JP2023104044A - Antenna selection apparatus - Google Patents

Abstract

To improve resolution for a low-resolution A/D converter and allow each of antennas to be weighted.SOLUTION: An antenna selection apparatus includes two receiving antennas 10A, 10B, a switch 11, a noise source 12, an adder 13, an A/D converter 14, an estimation apparatus 15, and a control apparatus 16. The A/D converter 14 is a device configured to quantize noise addition signals from the adder 13 and convert the signals into digital signals to be output. The A/D converter 14 is a low-resolution converter having a resolution of 1-9 bits. The estimation apparatus 15 acquires multiple samples from the digital signals, estimates a distribution of the noise addition signals from a distribution of voltage values of the samples, and estimates a voltage value of a receiving signal from the distribution. The control apparatus 16 is configured to control selection of the receiving antennas 10A, 10B in the switch 11 on the basis of the voltage value of the receiving signal estimated by the estimation apparatus.SELECTED DRAWING: Figure 1

Description

The present invention relates to an antenna selection device that weights a plurality of antennas so as to optimize communication.

In recent years, efforts have been made to positively utilize noise, and one representative example is nonlinear signal processing based on stochastic resonance phenomena. By using the stochastic resonance phenomenon, even weak signals buried in noise may be detected.

On the other hand, a technique called adaptive array antenna is known (Patent Documents 1 to 3). This is a technique for improving reception sensitivity by arranging a plurality of reception antennas and controlling the weighting of each reception antenna, and is an important technique for realizing multifunctional devices. The weighting here also includes selectively receiving any one of the receiving antennas. In particular, antenna selection technology, which selects the optimum antenna in a fading environment and receives signals, is widely used because it is expected to significantly improve reception sensitivity in spite of being easy to control.

For nanosensors and the like, it is assumed that communication modules with low performance are used for cost reduction. For example, it is conceivable to use a low resolution A/D converter of several bits. Further, since next-generation wireless communication terminals have high operating frequencies, it is assumed that low-cost, low-resolution A/D converters will be used. In the case of using a communication module with such low performance, attempts to improve the performance using communication technology are being considered.

Japanese Patent No. 5377711 Japanese Patent No. 4157530 Japanese Patent No. 4073148

However, if a low-resolution A/D converter is used in the adaptive array antenna, there is a problem that the optimum antenna may not be selected. For example, in an A/D converter that outputs a signal of 3.75V for an input signal of 2.5V to 5V, when a certain antenna A is 3V and a certain antenna B is 4.5V, the output is 3.5V. Since it becomes 75V, the antenna cannot be selected. However, in practice, antenna B should be selected because it can obtain more energy from the space and is more resistant to environmental noise.

SUMMARY OF THE INVENTION Accordingly, it is an object of the present invention to enable weighting of each receiving antenna even when a low-resolution A/D converter is used.

The present invention comprises a switch that sequentially selects and outputs one of signals received from each receiving antenna, and an addition that adds noise with a known distribution to the received signal from the switch and outputs a noise addition signal. a 1- to 9-bit A/D converter for quantizing the noise addition signal from the adder and outputting a digital signal; and measuring the digital signal from each A/D converter multiple times. an estimation device for estimating the distribution of the noise-added signal from the measured distribution of the digital signal and the distribution of the noise, and estimating the received signal from the estimated distribution of the noise-added signal; and a control device for controlling the weighting of each of the receiving antennas based on the estimated received signal of each of the receiving antennas.

In the present invention, the noise may be normally distributed. At this time, the estimation apparatus may obtain a normal distribution that minimizes the error from the distribution of the measured digital signal by the method of least squares, and estimate the normal distribution at that time as the distribution of the noise addition signal.

According to the present invention, each receiving antenna can be weighted even when a low-resolution A/D converter is used.

The figure which showed the structure of the antenna selection apparatus of embodiment. 4 is a diagram showing input/output characteristics of an A/D converter 14; FIG. The figure which showed the distribution of a digital signal. FIG. 4 is a diagram showing an estimated distribution of noise-added signals and an actual distribution of noise-added signals; FIG. 4 is a diagram showing the relationship between the voltage value of the signal received by the receiving antenna 10B and the rate at which the optimum antenna is selected;

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a diagram showing the configuration of an antenna selection device according to an embodiment. As shown in FIG. 1, the antenna selection device of the embodiment includes two receiving antennas 10A and 10B, a switch 11, a noise source 12, an adder 13, an A/D converter 14, an estimating device 15, and a controller 16 . The antenna selection device of the embodiment is a device that controls antenna selection in an adaptive array antenna, and selects the optimum antenna from the two reception antennas 10A and 10B to improve reception sensitivity.

The receiving antennas 10A and 10B are devices that receive electromagnetic waves and output received signals (voltage signals). The receiving antennas 10A and 10B are arranged at predetermined intervals.

The switch 11 is a device that selects and connects one of the receiving antennas 10A and 10B and outputs a received signal from the selected receiving antenna. Selection of the receiving antennas 10A and 10B by the switch 11 is controlled by the control device 16. FIG. A received signal output from the switch 11 is input to the adder 13 .

The noise source 12 is a device that generates and outputs noise. The noise is Gaussian noise with a known normal distribution. A known normal distribution is a normal distribution with known mean and variance. Noise generated by noise source 12 is input to adder 13 . Note that if the distribution is known, it does not have to be a normal distribution. Also, the noise may be obtained from the surrounding environment.

The adder 13 is a device that adds noise from the noise source 12 to the signal received from the switch 11 and outputs the result. A noise addition signal output from the adder 13 is input to the A/D converter 14 . Note that the input of noise from the noise source 12 is controlled to be off except during the antenna selection operation (that is, during normal communication).

The A/D converter 14 is a device that quantizes the noise addition signal from the adder 13, converts it into a digital signal (a signal of discrete voltage values), and outputs the digital signal. The A/D converter 14 uses one with a low resolution of 1 to 9 bits. A portion of the digital signal from A/D converter 14 is input to estimator 15 .

Conventionally, there were cases where it was difficult to select an antenna with a resolution of 1 to 9 bits, but in this embodiment antenna selection is possible even in such a case. The sampling rate of A/D converter 14 is arbitrary.

The estimator 15 acquires a plurality of samples from the digital signal, estimates the distribution of the noise addition signal from the distribution of the voltage values, and estimates the voltage value of the received signal from the distribution. The estimation method will be described later.

The control device 16 is a device that controls selection of the receiving antennas 10A and 10B in the switch 11 based on the voltage value of the received signal estimated by the estimating device. Specifically, of the receiving antennas 10A and 10B, the one with the larger estimated voltage value of the received signal is selected.

Next, the operation of the antenna selection device of the embodiment will be described. It should be noted that the antenna selection operation should be performed during the preamble period of communication. Antenna selection can be performed without effectively interfering with communication. Of course, a period for antenna selection may be separately provided for communication.

First, the switch 11 is controlled by the control device 16 to connect to the receiving antenna 10A and receive the receiving signal r01 from the receiving antenna 10A. Noise n of normal distribution (mean 0, variance σ ² ) from noise source 12 is added to received signal r01 to obtain r01+n, which is then converted into digital signal r1 by A/D converter 14 . The variance of the noise n suitable for estimating the received signal r01 depends on the received signal r01, but is for example 0.2-3.

Next, the estimating device 15 samples the voltage value of the digital signal r1 from the A/D converter 14 and obtains the distribution f(r1) of the voltage value of the digital signal r1. The number of samples may be such that a distribution is formed, and is, for example, 100 or more. The distribution f(r1) is a probability density distribution, and is a histogram-like distribution with a constant probability density for each range of voltage values to be quantized. For example, the dynamic range of the A/D converter 14 is 0 to 10 V, the resolution is 2 bits, and the input 0 to 2.5 V, 2.5 to 5 V, 5 to 7.5 V, and 7.5 to 10 V has four values. In the case of output input/output characteristics (see Fig. 2), a histogram with a constant probability density is obtained in each range of 0 to 2.5 V, 2.5 to 5 V, 5 to 7.5 V, and 7.5 to 10 V (Fig. 3).

Here, since the input r01+n to the A/D converter 14 fluctuates due to normal distribution noise n, the distribution f(r1) of the digital signal r1 from the A/D converter 14 is also the normal distribution probability density distribution g( r1). Let g(r1) be mean a and variance ^b2 .

Therefore, the estimation device 15 determines the mean a and the variance ^b2 of g(r1) so that the distribution f(r1) best fits the normal distribution g(r1). Then, let a and ^b2 at this time be the mean estimated value ^r01 and the variance estimated value ^ ^σ2 for the distribution of r01+n.

̂r01 and ^̂σ2 are calculated by the method of least squares. That is, as in the following equation (1), the sum of the squares of the difference between f(r1) and g(r1) is calculated for each quantized voltage section, and the average a, Determine the variance ^b2 .

In the above equation (1), i represents the i-th smallest quantized voltage interval, and r[i] represents the i-th voltage interval. Also, M is the number of bits of the A/D converter 14, and the number of quantized voltage sections is ^2M .

In addition to the least squares method, a maximum likelihood estimation method, machine learning, or the like may be used to estimate ^r01 and ^ ^σ2 .

Since the noise n has a normal distribution with mean 0 and variance ^σ2 , the estimated value of the received signal r01 before the noise n is added is r01. As described above, the estimation device 15 estimates the value of the received signal r01 received by the receiving antenna 10A as ^r01.

Next, the controller 16 controls the switch 11 to connect to the receiving antenna 10B and receive the receiving signal r02 from the receiving antenna 10B. Then, similarly to the case of the received signal r01, the estimating device 15 estimates the voltage value of the received signal r02 as ^r02.

Next, the control device 16 compares the estimated value ^r01 of the received signal r01 estimated by the estimating device 15 with the estimated value ^r02 of the received signal r02, and determines which of the receiving antennas 10A and 10B has the larger voltage value. Switch connection. In this manner, the antenna selection device of the embodiment selects the optimum one of the receiving antennas 10A and 10B.

As described above, according to the antenna selection device of the embodiment, even when the low-resolution A/D converter 14 is used, it is possible to select the optimum one of the receiving antennas 10A and 10B for improving the receiving sensitivity. . In the low-resolution A/D converter 14, even if the received signal r01 and the received signal r02 from the receiving antenna 10A have different voltage values, they are converted into the same voltage value by quantization in the A/D converter 14. more likely to get lost. In that case, conventionally, the antenna could not be selected. However, in the antenna selection device of the embodiment, by adding noise, it is possible to accurately estimate the voltage values of the received signals r01 and r02 from the digital signal. Therefore, even if the voltage values of the digital signals from the A/D converter 14 are the same for the received signal r01 and the received signal r02, the most suitable one of the receiving antennas 10A and 10B is selected based on the estimated values of the received signals r01 and r02. You can choose which one.

Next, simulation results regarding the antenna selection device of the embodiment will be described.

(Experiment 1)
The following simulation was performed for the estimation of the received signal in the embodiment, and its effect was confirmed. A received signal was estimated from a digital signal by the method of the first embodiment using a 2-bit A/D converter 14 with an input voltage dynamic range of 0 to 10V. The input/output characteristics of the A/D converter 14 at this time are as shown in FIG. The actual received signal was set to 4 V, and the number of samples for obtaining the distribution of the digital signal was set to 500. The noise has a normal distribution with an average of 0 and a variance of 1.

FIG. 3 is a diagram showing the distribution of the obtained digital signals. The horizontal axis is the voltage value (V) of the digital signal of the A/D converter 14, and the vertical axis is the probability density. FIG. 4 is a diagram showing the distribution of the estimated noise addition signal (the signal input to the A/D converter 14) and the distribution of the actual noise addition signal. The horizontal axis is the voltage value (V), and the vertical axis is the probability density. The distribution of the estimated noise sum signal was assumed to be a normal distribution with a mean of 3.96 V and a variance of 1.02. Since the noise distribution is a normal distribution with an average of 0V and a variance of 1, the received signal was estimated to be 3.96V. As a result, it was found that the estimation can be performed accurately with an error of 1%.

(Experiment 2)
Regarding the selection of the receiving antennas 10A and 10B in the embodiment, the following simulation was performed to confirm the effect. The characteristics of the A/D converter 14 were assumed to be the same as in Experiment 1, and the output was saturated at 10 V when the input was 10 V or higher. The number of samples for obtaining the distribution of digital signals was 500 for each. The signal received from the receiving antenna 10A was set to 10V, and the signal received from the receiving antenna 10B was changed from 8.2V to 11.8V at intervals of 0.4V. The noise has a normal distribution with an average of 0 and a variance of 2.25. As shown in FIG. 3, when no noise is added, the signals received from the receiving antennas 10A and 10B are both converted to 8.75V signals by the A/D converter 14. FIG. Therefore, conventionally, this is the case where the antenna cannot be selected.

FIG. 5 is a diagram showing the relationship between the voltage value of the signal received by the receiving antenna 10B and the rate at which the optimum antenna is selected. The ratio of selecting the optimum antenna was calculated after 10,000 trials. It was determined that the optimum antenna was selected when the signal with the larger voltage value was selected from the signals received by the receiving antennas 10A and 10B. That is, when the voltage value of the signal received by the reception antenna 10B is less than 10 V, the reception antenna 10A is the optimum antenna, and when the voltage value of the signal received by the reception antenna 10B is 10 V or more, the reception antenna 10B is the optimum antenna. be.

As shown in FIG. 5, when the difference in voltage between the signal received from the receiving antenna 10A and the signal received from the receiving antenna 10B is large (difference of 0.6 V or more), the optimal antenna has a ratio of almost 100%. Even when the voltage value difference was small (0.2 V difference), the optimum antenna could be selected at a rate of 94%.

(Modification)
In the embodiment, one of the signals received from the receiving antennas 10A and 10B is selected and received, but it is also possible to weight the signals received from the receiving antennas 10A and 10B. The weighting of the receive antennas 10A, 10B includes selective reception of the receive antennas 10A, 10B. In weighting, if one of the receiving antennas 10A and 10B is given a weight of 1 and the other is given a weight of 0, this is the same as selective reception by the receiving antennas 10A and 10B.

Although one type of noise is used in the embodiment, it is possible to improve the estimation accuracy by estimating the received signal by changing the type of noise. For example, the voltage value of the received signal may be estimated using noise of normal distribution, the voltage value of the received signal may be estimated using noise of Poisson distribution, and both may be averaged.

Although the embodiment selects one out of two receive antennas, it can be easily extended to three or more receive antennas.

INDUSTRIAL APPLICABILITY The present invention can be used for antenna selection in adaptive array antennas.

10A, 10B: Receiving Antenna 11: Switch 12: Noise Source 13: Adder 14: A/D Converter 15: Estimation Device 16: Control Device

Claims (3)

a switch for sequentially selecting and outputting one of the signals received from each receiving antenna;
an adder that adds noise with a known distribution to the received signal from the switch and outputs a noise addition signal;
a 1- to 9-bit A/D converter that quantizes the noise addition signal from the adder and outputs a digital signal;
measuring the digital signal from each A/D converter a plurality of times, estimating the distribution of the noise sum signal from the measured distribution of the digital signal and the noise distribution, and estimating the estimated noise sum signal; an estimating device for estimating the received signal from the distribution of
a control device for controlling the weighting of each of the receiving antennas based on each received signal of each of the receiving antennas estimated by the estimating device;
An antenna selection device comprising: 2. The antenna selection device according to claim 1, wherein the noise has a normal distribution. The estimating device obtains a normal distribution that minimizes an error from the measured distribution of the digital signal by the method of least squares, and estimates the normal distribution at that time as the distribution of the noise addition signal. 3. The antenna selection device according to claim 2.