JP4445358B2 - Fading signal generator and multipath fading simulator - Google Patents

Description

  The present invention relates generally to the technical field of wireless communication, and more particularly to a fading simulator and a multipath fading simulator that simulate the effects of fading on a received signal.

  The frequency band used in the mobile communication system is becoming wider in the future. Along with this, fading due to the Doppler effect causes certain frequency characteristics to be observed within the frequency band, which greatly affects the received signal quality. Therefore, in designing a mobile communication system, it is important to deal with not only multipath fading but also fading due to Doppler shift.

FIG. 1 shows the relationship between the power spectrum of the received signal and the Doppler shift. It is assumed that the moving body travels in a certain traveling direction, and radio waves arrive from isotropically various directions. As shown on the left side of FIG. 1, the spectrum is distributed in the range of ± f _D around the carrier frequency f _C. f _D is the maximum Doppler frequency, and represents a frequency shift or Doppler shift observed when the angle α between the traveling direction of the moving object and the arrival direction of the radio wave is zero. k (α) is a parameter that depends on the angle α. For example, when using a vertical whip antenna and assuming uniform arrival, k (α) = 3 / 2π. The frequency shift for an arbitrary angle α is expressed as f _D × cos (α). Therefore, if the angle α changes according to a certain probability distribution within a given range (for example, 0 ≦ α ≦ 2π), many incoming waves having a large Doppler shift of about ± f _D are generated, There are few incoming waves with small Doppler shifts. Therefore, the power density of the reception signal having the Doppler shift of about ± f _D is large. In the figure, a plurality of lines drawn adjacent to the cosine wave graph show how the frequency interval of the Doppler frequency is widened when the angle α is changed at a constant interval.

  Such evaluation for fading is often performed using a fading simulator in addition to or instead of receiving radio waves actually transmitted to space. The fading simulator generates a fading signal representing the influence of fading, and multiplies the fading signal by the signal input to the fading simulator and outputs the result. In this case, as shown on the left side of FIG. 1, what simulates the effect of Doppler shift in the radio frequency band (band expression), and equivalently simulated on the baseband as shown on the right side of FIG. (Equivalent low-frequency expression). The latter can simulate the influence of fading without constructing a processing circuit for the radio frequency band. This is advantageous from the viewpoint of evaluating the performance of the system by dividing it into a baseband band and a radio frequency band. Further, since a high frequency circuit is not required for the simulation, the fear that the fading evaluation accuracy deteriorates due to imperfection or deterioration of the characteristic of the circuit including the analog high frequency element is also eliminated. Furthermore, the latter fading simulator is advantageous from the viewpoint of, for example, modeling and investigating a plurality of characteristics deterioration causes given to the received signal for each cause. In the embodiments described later, an equivalent low-frequency representation method is used, but the present invention can be applied to any representation method.

  A first method for simulating the influence of fading as shown in FIG. 1 (particularly the left side) is to superimpose a large number of sine waves (elementary waves). In this method, the fading signal T (t) is

で表現され、ここで、ω_Ｄ＝２πｆ_Ｄは最大ドップラー角周波数を表し、ｆ_Ｄは最大ドップラー周波数を表し、φ_ｎはｎ番目の素波の位相を表し、α_ｎはｎ番目の素波の到来角（移動体の進行方向及び電波の到来方向のなす角）を表し、ｃ_ｎ又はＥ_０ｃ_ｎはｎ番目の素波の振幅を表し、Ｅ_０は所与の定数を表し、Ｎは素波の総数を表す。素波の振幅データは、テーブルに格納される。素波の振幅データは、１周期の期間にわたる振幅値、即ち１周期中の全てのサンプリング時点に対する振幅値として、ルックアップテーブル（ＬＵＴ）に格納される。それらの振幅値は、所定のサンプリング周波数に応じて抽出され、（１）式に基づいて合成され、フェージング信号Ｔ（ｔ）として出力される。この種の技術は、例えば特許文献１及び非特許文献１に記載されている。

Where ω _D = 2πf _D represents the maximum Doppler angular frequency, f _D represents the maximum Doppler frequency, φ _n represents the phase of the n th elementary wave, and α _n represents the n th elementary wave. represents the angle of arrival (angle between the traveling direction and the radio wave arrival direction of the moving object), c _n or E ₀ c _n represents the amplitude of the n-th elementary wave, E ₀ represents the given constants, n Represents the total number of rays. The amplitude data of the elementary wave is stored in a table. The amplitude data of the elementary wave is stored in a look-up table (LUT) as an amplitude value over a period of one cycle, that is, an amplitude value for all sampling points in one cycle. Those amplitude values are extracted according to a predetermined sampling frequency, synthesized based on the equation (1), and output as a fading signal T (t). This type of technology is described in Patent Document 1 and Non-Patent Document 1, for example.

A second method for simulating the influence of fading shapes a white Gaussian noise (WGN) signal with a low-pass filter (LPF) having a special transfer function. The transfer function is flat in the vicinity of the center frequency as shown in FIG. 1, but increases sharply as it deviates from the center frequency, and becomes zero when exceeding the maximum Doppler frequency. Such a fading simulator is called a JTC fader. The JTC fader is described in, for example, Patent Document 2 and Non-Patent Document 2.
JP-A-6-140950 JP 2003-298536 A W. C. Jakes, Jr. "Microwave Mobile Communications", John Wiley & Sons, pp. 67-76 (1974) “The JTC Fader”, 3GPP2 FYI, Nov. 8, 2002, Sandip Sarker, Qualcomm

According to the first method, the amplitude data stored in the table is prepared so that all amplitude values are extracted at a predetermined sampling frequency f _S during one period of the elementary wave. Such amplitude data must be prepared for each elementary wave. For example, it is assumed that the sampling frequency f _S is 3.8 MHz and the maximum Doppler frequency f _D is 100 Hz. The Doppler frequency f _D cos α varies in the range of −100 to 100 Hz. In this case, for an elementary wave having a maximum Doppler frequency of 100 Hz, 3.8 MHz / 100 Hz = 38000 amplitude values need to be stored in the table. For a 10 Hz elementary wave, 3.8 MHz / 10 Hz = 380,000 amplitude values need to be stored in the table. Furthermore, for the 1 Hz, 0.1 Hz, and 0.01 Hz elementary waves, 3.8 M, 38 M, and 380 M amplitude values need to be stored in the table. As described above, the amount of amplitude data of the elementary wave increases as the frequency decreases. In order to improve fading accuracy or to increase the periodicity of the fading to be simulated, it is desirable to prepare and simulate many elementary waves. However, if this is done, a very large amount of data must be prepared in advance, which is inconvenient from the viewpoint of simply simulating fading.

According to the second method, there is no problem that the amount of data increases remarkably, but there is a possibility that inconvenience may arise in terms of constructing a special low-pass filter. For example, when a filter having such a transfer characteristic is constructed by a finite impulse response (FIR) filter, a very large number of taps is required, and the circuit scale becomes large. When a filter is constructed with an infinite impulse response (IIR) filter, the circuit scale may not be as large as that of an FIR filter, but another problem arises that it is difficult to find an optimal solution for the tap coefficient. Furthermore, it is desirable that the spectrum having a frequency exceeding ± f _D is completely blocked. However, in an actually constructed low-pass filter, a frequency exceeding ± f _D is slightly reduced due to imperfection of the transfer characteristic. in some but would allow, or steepness in the vicinity of ± f _D is deteriorated. Therefore, there is a concern that fading cannot be accurately simulated, and the fading evaluation accuracy is deteriorated.

  The present invention has been made to cope with at least one of the above-mentioned problems, and the problem is to obtain a fading simulator and a multipath fading simulator that simulate the influence of Doppler shift in a received signal with high accuracy. It is.

The fading signal generator used in one embodiment is:
Noise generating means for outputting a Gaussian noise signal;
Low pass filter means connected to the noise generating means;
Memory means for storing amplitude data over a predetermined period for each of a plurality of predetermined frequencies belonging to the transition region of the low-pass filter means ;
A fading signal generator for simulating the effect of Doppler shift, comprising: a synthesizing unit that synthesizes the signal from the low-pass filter unit and the signal extracted from the memory unit and outputs a fading signal. There is .

  According to the present invention, the influence of the Doppler shift in the received signal can be simulated with high accuracy.

  In one aspect of the present invention, low-pass filter means for band-limiting a Gaussian noise signal, memory means for storing amplitude data over a period of ¼ period or more for each of a plurality of predetermined Doppler frequencies, and the low-pass filter Combining means for synthesizing the first signal from the band-pass filter means and the second signal from the memory means and outputting a fading signal is provided to simulate the effect of Doppler shift.

  Since the fading signal is generated by combining the first and second signals, the fading state can be more accurately simulated than when the fading signal is generated by only one of the first and second signals. it can.

  According to an aspect of the present invention, the plurality of frequencies are frequencies belonging to a transition region of the low-pass filter. Since it is not necessary to prepare amplitude data related to the elementary wave in the pass band on the lower frequency side than the transition band, the amount of data to be stored is small. Further, since the first signal only needs to represent a signal in the vicinity of a flat passband, the characteristics of the transition band need not be matched to Doppler fading. That is, the low-pass filter characteristic can be formed in a trapezoidal shape like a normal band-limiting filter, so that the low-pass filter can be easily created with a digital filter having a relatively small circuit scale.

  The second signal represents a signal having a deviation of about the maximum Doppler frequency. According to an aspect of the present invention, the amplitude data is set such that signal power changes according to characteristics of the low-pass filter. According to an aspect of the present invention, the amplitude data is set such that the maximum amplitude value increases as the maximum Doppler frequency is approached. According to an aspect of the present invention, the plurality of frequencies are set such that frequency intervals change according to characteristics of the low-pass filter. According to an aspect of the present invention, the plurality of frequencies are set such that the frequency intervals become closer as the maximum Doppler frequency is approached. Adjustment of the amplitude, frequency interval, and the like can be easily performed without requiring a significant increase in the amount of stored data and the circuit scale.

According to an aspect of the invention, an upsampler means connected to at least one of the low pass filter, the memory means and the synthesis means is provided. The up-sampler means outputs a signal in which data points having a predetermined value are inserted into the input signal at a predetermined frequency. As a result, the first signal and the second signal can be generated at a frequency lower than the sampling frequency f _S of the main signal. For this reason, the fading cycle can be prolonged, the amount of stored data can be reduced, and the like.

  According to one aspect of the present invention, a multipath fading simulator that adjusts and synthesizes the amplitude and phase of a plurality of signals from a plurality of fading simulators to simulate the effect of Doppler shift in a multipath propagation environment is used. One or more of the plurality of fading simulators are the above fading simulators. Thereby, a signal affected by both multipath fading and Doppler fading can be simulated with high accuracy.

  According to one aspect of the present invention, in the multipath fading simulator, one noise generating unit is commonly used for two or more fading simulators. According to one aspect of the present invention, one noise generating means and one low-pass filter are commonly used for two or more fading simulators.

  FIG. 2 shows an overall view of a multipath fading simulator according to an embodiment of the present invention. The multipath fading simulator apparatus 200 includes N delay units 202-1 to N, attenuators 204-1 to N connected to the N delay units, and fading connected to the N attenuators, respectively. Simulator devices 206-1 to 206-N and synthesis units 208-R and I that synthesize real and imaginary components of the complex signal, respectively. Each of fading simulator apparatuses 206-1 to 206-N includes a fading signal generator 210 and combining sections 212-R and I.

Each of the delay devices 202-1 to N receives an input signal for a predetermined time τ ₁ , τ ₂ ,. . . , Τ _N respectively. The input signal is also called a main signal, is expressed in a complex number format, and has a real component and an imaginary component. The real and imaginary components of the input signal are delayed separately. Thereafter, similarly, the real component and the imaginary component are processed separately.

Each of the attenuators 204-1 to 204-N converts the amplitudes of the signals input to them into predetermined adjustment amounts L ₁ , L ₂ ,. . . , L Adjust to match _N. The delay amounts τ ₁ , τ ₂ ,. . . , Τ _N and predetermined adjustment amounts L ₁ , L ₂ ,. . . , L _N are set so as to reproduce the multipath propagation characteristics of the received signal. The number of paths assumed is N.

  Although the input / output signals of the fading simulator apparatuses 206-1 to 206-1 are different from each other, since they have the same configuration and function, the fading simulator apparatus 206-1 will be described as a representative example. Fading simulator apparatus 206-1 receives a signal whose amplitude and phase are appropriately adjusted by delay unit 202-1 and attenuator 204-1. Further, the fading signal generator 210 in the fading simulator device 206-1 generates a fading signal that represents the influence of fading. The signal whose amplitude and phase are adjusted and the fading signal are synthesized by the synthesis unit 212 for each real component and imaginary component.

  Combining sections 208-R and I synthesize the signals output from the fading simulator devices 206-1 to 206-N for each real number component and imaginary number component, and output the combined signal. This output signal represents a received signal that is affected by fading due to Doppler shift. Thereafter, system design and the like are performed using this output signal.

  FIG. 3 shows a detailed view of the fading simulator device 206-1. As described above, the main signal whose phase and amplitude are adjusted is input to the fading simulator device 206-1, and the influence of fading is reflected on the main signal by synthesizing the main signal with the fading signal.

The fading signal generator 210 generally includes a first signal generation unit 302, a second signal generation unit 304, and synthesis units 306-R and I. The first signal generation unit 302 includes Gaussian noise generation units 322-R, I and low-pass filters (LPF) 324-R, I. The second signal generation unit 304 includes M look-up tables (LUTs) 340-1 to 340-1 M having amplitude data related to each of the M elementary waves, and 2M amplitude adjustment units A _{1R, I 1 to} A _{MR, I.}

  The first signal generation unit 302 generates and outputs a first signal that is a complex signal. The real component of the first signal is created by waveform shaping or band limiting the white Gaussian noise (WGN) signal from the Gaussian noise generator 322-R with the low-pass filter 324-R. Similarly, the imaginary component of the first signal is created by shaping the waveform of the output from the Gaussian noise generator 322-I using the low-pass filter 324-I. Since the Gaussian noise generators 322-R and I have the same configuration and function, both can be shared. However, from the viewpoint of extending the fading period and improving the fading evaluation accuracy, it is desirable to prepare them separately for each of the real number component and the imaginary number component.

  The low-pass filters 324-R and I have a normal trapezoidal transfer characteristic as shown in FIG. A low-pass filter having a transfer characteristic as shown in FIG. 4A can be easily configured with a small FIR filter or IIR filter having a relatively small number of taps. Therefore, the low-pass filters 322-R and I used in the present embodiment are greatly different from the low-pass filter having a special transfer characteristic as shown in FIG.

The second signal generation unit 304 outputs the real and imaginary components of the M elementary waves. In the present embodiment, a signal is output according to the sampling frequency f _S = 3.8 MHz. Each table 304-1 to M stores amplitude values of sine waves (elementary waves) corresponding to the respective frequencies. These amplitude values are prepared so that amplitude values for one period are sequentially output according to the sampling frequency f _S during one period of the elementary wave. For example, for the rays of frequency _{f i, k i = f S} / f i number of amplitude values are stored in the table 304-i. However, i = 1,. . . , M. The frequency f _i of each individual wave is determined by:
f _i = f _D × cos (α _i )
Here, f _D is the maximum Doppler frequency, and α _i is an angle formed by the traveling direction of the moving body and the direction in which the radio wave arrives. For convenience, α ₁ ,. . . , Α _{M in} the order of decreasing angles (the Doppler frequency increases in the order of f ₁ ,..., F _M ). Since the number of stored amplitude values k _i = f _S / f _i is inversely proportional to the frequency f _i of the elementary wave, the smaller the frequency f _i , the more amplitude values need to be stored in the table 304-i. . In this example, at most M Doppler frequencies f _M , f _M−1 ,. . . , F ₁ are merely prepared. Therefore, it is not necessary to prepare an amplitude value for a very small Doppler frequency in the table. As will be described later, the frequencies of the M elementary waves belong to the transition region of the low-pass filters 324-R and 324, and a table of elementary waves belonging to the pass region (frequency band lower than the transition region) is unnecessary. For this reason, the scale or storage capacity of the table prepared for each elementary wave can be smaller than that required in the first conventional method. Where to set the minimum frequency f ₁ can be set as appropriate according to the application. However, it should be noted that the amount of data stored in the table will increase excessively if the frequency is set too low.

  In this embodiment, from the viewpoint of reducing the amount of data stored in the table, a table common to the real number component and the imaginary number component is referred to using the property that the sine wave and the cosine wave are shifted in phase by π / 2 radians. Has been. Furthermore, in consideration of the symmetry of the sine wave, it is only necessary to store sine wave data having a quarter period or more in the table.

Amplitude adjusters A _{1R, I 1,.} . . , A _{MR, I} appropriately adjust the amplitude values stored in the table and output them. When the adjustment contents for the real number component and the imaginary number component are the same, the values after amplitude adjustment may be stored in the respective tables 340-1 to 340 -M. By doing so, the amplitude adjusting unit can be omitted.

The operation in this embodiment will be described with reference to FIGS. In general, the fading simulator needs to operate at a sampling frequency f _S or higher of the main signal. In this embodiment, it is assumed that the operation is performed at the sampling frequency f _S or more of the main signal. Gaussian noise generation portion 322-R in the first signal generating unit 302, I, respectively, and output the noise signal in accordance with the sampling frequency _{f S.} Noise signal is distributed over the entire frequency range including a ± f _D. A waveform of this noise signal is shaped by the low-pass filters 324-R, I, whereby the first signal is output and input to the synthesis units 306-R, I. The first signal has a power spectrum as shown in FIG.

The second signal generation unit 304 extracts an amplitude value from the tables 340-1 to 340 -M for each of the M elementary waves near the maximum Doppler frequency f _D according to the sampling frequency f _S. The extracted amplitude values are stored in the amplitude adjustment units A _{1R, I 1,.} . . , _{AMR, I} are output after being adjusted to an appropriate value, and input to the synthesis unit 306-R, I. Amplitude adjusters A _{1R, I 1,.} . . , A _{MR, I} are adjusted in consideration of the shape of the filter characteristics (FIG. 4A) in the transition region of the low-pass filters 324-R, I. In the example shown in the simplified illustration, the amplitude is adjusted so that the power of the M elementary waves gradually increases in accordance with the filter characteristic that gradually decreases linearly in the transition region (FIG. 4A). Has been.

  The first and second signals from the first and second signal generation units 302 and 304 are combined by the combining units 306-R and I. FIG. 4C shows the power spectrum of the combined signal. As a result, as shown in FIG. 1, a signal is generated that has a power spectrum that is flat in the vicinity of the center frequency but increases steeply as it deviates from the center frequency, and becomes 0 when the maximum Doppler frequency is exceeded. Such a signal accurately simulates the influence of the Doppler shift, and this signal affects the main signal by the synthesizer 212, and a signal affected by the Doppler shift is output.

  In the above operation, it has been described that the first signal is created, the second signal is created, the first and second signals are combined, and introduced into the main signal. However, the order of such operations is not essential, and the order of generating the first and second signals may be reversed, and all or part of the signal generation procedure may be performed simultaneously. Further, all or a part of the first signal, the second signal, and the main signal may be synthesized simultaneously.

FIG. 5 illustrates another operation example in the present embodiment. In this embodiment, the amplitude adjusters A _{1R, I 1,.} . . , _{AMR, I} are not essential, so they are omitted. As described with reference to FIG. 4, the first signal generation unit 302 generates a first signal. Second signal generating unit 304, for each of the rays in the vicinity of the maximum Doppler frequency _{f D,} extracts an amplitude value from the table 340-1～M. In this case, the number of elementary waves synthesized by the synthesis unit 306 is adjusted in accordance with the characteristic shape of the transition region of the low-pass filters 324-R, 324 (FIG. 5A). That is, M elementary waves f ₁ ,. . . , F _M of P elementary waves smaller than M are selected. Alternatively, only P tables may be prepared from the beginning. In the example shown in the simplified diagram, the elementary waves are selected so that the number density of the elementary waves changes from sparse to dense in accordance with the filter characteristics that gradually decrease linearly in the transition region (FIG. 5). (B)). Thus, even if the number of elementary waves is adjusted, the power density spectrum obtained by combining the first signal and the second signal can be approximated as shown on the right side of FIG.

FIG. 6 is a diagram illustrating a fading signal generator according to an embodiment of the present invention. The fading signal generator 610 may be used in place of the fading signal generator 210 of FIG. The fading signal generator 610 generally includes a first signal generation unit 602, a second signal generation unit 604, and addition units 606-R and I. The first signal generation unit 602 includes Gaussian noise generation units 622-R, I and low-pass filters (LPF_L0) 624-R, I. The second signal generation unit 604 includes M tables 640-1 to 640 -M having amplitude data regarding each of M elementary waves, and 2M amplitude adjustment units A _{1R, I to} AMR _{, I.} Since these elements have already been described in the first embodiment, they will not be further described.

The first signal generation unit 602 in the present embodiment is provided with L interpolators for each of the real component and the imaginary component, and each of the interpolators includes an upsampler UR _Li and a low-pass for interpolation processing. It consists of a filter LPF_Li. However, i = 1,. . . , L. Similarly, the second signal generator 604 is also provided with H interpolators for each of the real component and the imaginary component, and each of the interpolators includes an upsampler UR _Lj and a low-pass filter for interpolation processing. It consists of LPF_Lj. However, j = 1,. . . , H. Further, after the adders 606-R and I, T interpolators are provided for each of the real component and the imaginary component, and each of the interpolators includes an upsampler UR _Lk and a low frequency for interpolation processing. It consists of a pass filter LPF_Lk. However, k = 1,. . . , T. Since these interpolators have the same configuration and function, the interpolator having the upsampler UR _L1 and the low-pass filter LPF_L1 among the interpolators in the first signal generation means 602 represents them. Explained.

The interpolator generally increases the number of sampling data of the signal input thereto and outputs a signal equivalent to the signal sampled at the higher frequency R _L1 × f _SL while appropriately band-limiting. The upsampler UR _L1 connected to the low-pass filter 624-I (LPF_0) adds (R _L1 −1) “0” s to a series of sampling data representing an input signal. Increase the number of sampling data. When the sampling frequency of the input signal and _{f SL,} frequency after up-sampling _{is R L1} × _{f SL.} The low-pass filter LPF_L1 functions to remove frequency components that are larger than f _SL / 2 and smaller than (R _L1 −1/2) f _SL .

FIG. 7 is an explanatory diagram for explaining the operation of the interpolator (the negative region is omitted). If the maximum Doppler frequency _{f D} is 100 Hz, the sampling frequency _{f SL} is 400 Hz, the frequency division ratio _{R L1} is 2. The signal generated from the Gaussian noise generator 622-I and filtered with LPR_LO has frequency components up to 100 Hz. This signal, when reproduced sampled at 400Hz sampling frequency _{f SL,} similar power spectrum for each sampling frequency _{f SL} is generated. By adding one “0” between two sampling data using the upsampler UR _L1 , the number of sampling data is doubled. Then, the sideband or image generated at an odd multiple of the sampling frequency f _SL = 400 Hz is removed by the low-pass filter LPF_L1 for interpolation processing. When the low-pass filter removes frequency components larger than f _SL / 2 = 200 and smaller than (R _L1 −1/2) f _SL = 600 Hz, a power spectrum as shown in the lower side of FIG. 7 is obtained. This power spectrum is equivalent to that obtained when the signal related to the Doppler shift is sampled at a frequency of R _L1 × f _SL = 2 × 400 = 800 Hz. By further increasing the number of stages of interpolators that operate in this way, a signal equivalent to a signal obtained when sampling at a higher frequency can be generated.

In the first signal generation unit 602 of FIG. 6, the output of the low-pass filters 624-R and _624I is raised from f _SL to f _ST by L interpolators as described above, and the final stage The output of the interpolator is input to the synthesis units 606-I and R, respectively. The division ratios of the individual interpolators are set to R _L1 , R _L2,. . . , R _LL
f _ST = R _L1 × R _L2 ×... × R _LL × f _SL
It is. In the second signal generation unit 604, the amplitude value is sampled from each table 640-1 to M at the frequency f _SH . The output of the synthesis unit 607-R, I is also raised from f _SH to f _ST by the H interpolators as described above, and the output of the interpolator at the final stage is the synthesis unit 606-I, R. Respectively. The division ratios of the individual interpolators are set to R _H1 , R _H2,. . . , R _HH ,
f _ST = R _H1 × R _H2 × ・ ・ ・ × R _HH × f _HL
It is. Thus, the combining unit 606-R, the I, so that the signals respectively sampled at the frequency _{f ST} is synthesized. Furthermore, the combining unit 606-R, to the output of the I, the interpolator to T as described above, the sampling frequency is raised from f _ST to f _S, the output of the interpolator of the last stage output as the fading signal Is done. The division ratios of the individual interpolators are expressed as R _T1 , R _T2,. . . , R _TH
f _S = R _T1 × R _T2 ×... × R _TT × f _ST
It is. The initial sampling frequencies f _SL and f _SH , the value of the division ratio, and the number of stages of interpolators can be set arbitrarily. It is not difficult to construct a low-pass filter that blocks a band around an integer multiple of a certain sampling frequency. Therefore, with respect to the first signal, the Gaussian noise generators 622-R and I can be operated at a low frequency by appropriately setting the number of stages of the interpolator. This means that it can contribute to the extension of the fading cycle. For the second signal, the sampling frequency _{f SH} that is low, it is possible to save the amount of data to be stored in a table 640-1～M. This is because the amount of data needed to build the sine wave of a frequency f _i is because becomes (sampling frequency f _{SH) /} (the frequency f _i) pieces. The value of the frequency division ratio may be an arbitrary value, but from the viewpoint of simplifying the configuration, all the frequency division ratios may be set to 2.

  FIG. 8 shows a multipath fading simulator according to an embodiment of the present invention. A multipath fading simulator 800 includes N delay units 802-1 to 802-1, attenuators 804-1 to N connected to the N delay units, and a fading simulator connected to each of the N attenuators. Devices 806-1 to N and combining units 808-R and I are included. Each of the fading simulators 806-1 to 806 -N includes a fading signal generator 810 and a combining unit 812 -R, I. In the present embodiment, one Gaussian noise generator 814 and a serial-parallel converter (S / P) 816 used for N fading simulator apparatuses are provided. The first signal generation unit is provided with low-pass filters 824-R, I and interpolators 826-R, I. The first and second signals from the first and second signal generation units are combined by the combining units 828-R and I and then reflected on the main signal by the combining units 818-R and I.

Each of the delay devices 802-1 to 802 -N receives an input signal for a predetermined time τ ₁ , τ ₂ ,. . . , Τ _N respectively. Each of the attenuators 804-1 to 804 -N converts the amplitudes of the input signals into predetermined adjustment amounts L ₁ , L ₂ ,. . . , L Adjust to match _N. The delay amounts τ ₁ , τ ₂ ,. . . , Τ _N and predetermined adjustment amounts L ₁ , L ₂ ,. . . , L _N are set so as to reproduce the multipath propagation characteristics of the received signal. The number of paths assumed is N.

  Although the fading simulator devices 806-1 to 806-1 are different from each other in input / output signals, the fading simulator device 806-1 will be described as a representative because they have the same configuration and function. Fading simulator apparatus 806-1 receives a signal whose amplitude and phase are appropriately adjusted by delay device 802-1 and attenuator 804-1. Further, the fading signal generator 810 in the fading simulator device 806-1 generates a fading signal representing the influence of fading. The amplitude and phase-adjusted signal and the fading signal are synthesized for each real component and imaginary component by the synthesis units 812-R and I.

  The synthesizing units 808-R and I synthesize the signals output from the respective fading simulator devices 806-1 to 806-1N for each real number component and imaginary number component, and output the synthesized signal.

Gaussian noise generator 814, with 2N multiple of the sampling frequency of a predetermined frequency f _SL, and generates a white Gaussian noise signal.

  The serial-parallel converter 816 is provided between the Gaussian noise generator 814 and the N fading simulator apparatuses 806-1 to 806-1. The serial-to-parallel converter 816 converts the serial signal input thereto into 2N parallel signals, and supplies the parallel signals to the fading simulator devices 806-1 to 806-1-N two by two. .

Also in this embodiment, as in the multipath fading simulator of FIG. 3, the first signal and the second signal are generated and introduced into the main signal, so that the signal affected by fading can be simulated. it can. The Gaussian noise generator 814 used when generating the first signal is commonly used for the N fading signal generators 810. However, in order to ensure the independence or non-correlation of the Gaussian noise signal given to different fading simulator devices (further, the real and imaginary component sides of the first signal generation unit) They are distributed to each fading simulator device. Since the Gaussian noise signal generator generates a signal at a sampling frequency of 2N × f _SL , the sampling frequency of the signal applied to the low-pass filters 824-R and I of the first signal generation unit is f _SL respectively. . The low-pass filters 824-R, I perform band limitation and provide outputs to the interpolators 826-R, I. Interpolators 826-R, I function as one or more of the interpolators described with respect to FIG. Accordingly, the sampling frequency f _SL of the output signal of the low-pass filter 824-R, I is appropriately raised by the interpolator 826-R, I. The sampling frequency _{f SL} is equal to the sampling frequency _{f S} of the main signal, interpolator 826-R, I is omitted. The output of the interpolator 826-R, I is combined with the output signal from the second signal generation unit by the combining unit 828-R, I, and further combined with the main signal by the combining unit 812.

  FIG. 9 shows a multipath fading simulator according to an embodiment of the present invention. In the illustrated example, an element corresponding to the first signal generation unit is commonly used for N fading signal generators. The multipath fading simulator 900 includes N delay units 802-1 to 802-1, attenuators 804-1 to N connected to the N delay units, and fading units connected to the N attenuators, respectively. Simulator devices 806-1 to 806 -N and synthesis units 808 -R and I that synthesize real and imaginary components of complex signals, respectively. Each of the fading simulators 806-1 to 806 -N includes a fading signal generator 810 and a synthesis unit 812. In this embodiment, a Gaussian noise generator 814, a low-pass filter (LPF) 825, an interpolator 827, and a serial-parallel converter (S / P) 816 are commonly used for N fading simulator apparatuses. Is done.

The Gaussian noise generator 814 outputs a signal at a sampling frequency of 2N × f _SL . The low-pass filter 825 limits the band of the input signal and supplies the output to the interpolator 827. The low-pass filter 825 can be composed of an FIR filter as shown in FIG. 10, for example. In FIG. 10, D ^CH represents a delay element that realizes a delay of ^CH clocks, and k ₀ ,. . . , K _Q−1 represents a tap coefficient. The interpolator 827 receives a signal sampled at a frequency of 2N × f _SL , and outputs a signal equivalent to the signal sampled at R × 2N × f _S. R represents the frequency division ratio of upsampling performed by the interpolator (R = f _S / f _SL ).

As an example, signals of A _i , B _i (i = 1, 2,...) 2 series or 2 channels as shown in FIG. 11 are input to the interpolator 827 and are upsampled by R = 3 times. Shall. The number of channels and the frequency division ratio can be arbitrarily set. In the interpolator 827, as shown in FIG. 12, (number of channels) × (R−1) = 2 × (3-1) = 4 “0” points are respectively present after the signals A _i and B _i. Be placed. Note that the time interval between signal A _i and signal B _i is also narrowed. This signal is appropriately band-limited and then supplied to each fading signal generator 810 via a serial-to-parallel converter 816. As a result, a signal equivalent to a signal sampled at a frequency of R × 2N × f _SL = 6N × f _SL is formed. In fading signal generator 810, a signal equivalent to the signal sampled at frequency f _S is combined with the second signal by combining sections 828-R and I, and the combined signal is combined in combining section 812 of the fading simulator device. And synthesized with the main signal. Each output of the fading simulator device represents the influence of fading for each path, and they are combined by the combining sections 808-R and I, and finally, a received signal affected by the fading is output.

  Hereinafter, the means taught by the present invention will be listed as an example.

(Appendix 1)
Noise generating means for outputting a Gaussian noise signal;
Low pass filter means connected to the noise generating means;
Memory means for storing amplitude data over a predetermined period for each of a plurality of predetermined frequencies;
A fading simulator for simulating the influence of Doppler shift, comprising: a synthesizing unit that synthesizes a signal from the low-pass filter unit and a signal extracted from the memory unit and outputs a fading signal.

(Appendix 2)
The fading simulator according to claim 1, wherein the plurality of frequencies are frequencies belonging to a transition region of the low-pass filter.

(Appendix 3)
The fading simulator according to claim 1, wherein the amplitude data is set such that signal power changes according to characteristics of the low-pass filter.

(Appendix 4)
The fading simulator according to appendix 3, wherein the amplitude data is set such that the maximum amplitude value increases as the maximum Doppler frequency is approached.

(Appendix 5)
The fading simulator according to claim 1, wherein the plurality of frequencies are set such that frequency intervals change according to characteristics of the low-pass filter.

(Appendix 6)
The fading simulator according to appendix 5, wherein the plurality of frequencies are set such that the frequency interval becomes closer as the maximum Doppler frequency is approached.

(Appendix 7)
Up-sampler means connected to at least one of the low-pass filter, the memory means and the synthesis means, the up-sampler means comprising:
The fading simulator according to claim 1, wherein a signal in which data points having a predetermined value are inserted into the input signal at a predetermined frequency is output.

(Appendix 8)
A multipath fading simulator that adjusts and synthesizes the amplitude and phase of a plurality of signals from a plurality of fading simulators to simulate the effect of Doppler shift in a multipath propagation environment,
One or more of the plurality of fading simulators are the fading simulator according to appendix 1, wherein the multipath fading simulator is characterized in that

(Appendix 9)
The multipath fading simulator according to appendix 8, wherein one noise generation means is commonly used for two or more fading simulators.

(Appendix 10)
The multipath fading simulator according to appendix 8, wherein one noise generating means and one low-pass filter are commonly used for two or more fading simulators.

It is a figure which shows the influence of fading by a Doppler shift. 1 shows an overall view of a multipath fading simulator according to an embodiment of the present invention. FIG. It is a figure which shows the fading simulator apparatus by one Example of this invention. It is explanatory drawing for demonstrating the operation | movement in a present Example. It is explanatory drawing for demonstrating another operation | movement in a present Example. FIG. 3 is a diagram illustrating a fading signal generator according to an embodiment of the present invention. It is explanatory drawing for demonstrating operation | movement of an interpolator. 1 shows an overall view of a multipath fading simulator according to an embodiment of the present invention. FIG. 1 shows an overall view of a multipath fading simulator according to an embodiment of the present invention. FIG. It is a figure which shows the FIR filter which comprises a low-pass filter. It is a figure which shows the signal sequence before up-sampling. It is a figure which shows the signal series after up-sampling.

Explanation of symbols

200 multipath fading simulator; 202-1 to N delay unit; 204-1 to N attenuator; 206-1 to N fading simulator apparatus; 208-R, I synthesis unit; 210 fading signal generator; 212-R, I Synthesis unit;
302 first signal generation unit; 304 second signal generation unit; 306-R, I synthesis unit; 3322-R, I Gaussian noise generator; 324-R, I low-pass filter; 340-1 to M lookup table A _{1I, 1R to} A _{MI, MR} amplitude adjustment unit;
610 fading signal generator; 602 first signal generation unit; 604 second signal generation unit; 606-R, I synthesis unit; 607-R, I synthesis unit; 624-R, I low-pass filter; M lookup table;
800 multipath fading simulator; 802-1 to N delay unit; 804-1 to N attenuator; 806-1 to N fading simulator apparatus; 808-R, I synthesis unit; 810 fading signal generator; 814 Gaussian signal generator; 816 serial-parallel converter; 824-R, I low-pass filter; 826-R, I interpolator; 825 low-pass filter; 827 interpolator; 828-R, I combiner

Claims (4)

Noise generating means for outputting a Gaussian noise signal;
Low pass filter means connected to the noise generating means;
Memory means for storing amplitude data over a predetermined period for each of a plurality of predetermined frequencies belonging to the transition region of the low-pass filter means ;
A fading signal generator for simulating the influence of Doppler shift, comprising: a synthesizing unit that synthesizes a signal from the low-pass filter unit and a signal extracted from the memory unit and outputs a fading signal . 2. The fading signal generator according to claim 1, wherein the amplitude data is set such that signal power changes in accordance with characteristics of the low-pass filter unit . The low-pass filter means, said memory means and the fading signal generator of claim 1, wherein the Ru with the upsampler means connected to at least one of said combining means. A multipath fading simulator that adjusts and synthesizes the amplitude and phase of a plurality of signals from a plurality of fading simulators to simulate the effect of Doppler shift in a multipath propagation environment,
Each of the plurality of fading simulators generates the plurality of signals by multiplying and outputting the fading signal from the fading signal generator according to claim 1 and the signal input to the fading simulator. A multi-path fading simulator.

Images