JP4566892B2 - Apparatus and method for calculating amplitude of complex signal - Google Patents

Description

  The present invention relates to digital signal processing, and more particularly to an apparatus for calculating the amplitude of a complex signal given real and imaginary components.

As an algorithm or circuit mounting method for calculating (I ² + Q ² ) ^1/2 which is the magnitude of the complex signal (amplitude of the complex signal) such as I + jQ, which is a general calculation in digital signal processing, I ² + Q ² A method of substituting (method 1) and a method using the αMAX + βMIN method (method 2) are known.

(Method 1) The circuit using the method 1 has high calculation accuracy but has a problem that the circuit scale is large. That is, in the circuit using Method 1, for example, when a 10-bit real component (real component) I and imaginary component (imaginary component) Q are input, a total of 21 adders are used to calculate I ² + Q ^2. Necessary. In addition, the maximum bit width of I ² + Q ² is required (bit width of input complex signal) × 2 + 1 bits, and the circuit scale increases.

  (Method 2) Method 2 (αMAX + βMIN method) is a method widely used in digital signal processing, and Non-Patent Document 1 discloses a complex signal amplitude approximation calculation circuit using this αMAX + βMIN method.

FIG. 7 is a block diagram showing a circuit configuration in the case where constants (α = 15/16, β = 15/32) in a complex signal amplitude approximation calculation circuit using the αMAX + βMIN method. In the αMAX + βMIN method, the constant α And β may be set to other values. When a complex signal I + jQ is input to this circuit 127, absolute values AI and imaginary component Q of real component I are first calculated by absolute value calculation units (abs) 128 and 129. The absolute value AQ is calculated, and the output AI of the absolute value calculation unit 128 is input to the register 133, the integration unit (1-bit shift means) 131, and the comparison unit 130. The integration unit 131 calculates the absolute value of the input real component I. The value AI is multiplied by 1/2 and output to the register 134.

  The output AQ of the absolute value calculation unit 129 is input to the register 135, the integration unit (1 bit shift means) 132, and the comparison unit 130. The accumulator 132 multiplies the absolute value AQ of the input imaginary component Q by 1/2 and outputs it to the register 136.

  The comparison unit 130 compares the absolute value AI of the real component I with the absolute value AQ of the imaginary component Q, and outputs a signal indicating whether AI> AQ to the registers 133, 134, 135, 136. . The register 133 outputs AI to the adder (adding means) 137 when AI> AQ. The register 134 outputs AI / 2 to the adder 137 when AI ≦ AQ. The register 135 outputs AQ / 2 to the adder 137 when AI ≦ AQ. The register 136 outputs AQ to the adder 137 when AI> AQ.

  Therefore, adder 137 receives AI and AQ / 2 when AI> AQ, and outputs AI + AQ / 2. Adder 137 receives AI / 2 and AQ when AI ≦ AQ, and outputs AI / 2 + AQ.

The output of the adder 137 is branched into two, one being input to the adder 139 and the other being input to the adder 139 via the integrating unit (4-bit shift means) 138. The accumulating unit 138 multiplies AI + AQ / 2 or AI / 2 + AQ input from the adder 137 by 1/16, inverts the sign, and outputs the result to the adder 139. As a result, the adder 139 calculates 15 / 16MAX + 15 / 32MIN as an approximate value of (I ² + Q ² ) ^1/2 .

  FIG. 8 is a graph showing the calculation accuracy of the complex signal amplitude approximation calculation circuit 127 shown in FIG. As shown in this figure, the average error is about 3.1% and the maximum error is 6.3%.

As shown in FIG. 7, the complex signal amplitude approximation calculation circuit 127 includes six bit shifts (absolute value calculation units 128 and 129, a comparison unit 130, integration units 131, 132, and 138), and two adders (adders). 137, 139). Therefore, the circuit scale of the complex signal amplitude approximation calculation circuit 127 is approximately 1/10 compared with the method (method 1) substituting I ² + Q ² , and the bit width may be the bit width of the input complex signal. .
Richard G. Lyons, “Understanding Digital Signal Processing (2nd Edition)”, Prentice Hall, 2004, pp.479 §13.2 HIGH-SPEED VECTOR APPROXIMATION

  However, when the method 1 is used, the calculation scale is high, but the circuit scale is large, and when the method 2 is used, the calculation precision is lowered although the circuit scale is small.

  The present invention has been made in view of such a problem, and an object of the present invention is to make the circuit scale smaller than that of the complex signal amplitude calculation apparatus using the method 1 and to detect the complex signal using the method 2. An object of the present invention is to provide a complex signal amplitude calculation device with higher calculation accuracy than the amplitude calculation device.

Amplitude calculation apparatus of the complex signal of the present invention, in order to solve the above problems, an amplitude calculation apparatus of the complex signal for calculating the amplitude approximation of the complex signal C _in, a complex signal C _in input a phase calculating means (the M to 2 or greater natural number) M pieces obtained by the phase rotation outputs a complex signal C ₀ to C _M-1 at different rotation amounts from each other, the complex signal C ₀ to C _M-1, respectively , The absolute values of the real and imaginary components were compared, and the larger component (MAX) multiplied by a constant α and the smaller component (MIN) multiplied by a constant β were added. Arithmetic processing means for calculating an approximate amplitude value (α × MAX + β × MIN), and average value calculating means for calculating an average value of the approximate amplitude values calculated for each of the complex signals C _{0 to} C _M−1. It is characterized by having. Note that the M complex signals whose phases are rotated by different rotation amounts may include complex signals whose phases are rotated by 0 °, that is, complex signals that are not subjected to phase rotation.

The complex signal amplitude calculation method of the present invention is a complex signal amplitude calculation method for calculating an approximate value of the amplitude of the complex signal C _{in in} order to solve the above-described problem. a phase calculation step of outputting _M (M is a natural number of 2 or more) complex signals C _{0 to} C _{M−1 in} which _in is phase-rotated by different rotation amounts; and the complex signals C _{0 to} C _M−1. For each of the above, compare the magnitudes of the absolute values of the real and imaginary components, and approximate the amplitude by adding the larger component multiplied by the constant α and the smaller component multiplied by the constant β. An arithmetic processing step for calculating, and an average value calculating step for calculating an average value of the approximate amplitude values calculated for each of the complex signals C _{0 to} C _M−1 .

In the conventional configuration _{in which} the approximate amplitude value of the complex signal C _in inputted without phase rotation is calculated by the α × MAX + β × MIN method, when C _in = I _in + jQ _in = cos θ + j sin θ, 0 The values were smaller than the ideal values in the vicinity of °, 45 °, and 90 °, and conversely, the values were larger than the ideal values in the vicinity of 22.5 ° and 67.5 °.

On the other hand, according to the amplitude calculation apparatus and the amplitude calculation method described above, amplitude approximation values of M complex signals obtained by rotating the phase of the input complex signal C _in by different rotation amounts are calculated. Is averaged. As a result, a value smaller than the ideal value and a value larger than the ideal value can cancel each other, so that the calculation accuracy can be improved and the fluctuation of the approximate amplitude value of the complex signal can be suppressed. In addition to the conventional configuration, only the phase calculation means and the average value calculation means need be provided, so that the circuit scale does not increase significantly.

For example, the phase calculating means is configured such that the rotation amount θ _{0 to} θ represented by θ _k = kπ / (2M) when _k is an integer from 0 to M−1 for the input complex signal C _{in. M-1} only M complex signal C ₀ through may output a C _M-1 obtained by phase rotation.

  According to said structure, the value smaller than an ideal value and the value larger than an ideal value in the amplitude approximate value of said M complex signal can be negated more suitably. Therefore, the calculation accuracy can be further improved, and fluctuation of the amplitude approximate value of the complex signal can be more suitably suppressed.

Further, the phase calculation means includes M phase rotation means, and the M complex signals C ₀ are rotated by phase rotation of the input complex signal C _in by different rotation amounts by the phase rotation means. Thru | or the structure which calculates _CM-1 may be sufficient.

  The complex signal after phase rotation has a rotation amount of θ,

Is obtained by calculating. These multiplications can be realized by bit shift addition. Therefore, each phase rotation means can be realized with a small circuit scale. That is, according to the above configuration, the calculation accuracy can be improved and the fluctuation of the approximate amplitude value of the complex signal can be suppressed without significantly increasing the circuit scale.

In addition, the phase calculation means includes phase rotation means capable of changing the rotation amount of the phase rotation, and sequentially changes the rotation amount when the input complex signal C _in is rotated in phase. The complex signals C _{0 to} C _M−1 may be sequentially calculated and output.

  According to the above configuration, by providing the phase rotation means capable of varying the rotation amount, the number of phase rotation means can be reduced as compared with the case where the phase rotation means is provided for each rotation amount without reducing the calculation accuracy. The circuit scale can be reduced.

In the above configuration, the same number of arithmetic processing means as the number of phase rotation means, and storage means for storing the approximate amplitude values calculated by the arithmetic processing means for each of the complex signals C _{0 to} C _M−1. The calculation means may be configured to sequentially calculate the amplitude approximation values of the complex signals C _{0 to} C _M−1 sequentially input from the phase rotation means and sequentially store them in the storage means. .

  According to the above configuration, it is only necessary to provide the same number of arithmetic processing means as the number of phase rotation means, so that the circuit scale of the amplitude calculation apparatus can be further reduced without reducing the calculation accuracy.

Further, there may be provided an intensity conversion means for multiplying the average value calculated by the average value calculation means by a constant γ, and a value obtained by multiplying the average value by the constant γ is calculated as the amplitude of the input complex signal C _in. .

  According to said structure, the calculation result of the amplitude of a complex signal can be brought close to an ideal value by multiplying the average value by the constant γ by the intensity conversion means. Therefore, the calculation accuracy of the amplitude of the complex signal can be further improved. Here, the value of the constant γ is not particularly limited, and may be set as appropriate so as to minimize the error, for example.

  Further, the value of the constant α may be 15/16, and the value of the constant β may be 15/32. By setting the constants α and β in this way, the amplitude of the complex signal can be calculated with high accuracy.

  Further, the constants α and β may have a relationship of α = 2 × β. By setting the constant α and the constant β so as to satisfy the relationship of α = 2β, the circuit scale can be made smaller than that in the case where the constants α and β are both arbitrary in the arithmetic processing means.

  In this case, for example, the value of the constant α may be 1. When the constant α = 1, the value of the constant β is ½. Since 1/2 multiplication can be realized by 1-bit shift, the circuit scale can be further reduced.

  Alternatively, the value of the constant α may be 2. Also in this case, the multiplication of 2 can be realized by 1-bit shift, so that the circuit scale can be reduced.

  The value of M may be 3. According to said structure, the fluctuation | variation of an amplitude approximate value can be made into 1/4 times or less of the case where M = 2. Further, since only three amplitude approximate values used for calculating the average value need be calculated, the calculation in the phase calculating means is easy, and the increase in circuit scale can be minimized.

As described above, the amplitude calculation apparatus according to the present invention has M (M is a natural number of 2 or more) complex signals C _{0 to} C _M− obtained by rotating the phase of the input complex signal C _in by different rotation amounts. _{For each} of the phase calculation means for outputting ₁ and the complex signals C _{0 to} C _M−1 , the absolute values of the real and imaginary components are compared, and the larger component is multiplied by a constant α, An arithmetic processing unit that calculates an approximate amplitude value obtained by adding a smaller component multiplied by a constant β, and an average value of the approximate amplitude values calculated for each of the complex signals C _{0 to} C _M−1. Average value calculating means.

In the complex signal amplitude calculation method of the present invention, M (M is a natural number of 2 or more) complex signals C _{0 to} C _M− obtained by rotating the phase of the input complex signal C _in by different rotation amounts. _{For each} of the phase calculation step of outputting ₁ and the complex signals C _{0 to} C _M−1 , the absolute values of the real component and the imaginary component are compared, and the larger component is multiplied by a constant α, An arithmetic processing step of calculating an approximate amplitude value obtained by adding a smaller component multiplied by a constant β, and calculating an average value of the approximate amplitude values calculated for each of the complex signals C _{0 to} C _M−1. And an average value calculating step.

In this way, by calculating the amplitude approximate values of the M complex signals obtained by rotating the phase of the input complex signal C _in by different rotation amounts, and averaging them, a value smaller than the ideal value is obtained. Values greater than the ideal value can cancel each other. Therefore, it is possible to improve the calculation accuracy and suppress fluctuations in the approximate amplitude value of the complex signal. In addition to the conventional configuration, only the phase calculation means and the average value calculation means need be provided, so that the circuit scale does not increase significantly.

Embodiment 1
An embodiment of the present invention will be described. FIG. 1 is a block diagram showing a configuration of a complex signal amplitude calculation apparatus (complex signal amplitude approximation calculation apparatus) 100 according to the present embodiment.

  As shown in this figure, the amplitude calculation apparatus 100 includes a phase calculation unit 104, a linear calculation unit 108, and an averaging unit 109.

  The phase calculation unit 104 includes phase rotation units (phase rotation means) 101, 102, and 103. The linear operation unit 108 includes operation processing units (linear sum operation processing means) 105, 106, and 107.

When the amplitude calculation apparatus 100 inputs the complex signal _{C in =} I in + _jQ in, the complex signal _{C in =} I in + _jQ in are inputted to the phase rotation section 101, 102 and 103.

The phase rotation unit (e ⁰ ) 101 calculates a complex signal C ₀ = C _in e ⁰ = I ₀ + jQ ₀ obtained by rotating the input C _in = I _in + jQ _in by θ = 0 °, and this C ₀ Is output to the arithmetic processing unit 105.

The phase rotation unit (e ^{jπ / 6} ) 102 has a complex signal C ₁ = C _in e ^{jπ / 6} = I ₁ + jQ ^obtained by rotating the input C _in = I _in + jQ _in by θ = π / 6 = 30 °. ₁ is calculated, and this C ₁ is output to the arithmetic processing unit 106.

The phase rotation unit (e ^{jπ / 3} ) 103 is a complex signal C ₂ = C _in e ^{jπ / 3} = I ₂ + jQ ^obtained by rotating the input C _in = I _in + jQ _in by θ = π / 3 = 60 °. ₂ is calculated, and this C ₂ is output to the arithmetic processing unit 107.

The configuration of the arithmetic processing units 105, 106, and 107 is the same as the configuration of the complex signal amplitude approximation calculation circuit 127 described above, and thus description thereof is omitted here. The arithmetic processing units 105, 106, and 107 calculate “15 / 16MAX + 15 / 32MIN” as an approximate value (amplitude approximate value) of the amplitude of the complex signal input from the phase rotation units 101, 102, and 103. That is, the arithmetic processing unit 105 calculates an approximate value a ₀ of the amplitude of the complex signal C ₀ which is input and outputs to the averaging unit 109, the arithmetic processing unit 106 input complex signal C ₁ of the amplitude approximation of calculate the a _1, the arithmetic processing unit 107 calculates an approximate value a ₂ of the complex signal C ₂ which is the input amplitude.

The averaging unit 109 calculates the average value A = (a ₀ + a ₁ + a ₂ ) / 3 of the three approximate values input from the arithmetic processing units 105, 106, and 107 and inputs the average value A to the amplitude calculation apparatus 100. The result is output as the amplitude calculation result of the complex signal C _in .

FIG. 2 is a graph showing calculation accuracy (calculation accuracy) when I _in = cos θ and Q _in = sin θ are input to the amplitude calculation apparatus 100 (when C _in = I _in + jQ _in = cos θ + jsin θ is input). The relative error err with respect to the ideal value is represented by err = A ₀ (θ) = (I _in ² + Q _in ² ) ^1/2 , and the approximate value (average value) of the complex amplitude is A _a (θ). abs (A _a (θ) −A ₀ (θ)). Here, abs (x) is the absolute value of x. Since A ₀ (θ) = 1, it can be considered that the relative error err with respect to the ideal value is normalized.

  As shown in this figure, according to the amplitude calculation apparatus 100, the calculation accuracy of the approximate value of the complex signal amplitude can be improved as compared with the measurement accuracy of the conventional complex signal amplitude approximation calculation circuit 127 shown in FIG. Specifically, the average error of the approximate value calculated by the conventional complex signal amplitude approximate calculation circuit 127 is about 3.1% and the maximum error is 6.3%, whereas the amplitude calculation apparatus 100 Accordingly, the average of the relative error err with respect to the ideal value can be 1.9%, and the maximum value can be 2.2%.

  Here, the reason why the calculation accuracy can be improved by the amplitude calculation apparatus 100 will be described. As shown in FIG. 8, the approximate value calculated by the conventional complex signal amplitude approximate calculation circuit 127 is smaller than the ideal value in the vicinity of 0 °, 45 °, and 90 °, and conversely, 22.5 °, 67. The value is larger than the ideal value in the vicinity of 5 °.

On the other hand, the amplitude calculation apparatus 100 calculates an approximate value of the complex signal amplitude when the complex signal C _in = I _in + jQ _in is rotated by 30 ° and 60 ° in phase, and 0 °, 30 °, 60 °. The average value of the approximate values when the phase is rotated is calculated. As a result, a value smaller than the ideal value and a value larger than the ideal value can cancel each other. Therefore, calculation accuracy can be improved and fluctuations in the approximate value of the complex signal amplitude can be suppressed.

  In addition, the calculation required in the phase rotation units 102 and 103 is

It is.
here,
cos (30 °) = sin (60 °) = √3 / 2 = 0.660602...
And
cos (60 °) = sin (30 °) = 1/2
It is.
For example, √3 / 2 = 0.11011101101 when √3 / 2 and ½ are expressed by a signed 12-bit binary number, for example.
1/2 = 0.100000000
These multiplications can be realized by bit shift addition. Therefore, the phase rotation units 102 and 103 can be realized with a small circuit scale.

In this embodiment, three approximate values (a ₀ , a ₁ , a ₂ ) are calculated and averaged, that is, the number of approximate values used for calculating the average value (in this embodiment, the phase rotation unit and the calculation) Although the case where the number of each processing unit (arithmetic processing circuit) M is 3 has been described, the present invention is not limited to this, and M may be a natural number of 2 or more.

FIG. 3 is a graph showing the relationship between the error of the approximate value and M. The horizontal axis is the number of approximate values used for calculating the average value (the number of combinations of the complex signal amplitude approximate calculation circuit 127) M, and the vertical axis is the complex. The fluctuation of the approximate signal amplitude is σ = max (A _a (θ) −A ₀ (θ)) − min (A _a (θ) −A ₀ (θ)). Here, max (x) is the maximum value of x, and min (x) is the minimum value of x. Since A ₀ (θ) = 1, it can be considered that σ is normalized.

As shown in this figure, it can be seen that if M is increased, the fluctuation σ of the complex signal amplitude approximate value is suppressed, and the accuracy of the complex signal amplitude approximate calculation is improved. Each phase rotation angle may be, for example, θ _k = kπ / (2M) (k is an integer from 0 to M−1). However, the rotation angle is not limited to this, and by averaging the M approximate values, a portion smaller than the ideal value and a portion larger than the ideal value of each approximate value cancel each other, and θ = 0 What is necessary is just to set suitably so that calculation accuracy may improve rather than the case where only the approximate value of is used.

  In this embodiment, each arithmetic processing unit calculates an approximate value of the amplitude of the complex signal using the α × MAX + βMIN method, and α = 15/16 and β = 15/32. The values of α and β are not limited to this, and may be changed as appropriate.

  For example, α = 1 and β = 1/2 may be set. In this case, the arithmetic processing units 105, 106, and 107 may be replaced with the arithmetic processing unit 127b shown in FIG. That is, in the conventional complex signal amplitude approximation calculation circuit 127 described above, the accumulating unit 138 and the adder 139 may be omitted, and the output of the adder 137 may be used as the approximate value calculation result.

  In this case, the circuit scale can be reduced because only one bit shift is required to calculate the multiplication of 1/2. When α = 1, β = 1/2, the relative error err from the ideal value is inferior to that when α = 15/16 and β = 15/32, but the fluctuation σ of the complex signal amplitude approximate value σ Are the same. For example, if it is only necessary to accurately detect the signal shape as in the case where the time variation of the complex baseband Z = I + jQ is detected and the peak position is detected, if the fluctuation σ of the complex signal amplitude approximation value is small, Well, the relative error err from the ideal value does not matter. Therefore, in such a case, the fluctuation σ of the complex signal amplitude approximate value can be made the same and the circuit scale can be reduced. Therefore, it is preferable to set α = 1 and β = 1/2. Alternatively, substantially the same effect can be obtained even when α = 2 and β = 1.

  Note that the values of α and β are not limited to the above numerical values, and may be set arbitrarily. For example, the configuration (bit shift amount or the like) of the integration unit 138 is changed, or a new integration unit (bit shift or the like) is added between the adder 137 and the adder 139 or between the integration unit 138 and the adder 139. For example, the values of α and β may be changed as appropriate.

  In the present embodiment, the values of α and β used by the arithmetic processing units 105, 106, and 107 are equal to each other, but the present invention is not limited to this. For example, α = 15/16 and β = 15/32 are used in a certain arithmetic processing unit, and α = 1 and β = 1/2 are used in other arithmetic processing units. The values of may be different. Thus, by setting α and β individually for each arithmetic processing unit, it is possible to improve the calculation accuracy at a specific angle. In other words, when it is desired to improve the calculation accuracy at a specific angle, the values of α and β in each calculation processing unit are set so that the average value of the calculation result of each calculation processing unit is closer to the ideal value at the angle. Should be set.

  Further, as shown in FIG. 5, an intensity conversion unit 115 that multiplies the output A of the averaging unit 109 by a constant (γ times) may be provided after the averaging unit 109. For example, as a result of performing the intensity conversion by multiplying the output A of the averaging unit 109 when M = 3 in the amplitude calculation apparatus 100 by a constant γ = 59/64, the approximate value γA after the intensity conversion is an ideal value. The average relative error err was reduced to 0.3% and the maximum value to 0.7%. Note that the value of the constant γ is not limited to 59/64, and may be changed as appropriate so that the relative error err with respect to the ideal value can be reduced.

[Embodiment 2]
Another embodiment of the present invention will be described. For convenience of explanation, members having the same functions as those shown in the first embodiment are denoted by the same reference numerals and description thereof is omitted.

  In the first embodiment, three phase rotation units and three arithmetic processing units are provided in order to calculate three approximate values. In contrast, in the present embodiment, a configuration in which the circuit scale is further reduced by providing only one phase rotation unit that can vary the amount of phase rotation (phase rotation angle) will be described.

  FIG. 6 is a block diagram showing a configuration of an amplitude calculation apparatus (complex signal amplitude approximate calculation apparatus) 110 according to the present embodiment. As shown in this figure, the amplitude calculation device 110 includes a phase rotation unit 111, a calculation processing unit (linear sum calculation processing unit) 112, a storage unit 113, and an averaging unit 114.

The phase rotation unit 111 rotates the phase of the input complex signal C _in = I _in + jQ _in and outputs it to the arithmetic processing unit 112. The phase rotation unit 111 can change the amount of phase rotation.

  Since the configuration of the arithmetic processing unit 112 is the same as the configuration of the complex signal amplitude approximation calculation circuit 127 described above, description thereof is omitted here. The arithmetic processing unit 112 calculates “15 / 16MAX + 15 / 32MIN” as an approximate value of the amplitude of the input complex signal and outputs it to the storage unit 113.

  The storage unit 113 stores approximate values that are sequentially input from the arithmetic processing unit 112, and includes, for example, a RAM or a flip-flop.

  The averaging unit 114 calculates and outputs the average value of each approximate value input from the storage unit 113.

When the amplitude calculation unit 110 inputs the complex signal _{C in =} I in + _jQ in, the complex signal _{C in =} I in + _jQ in is input to the phase rotation section 111.

The phase rotation unit (e ^jθ ) 111 first calculates a complex signal C ₀ = C _in e ⁰ = I ₀ + jQ ₀ obtained by rotating the input C _in = I _in + jQ _in by θ = 0 °. C ₀ is output to the arithmetic processing unit 112. The arithmetic processing unit 112 calculates “15 / 16MAX + 15 / 32MIN” as an approximate value of the amplitude of the input complex signal. That is, the arithmetic processing unit 112 calculates the approximate value a ₀ of the amplitude of the complex signal C ₀ input from the phase rotation unit 111 and outputs it to the storage unit 113. The storage unit 113 stores the approximate value a ₀ input from the arithmetic processing unit 112.

Next, the phase rotation unit (e ^jθ ) 111 sets C _in = I _in + jQ _in a complex signal C ₁ = C _in e ^{jπ / 6} = I ₁ + jQ ₁ obtained by rotating the phase by θ = π / 6 = 30 °. It computes, and outputs the _{C 1} to the processing unit 112. The arithmetic processing unit 112 calculates an approximate value a ₁ of the amplitude of the complex signal C ₁ input from the phase rotation unit 111 and outputs it to the storage unit 113. The storage unit 113 stores the approximate value a ₁ input from the arithmetic processing unit 112.

Furthermore, the phase rotation unit (e ^jθ ) 111 calculates a complex signal C ₂ = C _in e ^{jπ / 3} = I ₂ + jQ ₂ obtained by rotating C _in = I _in + jQ _in by θ = π / 3 = 60 °. Then, this C ₂ is output to the arithmetic processing unit 112. The arithmetic processing unit 112 calculates an approximate value a ₂ of the amplitude of the complex signal C ₂ input from the phase rotation unit 111 and outputs it to the storage unit 113. The storage unit 113 stores the approximate value a ₂ input from the arithmetic processing unit 112.

Thereafter, the storage unit 113 outputs the approximate values a ₀ , a ₁ , and a ₂ stored as described above to the average value calculation unit 114. When the approximate values a ₀ , a ₁ , and a ₂ are input, the average value calculation unit 114 calculates the average value A = (a ₀ + a ₁ + a ₂ ) / 3 and is input to the amplitude calculation apparatus 100. And output as an amplitude calculation result of the complex signal C _in .

  Thereby, the amplitude calculation apparatus 110 according to the present embodiment has substantially the same effect as the amplitude calculation apparatus 100 according to the first embodiment. In addition, the number of phase rotation units and arithmetic processing units can be reduced as compared with the amplitude calculation device 100, and the circuit scale can be made smaller than that of the amplitude calculation device 100.

In addition, although the amplitude calculation apparatus 110 concerning this embodiment is provided with only one arithmetic processing part (arithmetic processing part 112), it is not restricted to this. For example, instead of the arithmetic processing unit 112, the arithmetic processing units 105 to 107 shown in the first embodiment are provided, the output C ₀ of the phase rotation unit 111 is input to the arithmetic processing unit 105, and C ₁ is input to the arithmetic processing unit 106. Then, C ₂ may be input to the arithmetic processing unit 107. In this case, instead of the storage unit 113, a buffer unit is provided between the arithmetic processing units 105 to 107 and the averaging unit 114, and approximate values a ₀ , a ₁ , a output from the arithmetic processing units 105 to 107 are provided. _The timing at which ₂ is input to the averaging unit 114 may be adjusted.

  Further, in the present embodiment, only one phase rotation unit 111 that can change the rotation amount is provided, but the present invention is not limited to this, and a phase rotation unit that can change the rotation amount may be further provided. Alternatively, a phase rotation unit having a constant rotation amount and a phase rotation unit having a variable rotation amount may be used in combination.

  In the amplitude calculation apparatus 110 according to the present embodiment, the number M of approximate values used for calculating the average value is not limited to 3 as in the first embodiment, and may be a natural number of 2 or more. The constants α and β of the α × MAX + βMIN method used in the arithmetic processing unit 112 are not limited to α = 15/16 and β = 15/32, and may be changed as appropriate. Further, an intensity conversion unit 115 may be provided after the averaging unit 109.

  The present invention is not limited to the above-described embodiments, and various modifications can be made within the scope of the claims, and embodiments obtained by appropriately combining technical means disclosed in different embodiments. Is also included in the technical scope of the present invention.

  The present invention can be applied to an amplitude calculation apparatus that calculates an approximate value of the amplitude of a complex signal.

It is a block diagram which shows the structure of the amplitude calculation apparatus concerning one Embodiment of this invention. It is a graph showing a calculation accuracy when you enter _{I 0 = cosθ, Q 0 =} sinθ to the amplitude computing apparatus of FIG. 2 is a graph showing a relationship between an error of an approximate value calculated by the amplitude calculation device of FIG. 1 and the number M of approximate values used for calculating an average value. It is a block diagram which shows the modification of the arithmetic processing part with which the amplitude calculation apparatus of FIG. 1 is equipped. It is a block diagram which shows the intensity | strength conversion part added to the amplitude calculation apparatus of FIG. It is a block diagram which shows the structure of the amplitude calculation apparatus concerning other embodiment of this invention. In the conventional complex signal amplitude approximation calculation circuit using the αMAX + βMIN method, it is a block diagram showing a circuit configuration when constant α = 15/16 and β = 15/32. It is a graph which shows the calculation accuracy of the conventional complex signal amplitude approximation calculation circuit shown in FIG.

Explanation of symbols

100, 110 Amplitude calculation device (complex signal amplitude calculation device)
101, 102, 103 Phase rotation unit (phase rotation means)
104 Phase calculation unit (phase calculation means)
111 Phase rotation unit (phase rotation means, phase calculation means)
105, 106, 107, 112 Arithmetic processing unit (arithmetic processing means)
108 Linear operation unit 109, 114 Averaging unit (averaging means)
113, 133, 134, 135, 136 Storage unit (storage means)
115 Strength conversion unit (strength conversion means)
128,129 Absolute value calculation unit (absolute value calculation means)
130 Comparison unit (comparison means, bit shift means)
131, 132 Integration unit (integration means, 1-bit shift means)
137, 139 Adder (addition means)
138 Integration unit (integration means, 4-bit shift means)

Claims (12)

A complex signal amplitude calculation device for calculating an approximate value of an amplitude of a complex signal C _in , comprising:
Phase calculation means for outputting _M (M is a natural number of 2 or more) complex signals C _{0 to} C _M−1 obtained by rotating the phase of the input complex signal C _in by different rotation amounts;
For each of the complex signals C _{0 to} C _M−1 , the absolute values of the real and imaginary components are compared, and the larger component is multiplied by a constant α and the smaller component is multiplied by a constant β. An arithmetic processing means for calculating an approximate amplitude value obtained by adding together
An amplitude calculating apparatus comprising: an average value calculating means for calculating an average value of the amplitude approximate values calculated for each of the complex signals C _{0 to} C _M−1 . The phase calculation means is
The input complex signal C _in is rotated in phase by a rotation amount θ _{0 to} θ _M−1 represented by θ _k = kπ / (2M) where _k is an integer from 0 to M−1. The amplitude calculation apparatus according to claim 1, wherein _M amplitude signals C _{0 to} C _M−1 are output. The phase calculation means is
M phase rotation means are provided, and the M complex signals C _{0 to} C _M-1 are calculated by rotating the phase of the input complex signal C _in by a different rotation amount by each phase rotation means. The amplitude calculation apparatus according to claim 1 or 2, characterized in that The phase calculation means is
Consists phase rotation means capable of changing the amount of rotation of the phase rotation, by sequentially changing the amount of rotation when rotating the phase of the input complex signal C _in, the M complex signal C ₀ to C _M The amplitude calculation apparatus according to claim ₁ , wherein ₋₁ is sequentially calculated and output. The same number of arithmetic processing means as the number of phase rotation means, and storage means for storing the approximate amplitude values calculated by the arithmetic processing means for each of the complex signals C _{0 to} C _M−1 .
The arithmetic means sequentially calculates the amplitude approximation values of the complex signals C _{0 to} C _M−1 sequentially input from the phase rotation means and sequentially stores them in the storage means. 4. The amplitude calculation device according to 4. Intensity conversion means for multiplying the average value calculated by the average value calculation means by a constant γ, and a value obtained by multiplying the average value by a constant γ is calculated as the amplitude of the input complex signal C _in. Item 6. The amplitude calculation apparatus according to any one of Items 1 to 5.   The amplitude calculation apparatus according to claim 1, wherein the value of the constant α is 15/16, and the value of the constant β is 15/32.   The amplitude calculation apparatus according to claim 1, wherein the constants α and β have a relationship of α = 2 × β.   9. The amplitude calculation apparatus according to claim 8, wherein the value of the constant [alpha] is 1.   9. The amplitude calculation apparatus according to claim 8, wherein the value of the constant [alpha] is 2.   The amplitude calculation apparatus according to claim 1, wherein the value of M is three. A method for calculating an amplitude of a complex signal for calculating an approximate value of the amplitude of the complex signal C _in , comprising:
A phase calculation step of outputting _M (M is a natural number of 2 or more) complex signals C _{0 to} C _M−1 obtained by rotating the phase of the input complex signal C _in by different rotation amounts;
For each of the complex signals C _{0 to} C _M−1 , the absolute values of the real and imaginary components are compared, and the larger component is multiplied by a constant α and the smaller component is multiplied by a constant β. An arithmetic processing step for calculating an approximate amplitude value obtained by adding the sum,
And an average value calculating step of calculating an average value of the approximate amplitude values calculated for each of the complex signals C _{0 to} C _M−1 .

Images