzky Golay filter The received SNR is calculated from the channel estimation result of the noise smoothing processing step in which the channel estimation result in which the influence of wideband noise is mitigated is used for data signal demodulation, and the channel estimation result of the channel estimation processing step is calculated, and the calculated reception SNR is used as the parameter of the Bilateral filter. and possess a control step for setting a Savitzky Golay filter parameters, Bilateral filter parameters as a regional parameters functions as noise removal value is reduced is to work more, effective as noise removal value increases The spatial parameter that increases is set, and the Savitzky Golay filter sets the window size, which is the number of data used when approximating the input signal with a polynomial, as a parameter, and in the control step, the area parameter is set to a predetermined value or less. Then, the spatial parameter S and the window size L are set based on the following formulas .
S = Smin + {(Smax-Smin) / (SNRmax-SNRmin)} * SNRrx
L = Lmin + {(Lmax-Lmin) / (SNRmax-SNRmin)} * SNRrx
However, SNRrx is the received SNR, SNRmin is the minimum value of the received SNR, SNRmax is the maximum value of the received SNR, Smax is the maximum value of the spatial parameter, Smin is the minimum value of the spatial parameter, Lmax is the maximum value of the window size, and Lmin is the maximum value. The minimum window size.

The present invention focuses on the fact that the noise smoothing filter and the noise removal filter used in voice communication and the like are suitable for the wideband noise included in the channel estimation result, and removes the wideband noise by these processes to estimate the channel. Performance and overall performance can be improved.

In addition, by considering the SNR (signal to noise ratio) of the received signal when setting the parameters of these filters, the functions of the noise smoothing filter and the noise elimination filter can be combined with the wideband wireless communication system. It can be exerted appropriately, and the resistance to channel time fluctuation can be improved.

Further, since wideband noise can be reduced without increasing the number of known signals for channel estimation, the transmission capacity can be increased, the load of channel information acquisition can be reduced, and the calculation for interference elimination can be performed. The amount can be reduced.

It is a figure which shows the configuration example of the transmission / reception device in the broadband wireless communication system of this invention. It is a figure which shows the digital signal processing unit 23A of the 1st configuration example of a receiving apparatus. It is a figure which shows the example of the wide band noise. It is a figure which shows the frequency response after noise smoothing. It is a figure which shows the demodulation result by the channel estimation before noise smoothing. It is a figure which shows the demodulation result by the channel estimation after noise smoothing. It is a figure which shows the digital signal processing unit 23B of the 2nd configuration example of a receiving apparatus. It is a figure which shows the digital signal processing unit 23C of the 3rd configuration example of a receiving apparatus. It is a figure which shows the frequency response before noise removal. It is a figure which shows the frequency response after noise removal. It is a figure which shows the frequency response after noise smoothing. It is a figure which shows the demodulation result by the channel estimation before noise removal and noise smoothing. It is a figure which shows the demodulation result by the channel estimation after noise removal and noise smoothing.

FIG. 1 shows a configuration example of a transmission / reception device in the broadband wireless communication system of the present invention.
In FIG. 1 (1), the transmitting device includes a digital signal processing unit 11, an RF processing unit 12, and a transmitting antenna unit 13. The digital signal processing unit 11 performs digital signal processing necessary for modulation of a data signal to be transmitted and spatial multiplex transmission of a plurality of channels. The RF processing unit 12 performs analog processing such as frequency conversion and RF filtering. The transmitting antenna unit 13 transmits radio signals of a plurality of channels.

In FIG. 1 (2), the receiving device includes a receiving antenna unit 21, an RF processing unit 22, and a digital signal processing unit 23. The receiving antenna unit 21 receives radio signals of a plurality of spatially multiplexed channels. The RF processing unit 22 performs analog processing such as frequency conversion of the received signal and RF filtering. The digital signal processing unit 23 performs channel estimation of a plurality of channels and demodulates the data signal.

When performing channel estimation, the digital signal processing unit 11 of the transmission device periodically attaches a known signal to the front of the data signal sequence and transmits the known signal. The digital signal processing unit 23 of the receiving device separates a plurality of data signals and known signals simultaneously received by spatial multiplexing, performs channel estimation using the information of the known signals, and performs channel equalization processing using the channel information. Then, the data signal is demolished.

Hereinafter, as the digital signal processing unit 23 of the receiving device, the digital signal processing unit 23A of the first configuration example, the digital signal processing unit 23B of the second configuration example, and the digital signal processing unit 23C of the third configuration example will be described. ..

(Digital signal processing unit 23A of the first configuration example)
FIG. 2 shows the digital signal processing unit 23A of the first configuration example of the receiving device.
In FIG. 2, the digital signal processing unit 23A is composed of a known signal / data signal separation unit 31, a channel estimation device 32A, and a demodulation processing unit 33. Further, the channel estimation device 32A is composed of a channel estimation processing unit 321, a noise smoothing processing unit 322, and a control unit 323A.

The known signal / data signal separation unit 31 separates a plurality of data signals simultaneously received by spatial multiplexing and a known signal periodically attached to each data signal. Processing such as synchronous detection of received signals is performed by a conventional method. Here, the plurality of separated data signals are input to the demodulation processing unit 33, and using the channel estimation result output by the channel estimation device 32A described later, equalization processing, channel code decoding processing, demodulation determination processing, etc. Demodulation processing is performed.

On the other hand, the plurality of separated known signals are input to the channel estimation device 32A. The channel estimation processing unit 321 of the channel estimation device 32A performs channel estimation by a method such as ZF (zero forcing) or MMSE (minimum mean square error) using a plurality of input known signals. For example, in the case of ZF, the channel can be estimated by dividing the known signal into the received signal. The result of this channel estimation includes the above-mentioned wideband noise.

The noise smoothing processing unit 322 inputs the channel estimation result from the channel estimation processing unit 321 and performs noise smoothing processing using a Savitzky Golay filter (hereinafter referred to as SG filter) or the like to reduce the influence of wideband noise. Is output to the demodulation processing unit 33. The demodulation processing unit 33 performs demodulation processing using this channel estimation result.

The control unit 323A calculates the received SNR from the estimation result of the channel estimation processing unit 321, sets the parameters of the SG filter using the calculated received SNR, and outputs the set parameters to the noise smoothing processing unit 322. Even if the received SNR fluctuates due to time fluctuation, it can be dealt with by changing the setting of the SG filter parameter.

The noise smoothing processing unit 322 sets the SG filter using this parameter and performs the noise smoothing processing. For example, the SG filter sets the degree of the polynomial that approximates the input signal and the number of data used when approximating using the polynomial (hereinafter, referred to as window size L) as parameters. At this time, if the minimum value of the window size L is Lmin, the maximum value is Lmax, the minimum value of the target received SNR is SNRmim, the maximum value is SNRmax, and the received SNR is SNRrx, the window size L is as follows. Set as.
L = Lmin + {(Lmax-Lmin) / (SNRmax-SNRmin)} * SNRrx… (1)

Here, SNRmim, SNRmax, Lmax, and Lmin are input separately or set to predetermined values. The equation (1) is an example of determining the window size L, and the window size L may be determined by another method. Further, the degree of the polynomial of the SG filter is determined by a predetermined method among 2 to 4.

By inputting the result of channel estimation including wideband noise shown in FIG. 3 into the noise smoothing processing unit (SG filter) 322 and performing noise smoothing, as shown in FIG. 4, the result of alleviating the influence of wideband noise is obtained. Be done. 5 and 6 show demodulation results by channel estimation before noise smoothing in FIG. 3 and after noise smoothing in FIG. From FIGS. 5 and 6, it can be seen that the performance has been improved by the present invention.

(Digital signal processing unit 23B of the second configuration example)
FIG. 7 shows a digital signal processing unit 23B of a second configuration example of the receiving device.
In FIG. 7, the digital signal processing unit 23B has a configuration in which a noise removal processing unit 324 is arranged in place of the noise smoothing processing unit 322 of the channel estimation device 32A of the digital signal processing unit 23A of the first configuration example shown in FIG. Is. The noise removal processing unit 324 removes wideband noise included in the output result of the channel estimation processing unit 321 by using a Bilateral filter (hereinafter referred to as a B filter) for removing small noise contained in a large signal. ..

The control unit 323C calculates the received SNR from the estimation result of the channel estimation processing unit 321, sets the parameters of the B filter using the calculated received SNR, and outputs the set parameters to the noise removal processing unit 324. Even if the received SNR fluctuates due to time fluctuation, it can be dealt with by changing the setting of the B filter parameter.

The noise reduction processing unit 324 sets the B filter using this parameter and performs noise removal processing. For example, the B filter is adjusted by the range parameter and the spatial parameter. When the area parameter becomes large, the function of the B filter becomes close to noise smoothing, and when the area parameter becomes small, the function of the B filter becomes small. Is closer to noise removal. Here, since noise smoothing is performed by the SG filter, the area parameter is set to a predetermined parameter of 0.001 or less so that the noise removal function works more in the B filter.

The larger the value of the spatial parameter, the greater the effect of noise removal, but if it is too large, even the desired signal is removed. Therefore, the spatial parameters are adjusted according to the received SNR so that the function works to the maximum. For example, if the minimum value of the received SNR is SNRmim, the maximum value is SNRmax, the maximum value of the spatial parameter S is Smax, the minimum value is Smin, and the received SNR is SNRrx, the spatial parameter S is set as follows.
S = Smin + {(Smax-Smin) / (SNRmax-SNRmin)} * SNRrx ... (2)

Here, SNRmim, SNRmax, Smax, and Smin are input separately or set to predetermined values.

(Digital signal processing unit 23C of the third configuration example)
FIG. 8 shows the digital signal processing unit 23C of the third configuration example of the receiving device.
In FIG. 8, the channel estimation device 32C of the digital signal processing unit 23C has the noise smoothing processing unit 322 of the first configuration example shown in FIG. 2 after the noise removal processing unit 324 of the second configuration example shown in FIG. Is arranged, wideband noise is removed from the channel estimation result of the channel estimation processing unit 321, noise smoothing processing is performed, and the result is output to the demodulation processing unit 33. That is, since noise is first removed and then noise smoothing processing is performed, it is more effective when the SNR is small. The control unit 323C performs processing in which the control unit 323A of the first configuration example and the control unit 323B of the second configuration example are combined.

FIG. 10 shows the result of removing wideband noise by passing the result of channel estimation (before noise removal) shown in FIG. 9 through a B filter. Further, the result after noise smoothing by passing through the SG filter is shown in FIG. 12 and 13 show demodulation results by channel estimation before noise reduction / noise smoothing in FIG. 9 and after noise removal / noise smoothing in FIG. 11. From FIGS. 12 and 13, it can be seen that the performance has been improved by the present invention.

11 Digital signal processing unit 12 RF processing unit 13 Transmitting antenna unit 21 Receiving antenna unit 22 RF processing unit 23 Digital signal processing unit 31 Known signal / data signal separation unit 32A, 32B, 32C channel estimation device 33 Demodulation processing unit 321 Channel estimation processing Unit 322 Noise smoothing processing unit 323A, 323B, 323C Control unit 324 Noise removal processing unit

Claims (6)

It is placed in the receiver of a broadband wireless communication system with a bandwidth of several hundred MHz or more in the quasi-millimeter wave and millimeter wave bands, and channel estimation is performed using the known signal transmitted from the transmitter and used for demodulation of the data signal. In the channel estimator
A channel estimation processing unit that estimates channels from the known signals using a predetermined method, and
A noise smoothing processing unit that inputs the channel estimation result from the channel estimation processing unit and uses a Savitzky Golay filter to mitigate the influence of wideband noise over the entire bandwidth of the bandwidth and uses the channel estimation result for demodulation of the data signal. When,
It is provided with a control unit that calculates the received SNR from the channel estimation result of the channel estimation processing unit and sets the parameters of the Savitzky Golay filter using the calculated received SNR.
In the Savitzky Golay filter, the size of the window, which is the number of data used when approximating the input signal with a polynomial, is set as the parameter.
The control unit is a channel estimation device, characterized in that the size L of the window is set based on the following mathematical formula.
L = Lmin + {(Lmax-Lmin) / (SNRmax-SNRmin)} * SNRrx
However, SNRrx is the received SNR, SNRmin is the minimum value of the received SNR, SNRmax is the maximum value of the received SNR, Lmax is the maximum value of the window size, and Lmin is the minimum value of the window size. It is placed in the receiver of a broadband wireless communication system with a bandwidth of several hundred MHz or more in the quasi-millimeter wave and millimeter wave bands, and channel estimation is performed using the known signal transmitted from the transmitter and used for demodulation of the data signal. In the channel estimator
A channel estimation processing unit that estimates channels from the known signals using a predetermined method, and
A noise removal processing unit that inputs the channel estimation result from the channel estimation processing unit and uses a Bilateral filter to remove wideband noise over the entire bandwidth, and uses the channel estimation result for demodulation of the data signal.
It is provided with a control unit that calculates the received SNR from the channel estimation result of the channel estimation processing unit and sets the parameters of the Bilateral filter using the calculated received SNR.
In the Bilateral filter, as the parameters, a region parameter in which the noise removal function becomes more effective as the value becomes smaller and a spatial parameter in which the noise removal effect becomes larger as the value becomes larger are set.
The control unit is a channel estimation device, characterized in that the area parameter is set to a predetermined value or less, and the spatial parameter S is set based on the following mathematical formula.
S = Smin + {(Smax-Smin) / (SNRmax-SNRmin)} * SNRrx
However, SNRrx is the received SNR, SNRmin is the minimum value of the received SNR, SNRmax is the maximum value of the received SNR, Smax is the maximum value of the spatial parameter, and Smin is the minimum value of the spatial parameter. It is placed in the receiver of a broadband wireless communication system with a bandwidth of several hundred MHz or more in the quasi-millimeter wave and millimeter wave bands, and channel estimation is performed using the known signal transmitted from the transmitter and used for demodulation of the data signal. In the channel estimator
A channel estimation processing unit that estimates channels from the known signals using a predetermined method, and
A noise removal processing unit that inputs the channel estimation result from the channel estimation processing unit and uses a Bilateral filter to remove wideband noise over the entire bandwidth, and uses the channel estimation result for demodulation of the data signal.
Enter the channel estimation result of removing the broadband noise from the noise removing unit, and further Savitzky Golay noise smoothing processing unit subjected to demodulation of the channel estimation result of mitigating the effects of the broadband noise by using a filter the data signal ,
A control unit that calculates the received SNR from the channel estimation result of the channel estimation processing unit and sets the parameters of the Bilateral filter and the Savitzky Golay filter using the calculated received SNR is provided.
In the Bilateral filter, as the parameters, a region parameter in which the noise removal function becomes more effective as the value becomes smaller and a spatial parameter in which the noise removal effect becomes larger as the value becomes larger are set.
In the Savitzky Golay filter, the size of the window, which is the number of data used when approximating the input signal with a polynomial, is set as the parameter.
The control unit is a channel estimation device, characterized in that the area parameter is set to a predetermined value or less, and the spatial parameter S and the window size L are set based on the following mathematical formulas.
S = Smin + {(Smax-Smin) / (SNRmax-SNRmin)} * SNRrx
L = Lmin + {(Lmax-Lmin) / (SNRmax-SNRmin)} * SNRrx
However, SNRrx is the received SNR, SNRmin is the minimum value of the received SNR, SNRmax is the maximum value of the received SNR, Smax is the maximum value of the spatial parameter, Smin is the minimum value of the spatial parameter, and Lmax is the size of the window. The maximum value of, Lmin, is the minimum value of the window size. It is placed in the receiver of a broadband wireless communication system with a bandwidth of several hundred MHz or more in the quasi-millimeter wave and millimeter wave bands, and channel estimation is performed using the known signal transmitted from the transmitter and used for demodulation of the data signal. In the channel estimation method
A channel estimation processing step for estimating a channel from the known signal using a predetermined method, and
A noise smoothing processing step in which the channel estimation result of the channel estimation processing step is input, and the channel estimation result in which the influence of wideband noise over the entire bandwidth of the bandwidth is mitigated by using a Savitzky Golay filter is used for demodulation of the data signal. When,
Wherein the received SNR from the results of channel estimation of the channel estimation process step calculates, have a control step for setting a parameter of the Savitzky Golay filter with the calculated received SNR,
In the Savitzky Golay filter, the size of the window, which is the number of data used when approximating the input signal with a polynomial, is set as the parameter.
In the control step, the channel estimation method is characterized in that the size L of the window is set based on the following mathematical formula.
L = Lmin + {(Lmax-Lmin) / (SNRmax-SNRmin)} * SNRrx
However, SNRrx is the received SNR, SNRmin is the minimum value of the received SNR, SNRmax is the maximum value of the received SNR, Lmax is the maximum value of the window size, and Lmin is the minimum value of the window size. It is placed in the receiver of a broadband wireless communication system with a bandwidth of several hundred MHz or more in the quasi-millimeter wave and millimeter wave bands, and channel estimation is performed using the known signal transmitted from the transmitter and used for demodulation of the data signal. In the channel estimation method
A channel estimation processing step for estimating a channel from the known signal using a predetermined method, and
Enter the result of the channel estimation of the channel estimation process step, a noise removing step of subjecting the demodulation of the data signal to the channel estimation result of removing a broadband noise over the entire bandwidth of the bandwidth by using a Bilateral filter,
Wherein the received SNR from the results of channel estimation of the channel estimation process step calculates, have a control step for setting a parameter of the Bilateral Filter using the calculated received SNR,
In the Bilateral filter, as the parameters, a region parameter in which the noise removal function becomes more effective as the value becomes smaller and a spatial parameter in which the noise removal effect becomes larger as the value becomes larger are set.
In the control step, the channel estimation method is characterized in that the area parameter is set to a predetermined value or less and the spatial parameter S is set based on the following mathematical formula.
S = Smin + {(Smax-Smin) / (SNRmax-SNRmin)} * SNRrx
However, SNRrx is the received SNR, SNRmin is the minimum value of the received SNR, SNRmax is the maximum value of the received SNR, Smax is the maximum value of the spatial parameter, and Smin is the minimum value of the spatial parameter. It is placed in the receiver of a broadband wireless communication system with a bandwidth of several hundred MHz or more in the quasi-millimeter wave and millimeter wave bands, and channel estimation is performed using the known signal transmitted from the transmitter and used for demodulation of the data signal. In the channel estimation method
A channel estimation processing step for estimating a channel from the known signal using a predetermined method, and
A noise removal processing step in which the channel estimation result of the channel estimation processing step is input and the channel estimation result obtained by removing wideband noise over the entire bandwidth of the bandwidth using a Bilateral filter is used for demodulation of the data signal.
Enter the channel estimation result of removing the broadband noise of the noise removing step, and further the noise smoothing processing step of subjecting the influence channel estimation result of the relaxation of the broadband noise for demodulation of the data signal using a Savitzky Golay Filter ,
Said channel calculates the received SNR from the results of channel estimation estimation process step, have a control step of setting the parameters and parameter of the Savitzky Golay filter of the Bilateral Filter using the calculated received SNR,
In the Bilateral filter, as the parameters, a region parameter in which the noise removal function becomes more effective as the value becomes smaller and a spatial parameter in which the noise removal effect becomes larger as the value becomes larger are set.
In the Savitzky Golay filter, the size of the window, which is the number of data used when approximating the input signal with a polynomial, is set as the parameter.
In the control step, the channel estimation method is characterized in that the area parameter is set to a predetermined value or less, and the spatial parameter S and the window size L are set based on the following mathematical formulas.
S = Smin + {(Smax-Smin) / (SNRmax-SNRmin)} * SNRrx
L = Lmin + {(Lmax-Lmin) / (SNRmax-SNRmin)} * SNRrx
However, SNRrx is the received SNR, SNRmin is the minimum value of the received SNR, SNRmax is the maximum value of the received SNR, Smax is the maximum value of the spatial parameter, Smin is the minimum value of the spatial parameter, and Lmax is the size of the window. The maximum value of, Lmin, is the minimum value of the window size.

Images