JP2010181385A - Apparatus and method for sensing incoming wave - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reduce a calculation processing amount in a multi-static radar receiver. <P>SOLUTION: An apparatus for sensing an incoming wave includes: a RF processing unit 3 for receiving a plurality of electric waves from a plurality of channels with different frequency bands; frequency conversion units 5-1, 5-2, 5-3 for frequency-converting the plurality of electric waves into the same low frequency bands, respectively; an addition unit 6 for adding the plurality of frequency-converted electric waves to obtain addition signals; and an estimation unit 7 for sensing incoming waves using characteristic value analysis of correlation matrixes of the addition signals. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an incoming wave detection device and an incoming wave detection method for detecting incoming waves from a plurality of transmission sources.

  Bistatic radar or multistatic radar has been proposed as a method for detecting a target having a shape in which a transmission wave from a radar transmitter is unlikely to return to a radar receiver installed at the same location as the radar transmitter. In the multistatic radar, a configuration in which there is one receiving station for a plurality of transmitting stations has been proposed (for example, Patent Document 1 and Patent Document 2). When one receiving station is configured to detect radio waves from a plurality of transmitting stations, the target is more easily detected as the number of transmitting stations increases. In the conventional configuration, even if there is one receiving station for a plurality of transmitting stations, the received wave from each transmitting station is separated in the receiving station and then the reflected wave is detected.

  On the other hand, various algorithms such as direction-of-arrival estimation and delayed wave separation have been proposed as methods for analyzing radio waves arriving at a radar receiving station (for example, Non-Patent Document 1). In Non-Patent Document 1, eigenvalue analysis is performed on a correlation matrix of a received signal to detect an incoming wave.

  In the future, in a configuration in which radar waves from a plurality of transmitting stations are received by a single receiving station, a configuration in which arrival waves are detected using eigenvalue analysis of a correlation matrix for each radar wave is assumed.

JP-A-11-125684 JP 05-34447 A

Nobuyoshi Kikuma "Adaptive signal processing by array antenna" Science and Technology Publishing, April 2004 Chapters 10-13

  In a configuration in which radar waves from a plurality of transmitting stations are received by a single receiving station, when the number of transmitting stations is increased in order to make it easier to detect the target, the receiving station detects the number of incoming waves as many as the number of transmitting stations. It is necessary to have a system. Since eigenvalue analysis is a calculation with a large amount of processing, the amount of processing or the circuit scale increases at the receiving station.

  The present invention has been made in view of the above, and provides an incoming wave detection device and an incoming wave detection method that can reduce the amount of calculation processing in a radar receiver.

  In order to achieve the above object, an incoming wave detection device according to an embodiment of the present invention includes an RF processing unit that receives a plurality of radio waves from a plurality of channels having different frequency bands and the plurality of radio waves that are in the same low frequency band. A frequency conversion unit that converts the frequency into a frequency, an addition unit that adds a plurality of frequency-converted radio waves to obtain an addition signal, and an estimation unit that performs arrival wave detection using eigenvalue analysis of a correlation matrix of the addition signal It is characterized by that.

  The amount of calculation processing in the radar receiver can be reduced.

The block diagram which shows the structure of the incoming wave detection apparatus which concerns on 1st Embodiment. The figure which shows operation | movement of the incoming wave detection apparatus which concerns on 1st Embodiment. The figure which shows the incoming wave detection apparatus which concerns on 1st Embodiment. The figure which shows the incoming wave detection apparatus which concerns on the modification 1 of 1st Embodiment. The figure which shows operation | movement of the incoming wave detection apparatus which concerns on the modification 1 of 1st Embodiment. The block diagram which shows the structure of the incoming wave detection apparatus which concerns on the modification 1 of 1st Embodiment.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following drawings, only parts that are essentially related to the present invention are shown, and parts that are necessary in actual implementation but are not directly related to the operation of the present application are not shown.

(First embodiment)
FIG. 1 is a block diagram showing a configuration of an incoming wave detection device 1 according to the first embodiment. The incoming wave detection device 1 includes an antenna 2 that receives radio waves of a plurality of channels f1, f2,... Having different frequencies, and an RF unit that amplifies, shapes, and converts the received radio waves when the superheterodyne method is adopted. 3, filters 4-1, 4-2,... For separating a plurality of channels, respectively, and frequency converters 5-1, 5-5 for converting the received waves of the respective channels that have passed through the filter 4 into baseband frequencies. 2..., And an adder 6 that adds the waves of each channel converted into the baseband, and an estimation unit 7 that detects an incoming wave for the added signal added by the adder 6.

  FIG. 2 is a diagram illustrating an operation of the incoming wave detection device 1 according to the first embodiment. The case of 3 channels is described. The received signals in the frequency bands corresponding to the respective channels f1, f2, and f3 are converted into basebands in the form of complex signals. If the center frequency is a so-called baseband in the vicinity of 0 Hz, it is converted into an in-phase signal (I) and a quadrature signal (Q) separately. In this case, the subsequent processing is performed for each of I and Q until the digital part can handle the complex number. Alternatively, in the state of a real signal, it may be converted to a low intermediate frequency (Low IF) where the center frequency is not zero. In any case, the three channels are converted into substantially the same baseband signal and added by the adder 6. Note that “identical” does not necessarily have to be completely identical. As will be described later, a slight error may be intentionally given.

  The estimation unit 7 detects an incoming wave by a super-resolution algorithm based on the eigenvalue analysis of the correlation matrix. Specifically, the detection of the incoming wave includes wave number detection, arrival direction estimation, delay distribution estimation, Doppler frequency estimation, and the like for a plurality of incoming waves incident on the antenna. The detection of an incoming wave does not perform all of these, but in most cases, wave number detection and arrival direction estimation are performed, and other estimations are performed as necessary.

  Super-resolution algorithms based on eigenvalue analysis of correlation matrix are, for example, MUSIC (Multiple Signal Classification) method (Non-patent Document 1), ESPRIT (Estimation of Signal Parameters via Rotational Innovation Techniques Non-Patent Document 1) MODE (Method Of Direction Estimation) method (P. Stoica et al., “Novel evolutionary method for direction estimation”, IEEE Proceedings, Vol. 137, No. 19, Ft. Non-Patent Document 1) etc. .

  The estimation unit 7 estimates the number of incoming waves included in the added signal based on the implemented algorithm, and in what direction, further delay, or each wave included in the added signal Estimate if it arrives at the frequency and output the result.

  As described above, the incoming wave detection device 1 according to the first embodiment can estimate a plurality of channels with one estimation unit. Compared with the method of performing eigenvalue analysis in the state of a wideband signal with f1, f2, and f3 maintaining the frequency interval, the bandwidth of the signal to be estimated can be reduced. The received wave is converted into a digital signal at any stage before the estimation unit 7, but the sampling rate in the estimation unit 7 can be reduced by adding as in the present embodiment. It is possible to reduce the estimation processing amount of incoming waves that arrive over a plurality of frequency channels. When reflected waves from multiple channels arrive from the same direction, keep paying attention to the direction when evaluating the cost function. If reflected waves from one of the multiple waves arrive, reflected waves from other waves The goal can be tracked continuously even if the power is lost.

  In the following, using FIG. 3, the direction of arrival estimation using the MUSIC method is taken as an example to explain that such a method is possible. When the arrival direction is estimated, a plurality of antenna elements 11-1 to 11-K are used as shown in FIG. Each of the plurality of antenna elements 11-1 to 11 -K is connected to the RF unit 3, and each channel is filtered and added, and then the signals of all the antenna elements are input to the same estimation unit 7. The

FIG. 3 shows K linear arrays arranged at intervals d. A signal obtained by receiving each antenna element k (= 1, 2,..., K) at time t and adding a plurality of channels for each antenna is _represented by x _k (t). This vectorized in k direction

Put it. T represents transposition. If there are a total of L incoming waves in all the added signals,

It can be expressed as. However,

It is. F ₁ (t) is the complex amplitude of the l-th wave, θ _l is the arrival direction of the l-th wave, and n _k (t) is noise added to each antenna at the time of reception. a (θ _l ) is called a steering vector.

Λ is the RF wavelength of the incoming wave. There are several ways to handle frequency multiplexed signals. By defining Λ for the center frequency of each channel and making it multidimensional so that the change of the steering vector becomes independent with respect to each change of sin (θ _l ) and Λ, accurate estimation becomes possible. . The method will be briefly described later. If the ratio band in the RF of the entire band including a plurality of channels to be added is small, the accuracy is slightly degraded, but the approximation can be made with the center frequency of the entire band. For example, in a channel with a center frequency of 10 GHz and a channel with a center frequency of 10.02 GHz, if d is 3 cm and θ _l is 30 degrees, the difference of (2π / Λ) · d · sin θ _l is about 0.01 radians. Even if Λ is defined with a value corresponding to 10 GHz, a large error does not occur.

Equation 2 is transformed into a correlation matrix.

E [] is an expected value in the time direction, and is substantially a time average in a certain period. H indicates complex conjugate transposition. I is a unit matrix, and σ ² is noise power when noise power added to all antennas is equal, and is measured in advance. S = E [F (t) F ^H (t)], which is a correlation matrix between incoming waves. When incoming waves are independent of each other, a diagonal matrix is obtained. Here, it is assumed that it is independent, and the case where it is not independent will be briefly described later. R _xx is the correlation matrix of the antenna reception signal, but as shown in Equation 4, when calculating the correlation matrix, the complex conjugate is taken, so the component corresponding to the center frequency is canceled, and no matter how many No longer. The same applies to S. When Λ of channels having different frequencies are approximated with the same frequency, a frequency offset component remains in F (t). However, by multiplying the complex conjugate, the frequency offset is canceled for the diagonal component. (If L waves are independent, no component other than the diagonal remains.) At this stage, even if a plurality of channels having different frequencies are added, there are L waves in the same channel (SNR). It can be seen that there is no significant difference (except for).

  The only problem in estimating the direction of arrival is the difference in the phase difference between the arriving waves generated between the antenna elements. The difference in phase difference is expressed by a steering vector, and the direction of arrival is estimated by estimating the steering vector for each incoming wave. As described above, the steering vector treats the frequency difference of each channel as an error, or incorporates the frequency difference by making it multidimensional, so the steering vector of multiple waves contained in channels with different frequencies is roughly correct. Can be estimated.

When eigenvalue analysis is performed on R _xx , eigenvectors corresponding to K eigenvalues λ _k are obtained. When the eigenvalues are arranged in descending order, the last KL pieces are (approximately) equal to the noise power σ ² and become noise eigenvalues. The wave number of the signal is estimated by subtracting the number of eigenvalues equal to the noise power from the number of antenna elements. For the actual wave number determination, a determination method such as AIC (Akaike Information Criteria) or MDL (Minimum Descripton Length) is used. The wave number determination method conforms to the conventional method, and detailed description thereof is omitted. If KL eigenvectors corresponding to eigenvalues of KL noises are e _i (i = L + 1,..., K), the space formed by e _i is the noise subspace. The remaining eigenvalues and eigenvectors correspond to the signal, and the eigenvector of the signal is orthogonal to the noise subspace. The space due to the eigenvector of the signal is equal to the space spanned by the steering vector corresponding to the direction of arrival of the signal. Therefore, the reciprocal P _MU (θ) is calculated by normalizing the steering vector and the Euclidean distance between the noise subspaces by the magnitude of the steering vector.

This is called a MUSIC spectrum and shows a peak at θ that coincides with the incident angle of the incoming wave. In Equation 5, the angle of arrival θ is estimated by scanning θ to find a peak or calculating the root of the denominator. The latter method is called root-MUSIC. P _MU (θ) and the denominator of the _{P MU (θ)} used in the root-MUSIC is called a cost function.

As described above, the phase difference between the antenna elements of the incoming wave is expressed in the form of a steering vector. This is a calculation of R _xx and S when obtaining the noise subspace, and the center frequency is not relevant. From the nature of eigenvalue analysis, there is no problem even if a plurality of channels are included in X (t) as long as the phase difference between the antenna elements of the incoming wave can be maintained. For these reasons, it is possible to estimate each incoming wave even if signals of a plurality of channels having different frequencies are added at the same baseband frequency and then processed.

  However, unlike the case where each channel is processed individually, the signal-to-noise ratio (SNR) may deteriorate, and an error may occur in estimation. Furthermore, most of the arrival wave estimation methods can estimate only arrival waves up to the number of vector elements −1 of X (t). Therefore, it is difficult to add a large number of channels, and it is desirable to select a channel to be added depending on the situation used. Channels are selected and added so that at least the total number of incoming waves is equal to or less than the number of elements of the vector of X (t) −1. This will be described later in detail with reference to FIGS.

  Although the above is an example of MUSIC, as can be understood from the above description, this method is a method for estimating an incoming wave by obtaining a steering vector using eigenvalue analysis of a correlation matrix (for example, ESPRIT, maximum likelihood). If it is an estimation method, etc., it is applicable.

  In the above example, all the incoming waves are independent. However, if the same signal arrives from different directions due to multipath, the spatial averaging method (Non-patent Documents 1 and 12) is applied. Good. In the spatial averaging method, the number of antennas is increased, the correlation matrix is averaged between antennas, and components other than the diagonal of S are brought close to zero.

  In the spatial averaging method, a plurality of subarrays composed of a plurality of antennas are defined in all antennas. It is sufficient that the number of antenna elements included in the subarray is K, K is L + 1 or more, and the number of subarrays is at least L / 2 or more. If a plurality of subarrays are defined as an array in which antenna elements are shifted one by one, the total number of antenna elements is required to be (3/2) L at a minimum with respect to the number of incoming waves L. Detailed methods are not described here because they do not fall within the scope of conventional methods. Such a spatial averaging method and this method can be used together without difficulty.

  Signals having different frequencies are inherently incoherent. However, when frequency conversion is performed to completely the same low frequency, if the modulation components of the signal are the same, for example, a complete sine wave, the signals are regarded as coherent. Again, incoherent can be achieved by spatial averaging. If it is known in advance that coherent waves are not included in the same channel and you do not want to use spatial averaging, calculate the expected frequency at least slightly different from the same low frequency. It is preferable to down-convert to a frequency shifted so that the period to be longer than the reciprocal of the frequency difference.

  When the delay distribution and Doppler frequency estimation are further performed using this method, or when the center frequency of each channel is individually defined in order to increase the accuracy of the arrival direction, a large number of steering vectors are used. There is a need to dimension.

  In the steering vector defined by Equation 3, the arrival angle, delay, and frequency are all on the shoulder of e in the same dimension. In this state, for example, there is no difference between how the steering vector changes when the arrival angle changes slightly and how the steering vector changes when the delay changes slightly, and it is not possible to distinguish which has changed. Therefore, by changing the parameters to be taken in the vector direction of X (t) and the shape of the steering vector, when the plural parameters such as the arrival angle, the delay, and the frequency change, how the steering vector changes Many different methods have been proposed. For example, there are a method disclosed in Japanese Patent Application Laid-Open No. 2006-208172, a method disclosed in IEICE Technical Report SANE 2007-109, and the like. The details of the multi-dimensional method are the same as the conventional method and will not be described here. However, by devising the shape of the steering vector in this way, parameters other than the arrival angle can be measured simultaneously, and can be used in combination with this method.

  When multi-dimensionalization with frequency is performed, when scanning the parameters when obtaining the steering vector corresponding to the incoming wave in the final stage, the frequency is limited to the center frequency of each channel or its surroundings. Thus, it is desirable to reduce processing.

  FIG. 1 does not show the position of the A / D converter. Since the processing performed by the estimation unit 7 is digital signal processing and the antenna is an analog component, a portion for converting an analog signal into a digital signal is required between the antenna 2 and the estimation unit 7. The estimation unit 7 operates with a bandwidth (sampling rate) sufficient to process one channel. Accordingly, processing (decimation) for reducing the bandwidth is performed somewhere up to that point.

  There are roughly two types of A / D converter insertion positions and decimation positions.

  The first is a method in which an A / D converter is inserted immediately before or after the adder 6 and the sampling rate of the A / D converter is set to the sampling rate after decimation. According to this method, a decimation operation is substantially unnecessary and a low-speed A / D converter can be used, and the linearity and the number of bits of the A / D converter can be increased correspondingly.

  The other is that all channels are converted into digital signals with a high-speed A / D converter at once in RF, Low IF or baseband while maintaining the frequency spacing at RF, and digital filters and Convert to the same frequency with a frequency converter. Alternatively, the A / D converter output signal is duplicated by the number of additions, added with frequency offsets corresponding to the frequency intervals of the channels, and subjected to filtering and decimation. In this method, a filter or a frequency converter also serves as a decimation unit. In this configuration, the degree of freedom in selecting and adding channels is higher than in the method of tuning the analog filter. A configuration for digitizing the RF unit is being realized. In such a case, this method is desirable.

(Modification 1 of the first embodiment)
FIG. 6 is a block diagram illustrating a configuration of the incoming wave detection device 101 according to the first modification of the first embodiment. 4 and 5 are diagrams for explaining the premise and operation of the embodiment of FIG. A form corresponding to a plurality of antenna elements is not shown because the figure is complicated, and only one antenna is shown. There are actually a plurality of antenna elements, and as many RF units, filters, frequency converters, and adders as the number corresponding to the antenna elements are required.

  First, the premise will be described with reference to FIG. An arrival wave estimation apparatus 101 according to the first modification of the first embodiment is mounted on the radar receiver 10. The radar receiver 10 is a receiver of a multistatic radar system, and there are nine corresponding radar transmitters, radar transmitters 8-1 to 8-9 in the case of FIG. The radar receiver 10 is intended to detect the target 9, but all of the radar transmitters 8-1 to 8-9 have not succeeded in irradiating the target 9 with radar waves. Further, not all radar reflected waves from the target 9 of the radar transmitter irradiating the target 9 with radar waves necessarily reach the radar receiver 10.

  The radar transmitter 8-x (x = 1 to 9) determines whether or not the radar wave of its own transmitter is irradiated on the target 9, by determining whether the radar receiver 10 or another radar receiver (not shown) or radar transmission is used. It is detected by the reflected wave received by the radar receiver installed in the aircraft itself. However, since this part is not related to the operation of this method, the description is omitted. In FIG. 4, among the nine radar transmitters 8 -x, the radar waves of 8-1, 8-2, 8-3, 8-5 and 8-6 are irradiated on the target 9.

  The transmission frequencies of these radar transmitters 8-x are different from each other, and are f1 to f9 in order. FIG. 5 is a diagram schematically showing the detection status of these frequencies and the reflected wave at the radar receiver 10. The horizontal axis represents the RF frequency, and each mountain represents the channel of each radar wave. The darkness of the fill corresponds to the detection state of the reflected wave at the radar receiver 10. The reflected waves of the radar waves 8-2 and 8-5 (f2 and f5) are detected relatively stably, and 8-1, 8-3 and 8-6 (f1, f3, f6) are unstable. And others are rarely detected.

  In FIG. 4, the trajectory of the radar wave is drawn with a line, but actually it is reflected in all directions by the target. When the transmission radar wave is applied to the target 9, the radar reflection cross section for each incident direction and reflection direction is defined by the material and shape of the target. The reason why only two of the five waves irradiated on the target 9 can be detected stably depends on the position of the radar transmitter / radar receiver relative to the target, the power of the radar transmission wave, and the like.

  In order to cope with such a situation, in the form of FIG. 6, the radar wave that is detected in an unstable manner or the channel of the radar wave that is rarely detected is added with an appropriate number of channels and stably detected. The channels to be added are configured to be added by individual systems. In the form of FIG. 6, there are two systems that individually estimate for each channel, each having estimation units 7-1 and 7-2, and the other channels are added by adders 6-1, 6-2,. It is estimated by the estimation units 7-3, 7-4,. The process after the addition is the same as in FIG. In the example of FIG. 5, the channels f2 and f5 are processed by a system that performs individual processing, and the other channels are processed by a system that performs addition processing. For example, the channels of f1 and f3 are added by the adder 6-1 and estimated by the estimation unit 7-3, and the channels of f4, f6, and f7 are added by the adder 6-2, and the estimation unit 7- 4 and so on, and so on.

  The estimation results obtained by the estimation units 7-1,... Are sent to the integration processing unit 13 and the detection status database 14 that integrate the results.

  The integration processing unit 13 integrates the detection amounts of desired parameters (such as the arrival direction and the delay amount of the reflected wave) of the target 9 for the estimation results of the respective estimation units, determines the position and direction of the target 9, Output. The method of integration by the integration processing unit 13 is not the purpose of this method and will not be described in detail, but appropriate processing such as weighting processing is performed with priority given to the estimation result of the channel with higher reliability of the reflected wave.

  The detection status database 14 collects and organizes the detection status at each estimation unit. The detection status for each channel (the arrival probability of the reflected wave and the SNR of the reflected wave) is arranged and ranked in the long and short term. It is easy to collect such information for channels that are individually processed. The estimation unit also outputs the power of each incoming wave. Normally, since direct waves and clutter from the radar transmitter are received in some form, the power and detected SNR of each incoming wave such as reflected wave, direct wave, and clutter are calculated and sent simultaneously to the detection status database 14. . The power of each wave when a plurality of waves arrives can be calculated by an equation as shown in Chapter 13 (13.35) of Non-Patent Document 1, for example. The discrimination between the clutter and the target reflected wave may be made based on the arrival angle, the spread of the Doppler frequency being large, or the Doppler frequency itself, as in the prior art, and the approximate position and direction of the target 9 are given in advance. In some cases, it may be determined depending on whether or not it is suitable.

  In the system estimated after the addition, depending on the eigenvalue analysis and the calculation of the power from the eigenvalue analysis result as described above, even if a reflected wave is detected, it may not be known which channel is added. If the channel is semi-fixedly assigned to the grid, it is not necessary to know which channel is detected. However, when switching which channel is assigned to each channel depending on the detection status, it is necessary to know which channel is detected.

  When the frequency is detected in addition to the arrival angle in the multi-dimensionalization as described above, it is possible to easily know which channel is detected from the frequency. If multi-dimensionalization is not performed, the reflected wave of which channel depends on other methods, the waveform and spectrum of the signal before addition, or if the radar wave is a pulse wave, depending on the detection timing of the reflected wave, etc. You can tell if it is. Furthermore, as described above, when multidimensionalization is not performed, an error occurs in the arrival direction estimation due to the difference in the center frequency for each channel. If it is reflected from the same target 9, it should come from the same direction. If the direction of arrival of the target reflected wave can be estimated to some extent accurately by the individual processing system, the approximate frequency deviation can be estimated from the direction of arrival of the reflected wave detected by the addition system, that is, the steering vector error. Yes, it is good to estimate which channel is detected.

  In this way, the detection status of the target reflected wave for each channel is obtained, and the system selection control unit 12 can detect the channel stably and as accurately as possible based on the detection status collected in the detection status database 14. Select and assign to the system of individual processing. Whether it is stably detected is determined from the temporal detection probability. Whether it can be detected with high accuracy depends on the signal-to-interference and noise ratio (SINR), that is, the SNR of the target reflected wave, and the direct wave and the target reflected wave when the direct wave is not too close in angle or too close. Depending on the power ratio.

  Note that the detection status database 14 knows characteristics of a certain degree of operation of each radar transmitter, for example. For example, a radar transmitter that does not follow the target 9 always ranks lower. Also good. In this way, it is possible to avoid a situation in which a radar wave that does not follow due to a false detection comes to the top.

  With the configuration shown in FIG. 6, target detection with high accuracy is always performed from channels that can be stably detected, and processing for unstable channels with low probability of arrival and channels that are used in an auxiliary manner are performed. Can be reduced. Furthermore, even if there is a change in the detection status of these channels, the assignment of channels to the system can be changed in accordance with the detection status, and the target can be continuously detected with high accuracy.

  As for the number of systems to be individually processed, for example, all channels with good detection status may be processed individually, but if reduction of processing is taken into consideration, what is the highest number of ranking results according to detection status? Good. In the case of FIG. 6, the top two channels and two channels with good detection status are assigned to individual processing.

  For channels to be added, attention must be paid to the number of additions. In arrival wave estimation by eigenvalue analysis, the maximum number of waves that can be detected is limited to the number of elements of X (t) minus one. In the situation shown in FIG. 4, in most cases, the direct wave from the radar transmitter 8-x reaches the radar receiver 10 by the antenna side lobe. Therefore, the number of channels that can be added is not the number of elements of X (t) minus 1, but less. It also depends on how much clutter reaches each channel. Furthermore, it depends on the arrival probability of the target reflected wave of each channel. For example, in the case where clutter of about 2 waves on average (or equivalent to 2 waves) is constantly observed in any channel, the number of clutter waves (2 waves) is added to the direct wave (1 wave). If the channels are added with a very low probability of arrival, it is very rare that they are detected simultaneously in a plurality of added channels. Therefore, the number of elements of X (t) is -1 to about 1 or 2 The upper limit is the number obtained by subtracting it and dividing it by the sum of the direct wave number and the average clutter wave number (three waves). However, in the arrival wave estimation by eigenvalue analysis, the arrival wave detection accuracy deteriorates as the total number of arrival waves approaches the element number −1 of X (t), so a slight margin may be allowed. For example, the number subtracted initially from the element number −1 of X (t) may be a large number such as 3 or 4 instead of 1 or 2. Of course, how much margin is viewed depends on the number of elements of X (t) in the first place. This is because, if the number of elements of X (t) is a number such as 6, 7, for example, but if a number such as 3, 4 is subtracted, addition itself becomes impossible. In addition, although stable arrival cannot be expected, if a channel that arrives with a certain arrival probability is included in a channel to be added, it is preferable to set a large number in advance. Alternatively, the upper limit may be the number obtained by dividing the number of elements of X (t) minus 1 (minus the margin) by the total of the direct wave, average clutter wave number, and reflected wave (one wave).

  Note that the wave number of the clutter depends on the terrain between the radar transmitter and the radar receiver and the antenna structure, and therefore the tendency may be different for each channel. Therefore, the average clutter wave number used in the above method is preferably changed corresponding to the channels to be combined. Alternatively, the threshold value obtained by subtracting the expected value of the target reflected wave that may arrive at the margin from the number of elements of X (t) minus 1 as the sum of the number of incoming waves of the direct wave and the clutter in the channels to be combined. It is good to combine so as not to exceed.

  The combination of channels to be added may be, for example, a combination of channels having similar detection conditions as described above. In that case, it is better to reduce the number of additions in a combination with better detection status. It is better that the number of additions is smaller not only in terms of the total number of incoming waves but also in terms of SNR.

  As another combination method, channels having frequencies as close as possible may be combined. In the case of performing one-dimensional detection without performing multi-dimensional steering vectors, an error occurs in the estimation result due to the difference in the center frequency for each channel. This is a good way to reduce the error.

  Furthermore, in order to allow such an error and to prevent the error from affecting the estimation result, for example, a channel with a large error should not be combined with a channel with a close angular difference between the direct wave and the target reflected wave. May be. Alternatively, a combination in which the detection angle of the target reflected wave cannot be distinguished from the direct wave of another channel to be combined due to an error may be avoided. In this case, even if the angles of the direct wave and the reflected wave of the other channel to be combined are close, there is no problem if the target reflected wave is combined so as to be away from the direct wave of the other channel due to an error. In other words, by changing the center frequency when defining the steering vector in the estimation unit, it is possible to select the direction of approaching or moving away from the direct wave of another channel due to error, so the center frequency is devised. May be. However, in this case, it is necessary to know in advance the direct wave direction of each channel and the approximate direction of the target reflected wave. Since the direct wave is detected almost always, it is easy to know the direction of the direct wave, and the direction of the target reflected wave may be known from a system of individual estimation or by a method of notifying from the outside in advance. .

  In addition, when the detection status is good, that is, when the channel of individual processing is assigned to a channel with a high probability of arrival of the target reflected wave and the SNR is good and the other channels are estimated by addition, the SNR is increased in the added channel. Even if a reflected wave arrives, it may not be detected if the power is weak. However, in this embodiment, a channel having the best possible SNR of the target reflected wave is assigned to the individual processing system. In the addition processing system of the present embodiment, the detected SNR deteriorates. As a result, when the channel assignment for the processing system is switched, only the channels that can be detected even when the SNR slightly deteriorates are separated from the addition system. It is convenient to achieve the purpose.

  The timing for changing the channel assignment to the system may be, for example, about once every several seconds and periodically. However, it may be performed as needed when the detection status ranking in the detection status database 14 is changed and the combination is to be changed. That is, when it is determined that there is a channel in which the addition processing is performed and the detection status is improved over a certain period of time, or when the detection status of the individually processed system is deteriorated. . Even if the ranking is changed, it may be left as it is if it is changed within the same addition system.

  In the arrival wave estimation by eigenvalue analysis of the correlation matrix, the target reflected wave requires time required for calculating the expected value (actually the time average value) in Equation 4. Allocation change intervals shorter than this period are meaningless. The received signal always includes thermal noise, and approaches the expected value as the time averaging period becomes longer. On the other hand, if the period is lengthened, there is a possibility that the incoming signal other than the thermal noise, that is, the arrival direction of the direct wave or the target reflected wave may change during that period. Therefore, the average period cannot be made longer than necessary.

  As described above, the radar reflection cross-sectional area of the target 9 changes for each incident angle and reflection angle of the radar wave. Even in a channel where a target reflected wave has been detected to some extent, the detection state of the target reflected wave may be deteriorated due to changes in the incident angle with respect to the target 9 and the reflection angle when the radar receiver receives. How much time unit changes this depends on the position of the radar transmitter, target, radar receiver, moving speed, and the shape of the radar cross section of the target. However, if the position and speed are assumed to be a constant model, the speed of change can be determined to some extent depending on the target radar reflection cross-sectional shape. Accordingly, it is preferable to determine the replacement at a time when the detection changes to non-detection, from non-detection to detection, or at about half the interval. If it is changed each time, at least once, if it is assigned, it is preferable not to change the channel for that period.

  Note that the present invention is not limited to the above-described embodiment as it is, and can be embodied by modifying the constituent elements without departing from the scope of the invention in the implementation stage. In addition, various inventions can be formed by appropriately combining a plurality of components disclosed in the embodiment. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, constituent elements over different embodiments may be appropriately combined.

DESCRIPTION OF SYMBOLS 1, 101 ... Arrival wave detection apparatus 2 ... Antenna 3, 3-1, 3-2, 3-3, 3-4, 3-K ... RF part 4, 4-1, 4-2 4-3, 4-4, 4-5, 4-6, 4-7, 4-1-1, 4-1-2, 4-1-3, 4-K-1, 4-K-2 , 4-K-3... Filters 5, 5-1, 5-2, 5-3, 5-4, 5-5, 5-6, 5-7, 5-1-1, 5-1 2, 5-1-3, 5-K-1, 5-K-2, 5-K-3... Frequency converter 6, 6-1, 6-2, 6-K. , 7-1, 7-2, 7-3, 7-4... Estimator 8, 8-1, 8-2, 8-3, 8-4, 8-5, 8-6, 8-7 , 8-8, 8-9 ... Radar transmitter 9 ... Target 10 ... Radar receiver 11, 11-1, 11-2, 11-3, 11-4, 11-K ... Antenna element 12 · Channel selection controller 13 ... integration processing unit 14 ... detection status database

Claims (8)

An RF processing unit for receiving a plurality of radio waves from a plurality of channels having different frequency bands;
A frequency converter that converts the frequency of each of the plurality of radio waves to the same low frequency band;
An adder that adds a plurality of radio waves subjected to frequency conversion to obtain an addition signal;
An arrival wave detection apparatus comprising: an estimation unit that detects an arrival wave using eigenvalue analysis of a correlation matrix of the addition signal. A plurality of antenna elements;
The adding unit adds the plurality of radio waves for each of reception signals at the plurality of antenna elements,
The arrival wave detection apparatus according to claim 1, wherein the estimation unit generates the correlation matrix from reception signal vectors corresponding to the plurality of antenna elements and estimates arrival directions of the plurality of radio waves.   The incoming wave detection device according to claim 1, further comprising a detection status database that stores a detection status of a reflected wave from a target for each of the plurality of channels. A plurality of the estimation units, further comprising a system selection control unit,
The system selection control unit selects one or more channels that are likely to detect a reflected wave from a target based on the detection status,
For each channel with a high probability of detection, one of the plurality of estimation units independently detects an incoming wave,
The channel with a low detection possibility adds two or more in the said addition part, and an incoming wave is detected in one of these several estimation parts by an addition signal. The incoming wave detection apparatus as described. A plurality of the adding unit,
The system selection control unit has a low detection possibility so that a total value of the number of arriving waves included in each channel stored in the detection status database does not exceed a predetermined threshold for the plurality of channels with a low detection possibility. Creating a combination of channels, adding each by one of the plurality of addition units, and detecting an incoming wave for each of the obtained plurality of addition signals by one of the plurality of estimation units. The incoming wave detection device according to claim 4, characterized in that: The detection status database updates a detection status based on a result of detection of an incoming wave by the plurality of estimation units,
6. The incoming wave detection apparatus according to claim 4, wherein the system selection control unit selects a channel to be allocated to each system based on the updated detection status of the detection status database. .   The incoming wave detection apparatus according to claim 1, wherein the algorithm is MUSIC, ESPRIT, or maximum likelihood estimation. An incoming wave detection method for an incoming wave detection device,
Receive multiple radio waves from multiple channels with different frequency bands,
The frequency conversion of the plurality of radio waves to the same low frequency band,
Add a plurality of radio waves subjected to frequency conversion to obtain an addition signal,
An incoming wave detection method, wherein an incoming wave is detected using eigenvalue analysis of a correlation matrix of the addition signal.

Images