JP6053436B2  Detection apparatus, computer program, and detection method  Google Patents
Detection apparatus, computer program, and detection method Download PDFInfo
 Publication number
 JP6053436B2 JP6053436B2 JP2012220975A JP2012220975A JP6053436B2 JP 6053436 B2 JP6053436 B2 JP 6053436B2 JP 2012220975 A JP2012220975 A JP 2012220975A JP 2012220975 A JP2012220975 A JP 2012220975A JP 6053436 B2 JP6053436 B2 JP 6053436B2
 Authority
 JP
 Japan
 Prior art keywords
 acceleration
 unit
 signal
 target
 assumed
 Prior art date
 Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
 Active
Links
 238000004590 computer program Methods 0.000 title claims description 5
 238000001514 detection method Methods 0.000 title claims 4
 230000001133 acceleration Effects 0.000 claims description 219
 238000004364 calculation methods Methods 0.000 claims description 90
 239000000284 extracts Substances 0.000 claims description 6
 230000001131 transforming Effects 0.000 claims description 3
 230000001172 regenerating Effects 0.000 description 28
 230000008929 regeneration Effects 0.000 description 28
 238000010586 diagrams Methods 0.000 description 20
 238000000034 methods Methods 0.000 description 10
 229920003211 cis1,4polyisoprene Polymers 0.000 description 9
 238000009825 accumulation Methods 0.000 description 8
 230000000694 effects Effects 0.000 description 7
 230000000875 corresponding Effects 0.000 description 5
 238000007476 Maximum Likelihood Methods 0.000 description 4
 230000001427 coherent Effects 0.000 description 3
 230000036461 convulsion Effects 0.000 description 3
 238000005070 sampling Methods 0.000 description 3
 280000233134 Hardware Resources companies 0.000 description 2
 230000004069 differentiation Effects 0.000 description 2
 230000002123 temporal effects Effects 0.000 description 2
 230000005540 biological transmission Effects 0.000 description 1
 238000004422 calculation algorithm Methods 0.000 description 1
 230000002068 genetic Effects 0.000 description 1
 230000003287 optical Effects 0.000 description 1
 238000005457 optimization Methods 0.000 description 1
Images
Description
The present invention relates to a detection device that detects a target.
There is an apparatus for detecting a target by radiating an electromagnetic wave and receiving a reflected wave reflected by the radiated electromagnetic wave hitting the target.
In order to detect the target even when the reflected wave (target signal) is weak and buried in noise, the signal received in a predetermined period (signal processing cycle, SPI) is coherently integrated by fast Fourier transform, etc. There is.
However, when the target speed changes, the frequency of the reflected wave changes due to the Doppler effect.
Therefore, after detection, the SPI is divided into short periods (coherent processing period, CPI) in which the target speed can be regarded as almost constant, the signals received at each CPI are coherently integrated, and the coherently integrated signal is incoherently integrated. There is an integration (PDI) method.
There is also a shift PDI method in which the frequency of a coherently integrated signal is shifted to incoherently integrate.
Fukushima Fuyuki, Okamoto Kazuhisa, Shindo Shinji, Fujisaka Takahiko, Kobuchi Yoshio, "Improvement of Doppler Radar Acceleration Target Detection Performance Using PDI Method", IEICE Transactions B, Vol. J82B, no. 11, 211612169 pages, November 1999.
The PDI method is effective when the change in the target speed is small because the phase difference of the reflected wave between the CPIs is neglected for integration. However, if the change in the target speed is large, the signal cannot be integrated between the CPIs, and the target cannot be detected.
In the shift PDI method, since the frequency of the coherently integrated signal is shifted and integrated between the CPIs, the signal can be integrated between the CPIs even when the change in the target speed is larger than in the PDI method. However, when the change in the target speed in the CPI is so large that it cannot be ignored, the signaltonoise ratio cannot be improved by coherent integration. If the CPI is shortened so that the change in the target speed in the CPI can be ignored, the reflected waves to be coherently integrated are reduced, and the signaltonoise ratio cannot be improved. For this reason, the target cannot be detected even if the frequency of the coherently integrated signal is shifted and incoherently integrated.
For example, an object of the present invention is to detect a target even when a change in target speed is so large that the target cannot be detected by the PDI method or the shift PDI method.
The detection device according to the present invention includes:
In the detection device for detecting the target based on the signal intensity received by the reflected wave reflected by the radiated electromagnetic wave hitting the target,
An acceleration assumption unit, a compensation amount calculation unit, a compensation unit, a Fourier transform unit, an acceleration estimation unit, and a detection unit;
The acceleration assumption unit assumes a plurality of target accelerations,
The compensation amount calculation unit calculates a signal compensation amount for compensating the phase of the reflected wave for each acceleration assumed by the acceleration assumption unit,
The compensation unit compensates the phase of the reflected wave with the signal compensation amount for each acceleration assumed by the acceleration assumption unit,
The Fourier transform unit, for each acceleration assumed by the acceleration assumption unit, Fourier transform the signal intensity of the reflected wave compensated for the phase by the compensation unit,
The acceleration estimation unit estimates the target acceleration from the acceleration assumed by the acceleration assumption unit based on the signal intensity Fouriertransformed by the Fourier transformation unit,
The detection unit detects the target with respect to the acceleration estimated by the acceleration estimation unit, based on the signal intensity Fouriertransformed by the Fourier transform unit.
According to the detection apparatus of the present invention, even when the change in the target speed is so large that the target cannot be detected by the PDI method or the shift PDI method, the target can be detected.
Embodiment 1 FIG.
The first embodiment will be described with reference to FIGS.
FIG. 1 is a diagram illustrating an example of the overall configuration of a sensor signal detection system 10 according to this embodiment.
The sensor signal detection system 10 (speed measurement system) includes a sensor 11 and a sensor signal detection device 12.
The sensor 11 observes the target and outputs an observation signal 101. The sensor 11 is, for example, a high pulse repetition frequency (HPRF) Doppler radar or other radar.
The sensor signal detection device 12 (speed measurement device) analyzes the observation signal 101 to detect a target, and outputs signal detection information 102.
The sensor 11 (radiating unit / receiving unit), for example, emits a pulsed electromagnetic wave at a predetermined pulse repetition period (PRI) Δt, and receives the reflected wave reflected by the radiated electromagnetic wave hitting the target. The observation signal 101 output from the sensor 11 represents the signal intensity of the received reflected wave.
FIG. 2 is a diagram illustrating an example of hardware resources of the sensor signal detection device 12 in this embodiment.
The sensor signal detection device 12 is, for example, a computer. The sensor signal detection device 12 includes a control device 91, an input device 92, an output device 93, a storage device 94, and an arithmetic device 95.
The control device 91 controls the entire computer by executing the computer program stored in the storage device 94.
The arithmetic unit 95 performs arithmetic operations and logical operations. The arithmetic device 95 performs an operation using the data stored in the storage device 94 and generates data representing the operation result. For example, the storage device 94 stores the data generated by the arithmetic device 95.
The storage device 94 stores a computer program executed by the control device 91, data used by the calculation device 95 for calculation, and the like. The storage device 94 is, for example, an internal storage device such as a volatile memory or a nonvolatile memory, or an external storage device such as a magnetic disk device or an optical disk device.
The input device 92 inputs information from outside the computer and converts it into data. The data converted by the input device 92 is stored, for example, by the storage device 94 or used by the calculation device 95 for calculation. The input device 92 is, for example, an operation input device such as a mouse or a keyboard, an image input device such as a camera or a scanner, a voice input device such as a microphone, a physical quantity measuring device such as a temperature sensor or a pressure sensor, or a signal transmitted by another device. For example, a receiving device.
The output device 93 converts the data stored in the storage device 94 and the data generated by the arithmetic device 95 and outputs the converted data to the outside of the computer. The output device 93 is, for example, a display device that displays characters and images, an audio output device such as a speaker, a printing device that prints characters and images, and a transmission device that transmits signals to other devices.
In the functional block of the sensor signal detection device 12 described below, for example, the control device 91 executes a computer program stored in the storage device 94 to control the input device 92, the output device 93, the storage device 94, and the arithmetic device 95. Is realized. However, these functional blocks may be realized by other configurations such as other electrical configurations and mechanical configurations.
FIG. 3 is a diagram showing an example of a functional block configuration of the sensor signal detection device 12 in this embodiment.
The sensor signal detection device 12 includes, for example, an observation signal storage unit 21, an initial hypothesis generation unit 22, a signal compensation amount calculation unit 23, a signal compensation unit 24, a signal integration unit 25, a likelihood calculation unit 26, A signal detection unit 27 and a hypothesis regeneration unit 28 are included.
The observation signal storage unit 21 accumulates K observation signals 101 output from the sensor 11 for K pulses. K is an integer of 2 or more, and represents the number of pulse hits (sampling number). When PRI (sampling period) is Δt, the observation time is K × Δt.
The initial hypothesis generation unit 22 (acceleration assumption unit) generates a hypothesis group 103. The hypothesis group 103 includes, for example, N hypotheses. N is an integer of 2 or more. The hypothesis assumes a target motion model. For example, on the assumption that the target is moving at a constant acceleration, the initial hypothesis generation unit 22 sets the minimum value a _{min} based on the minimum value a _{min} and the maximum value a _{max} that can be taken by a predetermined target acceleration. N accelerations a _{n} (n is an integer of 1 to N) are selected from real numbers of _{min} or more and a maximum value a _{max} or less. The initial hypothesis generator 22, for example, a uniform random selects N pieces of acceleration a _{n.} Alternatively, the initial hypothesis generation unit 22, at regular intervals, may be configured to select the N number of acceleration a _{n.}
The signal compensation amount calculation unit 23 (compensation amount calculation unit) is based on the hypothesis group 103 generated by the initial hypothesis generation unit 22 (or the hypothesis group 109 generated by the hypothesis regeneration unit 28 described later), and the signal compensation amount 104. Calculate The signal compensation amount 104 compensates for a frequency or phase shift of the observation signal 101 due to the movement of the target during the observation time K × Δt. The signal compensation amount 104 varies depending on the target motion model. For this reason, the signal compensation amount calculation unit 23 calculates the signal compensation amount 104 for each of the N hypotheses included in the hypothesis group 103 (or the hypothesis group 109). The signal compensation amount calculation unit 23 calculates N signal compensation amounts 104.
The signal compensation unit 24 (compensation unit) compensates the observation signal sequence 105 (observation signals 101 for K pulses) stored in the observation signal storage unit 21 with the signal compensation amount 104 calculated by the signal compensation amount calculation unit 23. Thus, a compensated observation signal sequence 106 is generated. The signal compensation unit 24 generates a compensated observation signal sequence 106 for each of the N hypotheses included in the hypothesis group 103 (or hypothesis group 109) using the signal compensation amount 104 calculated by the signal compensation amount calculation unit 23. To do. The signal compensator 24 generates N compensated observation signal sequences 106.
The signal integration unit 25 (Fourier transform unit) integrates the compensated observation signal sequence 106 generated by the signal compensation unit 24 to generate an integrated observation signal sequence 107. For example, the signal integration unit 25 performs coherent integration of the compensated observation signal sequence 106 by fast Fourier transform (FFT). The signal integration unit 25 generates an integrated observation signal sequence 107 using the compensated observation signal sequence 106 generated by the signal compensation unit 24 for each of the N hypotheses included in the hypothesis group 103 (or hypothesis group 109). To do. The signal integration unit 25 generates N integrated observation signal sequences 107.
If the hypothesis matches the actual target motion, the deviation of the observation signal 101 due to the target motion disappears, and the signaltonoise ratio (SNR) of the integrated observation signal sequence 107 becomes high.
The likelihood calculation unit 26 (likelihood calculation unit) calculates a signal noise ratio SNR (integral SN ratio) of the integrated observation signal sequence 107 generated by the signal integration unit 25. The likelihood calculating unit 26 calculates, for example, the maximum amplitude value S _{n of} the integrated observation signal sequence 107 and the average amplitude value N _{n} of signals other than the maximum amplitude value S _{n} . The likelihood calculating unit 26 calculates a quotient obtained by dividing the maximum amplitude value S _{n} by the average amplitude value N _{n} to obtain a signaltonoise ratio SNR. The likelihood calculating unit 26 calculates the signaltonoise ratio SNR for each of the N hypotheses included in the hypothesis group 103 (or the hypothesis group 109) using the integrated observation signal sequence 107 generated by the signal integration unit 25. . The likelihood calculator 26 calculates N signaltonoise ratios SNR.
Likelihood calculation unit 26, based on the calculated signaltonoise ratio SNR, and calculates the likelihood w _{n.} The likelihood w _{n} represents the likelihood of hypotheses (plausibility). If the hypothesis matches the actual target motion, the signaltonoise ratio SNR increases. Therefore, it can be said that a hypothesis with a higher signaltonoise ratio SNR is more likely. For example, the likelihood calculator 26 calculates a difference obtained by subtracting a predetermined ideal signal noise ratio SNR _{ref} (ideal integral SN ratio) from the signal noise ratio SNR, and quotient obtained by dividing the calculated difference by the ideal signal noise ratio SNR _{ref.} then calculates the power of which was raised to the power of the Napier number in the calculated quotient, and the likelihood w _{n.} Likelihood calculation unit 26, for each of the N hypotheses included in the hypotheses 103 (or hypotheses 109), calculates the likelihood _{w n.} Likelihood calculation unit 26 calculates the N number of likelihood _{w n.} Likelihood calculation unit 26 outputs the likelihood information 108 representing the calculated N number of likelihood w _{n.}
Signal detection unit 27 (acceleration estimating unit, detection unit), when the maximum value of the likelihood w _{n} to the likelihood calculating unit 26 has calculated is larger than a predetermined threshold, the hypothesis corresponding to the likelihood w _{n} adopt. The signal detection unit 27 detects the target for the adopted hypothesis based on the integrated observation signal sequence 107 generated by the signal integration unit 25. For example, the signal detection unit 27 calculates a target speed and acceleration. The signal detection unit 27 outputs signal detection information 102 representing the detected result.
Hypothesis regenerating unit 28 (acceleration assuming unit), when the maximum value of the likelihood w _{n} to the likelihood calculating unit 26 has calculated is smaller than the threshold value, generates a new hypotheses 109. The hypothesis group 109 includes, for example, N hypotheses. For example, the hypothesis regenerating unit 28 selects N accelerations a _{n} (n is an integer not less than 1 and not more than N) from real numbers not less than the minimum value a _{min and not} more than the maximum value a _{max} .
Hypothesis regenerating unit 28, the likelihood w _{n} to the likelihood calculating unit 26 is calculated mainly selected hypotheses near the high hypothesis.
For example, the hypothesis regenerating unit 28, from among the previous Nnumber of hypotheses (the M, an integer. 1 or more and less than N) M pieces sequentially likelihood w _{n} is high is extracted hypotheses. Hypothesis regenerating unit 28 extracts likelihood calculating unit 26 for M hypotheses calculates the total value Σw which is the sum of likelihood w _{n} calculated. Hypothesis regenerating unit 28, the likelihood w _{n} for each M hypothesis selected, it calculates the calculated divided by the total value Σw was (normalized likelihood), the calculated quotient, select the hypothesis N hypotheses are selected from among the M hypotheses with overlap. The hypothesis regenerating unit 28 gives a fluctuation to each of the selected N hypotheses and generates new N hypotheses. For example, assuming that the maximum value to be given to acceleration as fluctuation is Δa (where Δa> 0), the hypothesis regeneration unit 28 is a real number (however, 0 is excluded) from −Δa to Δa. , A real number Δ _{n} is selected. The hypothesis regeneration unit 28 selects, for example, a real number Δ _{n} uniformly and randomly. Hypothesis regenerating unit 28, the acceleration in the selected hypothesis, and calculating a sum obtained by adding the real delta _{n} selected, the calculated sum to the acceleration a _{n} in the new hypothesis. However, the hypothesis regeneration unit 28, when the calculated sum is smaller than the minimum value _{a min,} the minimum value _{a min,} and the acceleration _{a n} in the new hypothesis, if the calculated sum is greater than the maximum value _{a max,} the maximum value a _{Let} _{max be the} acceleration an in the new hypothesis.
Based on the hypothesis group 109 generated by the hypothesis regenerating unit 28, the signal compensation amount calculating unit 23 calculates the signal compensation amount 104, the signal compensating unit 24 generates the compensated observation signal sequence 106, and the signal integrating unit 25 to produce an integrated already observed signal sequence 107, the likelihood calculation unit 26 calculates the likelihood w _{n.} _{If} the maximum value of the likelihood wn is equal to or greater than the threshold value, the signal detection unit 27 outputs the signal detection information 102.
If the maximum value of the likelihood w _{n} is less than the threshold, the hypothesis regenerating unit 28 compares the number of times that generated the hypotheses 109 and upper limit number of times. When the number of generations of the hypothesis group 109 is smaller than the upper limit number L, the hypothesis regenerating unit 28 generates a new hypothesis group 109.
When the number of generations of the hypothesis group 109 reaches the upper limit number L, the hypothesis regeneration unit 28 does not generate a new hypothesis group 109. Signal detector 27, taken from among N × (L + 1) number of hypotheses included in the hypotheses 103 and the L hypotheses 109, likelihood w _{n} to the likelihood calculating unit 26 has calculated the hypothesis is the maximum To do. The signal detection unit 27 detects a target for the adopted hypothesis based on the integrated observation signal sequence 107 generated by the signal integration unit 25, and outputs signal detection information 102 representing the detection result.
FIG. 4 is a diagram illustrating an example of the relationship between the electromagnetic wave radiated by the sensor 11 and the reflected wave reflected by the target 99 in this embodiment.
The vertical direction shows the flow of time. The horizontal direction indicates the distance between the sensor 11 and the target 99.
The sensor 11 repeatedly emits pulsed electromagnetic waves at a predetermined pulse repetition period. The target 99 performs a constant acceleration motion at a predetermined acceleration.
The electromagnetic wave radiated from the sensor 11 at time 141 hits the target 99 at the reflection position 151 and is reflected, and the sensor 11 receives the reflected wave at time 161.
The electromagnetic wave radiated from the sensor 11 at the time 142 strikes the target 99 at the reflection position 152 and is reflected, and the sensor 11 receives the reflected wave at the time 162.
The electromagnetic wave radiated from the sensor 11 at the time 143 hits the target 99 at the reflection position 153 and is reflected, and the sensor 11 receives the reflected wave at the time 163.
The electromagnetic wave radiated from the sensor 11 at the time 144 hits the target 99 at the reflection position 154 and is reflected, and the sensor 11 receives the reflected wave at the time 164.
The electromagnetic wave radiated by the sensor 11 at time 145 hits the target 99 at the reflection position 155 and is reflected, and the sensor 11 receives the reflected wave at time 165.
Although the interval at which the sensor 11 radiates electromagnetic waves is constant, the reflection positions 151 to 155 are different, so the interval at which the sensor 11 receives the reflected wave is not constant.
Moreover, since the target 99 is moving, the frequency of the reflected wave received by the sensor 11 is different from the frequency of the electromagnetic wave radiated by the sensor 11 due to the Doppler effect. In this example, since the target 99 is moving in the direction approaching the sensor 11, the frequency of the reflected wave is higher than the frequency of the radiated wave.
FIG. 5 is a diagram showing an example of the relationship between the electromagnetic wave radiated by the sensor 11 and the reflected wave reflected by the target 99 in this embodiment.
The horizontal axis represents time. The reason why there are a plurality of horizontal axes is that the display is folded and displayed at the pulse repetition period 140. The intersection of the horizontal axis and the line 148 represents the same time as the intersection of the horizontal axis and the line 149 one level above.
The electromagnetic wave 171 is an electromagnetic wave emitted by the sensor 11 at time 141 (see FIG. 4). The reflected wave 181 is a reflected wave received by the sensor 11 at time 161 when the electromagnetic wave 171 is reflected by the target 99 at the reflection position 151.
The electromagnetic wave 172 is an electromagnetic wave emitted by the sensor 11 at time 142. The reflected wave 182 is a reflected wave received by the sensor 11 at time 162 when the electromagnetic wave 172 is reflected by the target 99 at the reflection position 152.
The electromagnetic wave 173 is an electromagnetic wave emitted by the sensor 11 at time 143. The reflected wave 183 is a reflected wave received by the sensor 11 at time 163 when the electromagnetic wave 173 is reflected by the target 99 at the reflection position 153.
The electromagnetic wave 174 is an electromagnetic wave emitted by the sensor 11 at time 144. The reflected wave 184 is a reflected wave received by the sensor 11 at time 164 when the electromagnetic wave 174 is reflected by the target 99 at the reflection position 154.
Actually, since the reflected waves 181 to 184 are weak, they are buried in noise and cannot be clearly identified as shown in this figure.
FIG. 6 is a diagram illustrating an example of the signal intensity received by the sensor 11 in this embodiment in the frequency domain.
The horizontal axis indicates the frequency.
The top represents the signal intensity in the pulse repetition period 140 including the reflected wave 181.
The second represents the signal intensity in the pulse repetition period 140 including the reflected wave 182.
The third represents the signal intensity in the pulse repetition period 140 including the reflected wave 183.
The fourth represents the signal intensity in the pulse repetition period 140 including the reflected wave 184.
In this example, since the target 99 moves in a direction approaching the sensor 11, the frequency of the reflected waves 181 to 184 is higher than the frequency 170 of the electromagnetic wave radiated by the sensor 11.
In this example, since the target 99 is decelerated, the frequency of the reflected wave 182 is lower than the frequency of the reflected wave 181, the frequency of the reflected wave 183 is lower than the frequency of the reflected wave 182, and the reflected wave 183 The frequency of the reflected wave 184 is lower than the frequency of.
Since the reflected waves 181 to 184 are weak, they are buried in the noise 180 even in the frequency domain and cannot be clearly identified.
FIG. 7 is a diagram illustrating an example of the compensated observation signal sequence 106 generated by the signal compensation unit 24 in this embodiment.
The horizontal axis indicates time as in FIG.
Compensating the observation signal changes the frequency and phase of the reflected waves 181 to 184 (and noise not shown). However, it does not change that the reflected waves 181 to 184 are buried in noise.
FIG. 8 is a diagram showing an example of the signal strength represented by the compensated observation signal sequence 106 generated by the signal compensation unit 24 in this embodiment in the frequency domain.
The horizontal axis indicates the frequency.
The top represents the signal intensity in the pulse repetition period 140 including the reflected wave 181.
The second represents the signal intensity in the pulse repetition period 140 including the reflected wave 182.
The third represents the signal intensity in the pulse repetition period 140 including the reflected wave 183.
The fourth represents the signal intensity in the pulse repetition period 140 including the reflected wave 184.
When the target 99 is moving at a constant acceleration and the acceleration of the target 99 matches the acceleration assumed in the hypothesis, the frequencies and phases of the reflected waves 181 to 184 match.
FIG. 9 is a diagram illustrating an example of the signal strength represented by the integrated observation signal sequence 107 generated by the signal integration unit 25 in this embodiment.
The horizontal axis indicates the frequency.
If the target 99 is moving at an equal acceleration and the acceleration of the target 99 matches the acceleration assumed in the hypothesis, the reflected waves 181 to 184 have the same frequency and phase. The signal strength of 189 increases. Since the phase of the noise 180 is different, the signal intensity does not increase so much even if the noise 180 is integrated.
As a result, the reflected wave 189 can be clearly identified from the noise 180.
FIG. 10 is a diagram illustrating an example of the signaltonoise ratio 190 of the integrated observation signal sequence 107 in this embodiment.
The horizontal axis indicates the acceleration of the target 99 assumed in the hypothesis.
The vertical axis represents the signaltonoise ratio of the integrated observation signal sequence 107.
A broken line 191 represents a minimum value a _{min} of a predetermined range as a range that the acceleration of the target 99 can take. A broken line 192 represents a maximum value a _{max} of a predetermined range as a range that the acceleration of the target 99 can take.
A broken line 193 represents a threshold value of the signaltonoise ratio 190 corresponding to a likelihood threshold value for determining whether or not the signal detection unit 27 adopts a hypothesis. As the signaltonoise ratio 190 increases, the likelihood calculated by the likelihood calculating unit 26 increases. Therefore, the signal detecting unit 27 adopts the hypothesis when the signaltonoise ratio 190 is larger than the threshold value indicated by the broken line 193.
If the signaltonoise ratio 190 is calculated for all accelerations of the minimum value a _{min} or more and the maximum value a _{max} or less, it is easy to find the acceleration at which the signaltonoise ratio 190 is maximized, but the calculation amount for that is enormous.
In order to detect the target 99 in real time, it is necessary to suppress the calculation amount as much as possible.
The sensor signal detection apparatus 12 in this embodiment starts from the hypothesis group 103, calculates the likelihood of each hypothesis, repeats the regeneration of the hypothesis group 109, and selects hypotheses whose likelihood exceeds the threshold. Explore. It is good if the likelihood exceeds the threshold value, and not to search for a hypothesis that maximizes the likelihood is to prioritize detecting the target 99 with a small amount of calculation rather than accurately measuring the acceleration of the target 99. Because. Accordingly, the sensor signal detection device 12 may be configured to find a hypothesis whose likelihood exceeds a threshold and then search for a hypothesis with a higher likelihood if the calculation amount has a surplus, and further search is continued. .
Next, the signal compensation amount 104 calculated by the signal compensation amount calculation unit 23 will be described.
Let R (t) be the distance between the sensor 11 and the target 99 at time t (target distance from the sensor 11). Assuming that the time when the sensor 11 radiates the electromagnetic wave reflected at the target 99 at time t is t _{1} and the time when the sensor 11 receives the reflected wave reflected at the target 99 at time t is t _{2} ,
When this is differentiated by t, the following equation is obtained.
From this, the following equation is obtained.
Here, let us consider Taylor expansion in the vicinity of t = t _{0} in the expression expressing t _{1} as t _{2} .
Assume that the intensity v (t _{1} ) of the electromagnetic wave radiated by the sensor 11 at time t _{1} is expressed by the following equation.
At this time, the intensity z (t _{2} ) of the reflected wave received by the sensor 11 at time t _{2} can be expressed by the following equation.
Therefore, the signal compensation amount calculation unit 23 calculates a signal compensation amount z _{c} (t _{2} ) represented by the following equation.
The signal compensation unit 24 calculates a product of the reflected wave intensity z (t _{2} ) and the signal compensation amount z _{c} (t _{2} ) to obtain a compensated signal intensity z hat (t _{2} ).
If the third and higher order terms of t _{2} are ignored, the constant C _{2} is expressed by the following equation.
Therefore, the signal compensation unit 24 may be configured to calculate the signal compensation amount z _{c} (t _{2} ) represented by the following equation.
If the target 99 is detected, the speed of the target 99 can be known from the Doppler frequency. If the acceleration of the target 99 is to be known more accurately, the constant C _{2} in the adopted hypothesis and the speed of the target 99 are expressed by the following equation. And the acceleration of the target 99 may be obtained.
FIG. 11 is a diagram showing an example of the flow of the signal detection processing 120 in this embodiment.
In the signal detection process 120, the sensor signal detection device 12 detects the target 99 from the observation signal 101 output from the sensor 11. The signal detection processing 120 includes, for example, an observation signal accumulation step 121, an initial hypothesis generation step 122, a hypothesis selection step 123, a signal compensation amount calculation step 124, a signal compensation step 125, a signal integration step 126, a hypothesis likelihood. Degree calculation step 127, search end determination step 128, hypothesis regeneration step 129, and signal detection step 130. The sensor signal detection device 12 starts the signal detection processing 120 from the observation signal accumulation step 121.
In the observation signal accumulation step 121, the observation signal storage unit 21 receives and stores the observation signal 101 output from the sensor 11. When the observation signal storage unit 21 stores the observation signal 101 for the observation time K × Δt, the sensor signal detection device 12 proceeds to the initial hypothesis generation step 122.
In the initial hypothesis generation step 122, the initial hypothesis generation unit 22 generates a hypothesis group 103 composed of N hypotheses.
In the hypothesis selection step 123, the signal compensation amount calculation unit 23 generates the hypothesis group 103 generated by the initial hypothesis generation unit 22 in the initial hypothesis generation step 122 or the hypothesis group generated by the hypothesis regeneration unit 28 in the hypothesis regeneration step 129. A hypothesis that has not yet been processed is selected from the N hypotheses included in 109.
If all of the N hypotheses have been processed and there is no hypothesis that has not yet been processed, the signal compensation amount calculation unit 23 advances the processing to the search end determination step 128.
If there is a hypothesis that has not yet been processed among the N hypotheses, the signal compensation amount calculation unit 23 selects one hypothesis from among the hypotheses that have not yet been processed, and proceeds to the signal compensation amount calculation step 124. Proceed with the process.
In the signal compensation amount calculation step 124, the signal compensation amount calculation unit 23 calculates the signal compensation amount 104 based on the hypothesis selected in the hypothesis selection step 123.
In the signal compensation step 125, the signal compensation unit 24 includes the observation signal sequence 105 including the observation signal 101 stored in the observation signal storage unit 21 in the observation signal accumulation step 121, and the signal compensation amount calculation unit 23 in the signal compensation amount calculation step 124. The compensated observation signal sequence 106 is calculated based on the signal compensation amount 104 calculated by.
In the signal integration step 126, the signal integration unit 25 calculates the integrated observation signal sequence 107 based on the compensated observation signal sequence 106 calculated by the signal compensation unit 24 in the signal compensation step 125.
In the hypothesis likelihood calculation step 127, the likelihood calculation unit 26 selects the signal compensation amount calculation unit 23 in the hypothesis selection step 123 based on the integrated observed signal sequence 107 calculated by the signal integration unit 25 in the signal integration step 126. The likelihood of the hypothesis is calculated.
The signal compensation amount calculation unit 23 returns the process to the hypothesis selection step 123 and selects the next hypothesis.
In the search end determination step 128, the signal detection unit 27 determines whether or not to end the search. For example, the signal detection unit 27 determines whether or not the maximum value among the N likelihoods calculated by the likelihood calculation unit 26 in the hypothesis likelihood calculation step 127 is greater than a threshold value.
If the maximum likelihood value is greater than the threshold value, the signal detection unit 27 determines that the search is to be terminated, and proceeds to the signal detection step 130.
When the maximum likelihood value is less than or equal to the threshold value, the signal detection unit 27 determines whether or not the number of times the hypothesis regenerating step 129 has been executed has reached the upper limit number.
When the number of executions of the hypothesis regeneration step 129 reaches the upper limit number, the signal detection unit 27 determines to end the search, and proceeds to the signal detection step 130.
If the number of times the hypothesis regenerating step 129 has been executed is less than the upper limit number, the signal detection unit 27 determines that the search has not ended yet, and proceeds to the hypothesis regenerating step 129.
In the hypothesis regeneration step 129, the hypothesis regeneration unit 28, based on the N hypotheses at that time and the likelihood calculated by the likelihood calculation unit 26 in the hypothesis likelihood calculation step 127 for each hypothesis, A hypothesis group 109 composed of new N hypotheses is generated.
The signal compensation amount calculation unit 23 returns the process to the hypothesis selection step 123 and selects a hypothesis from the new hypothesis group 109.
In the signal detection step 130, the signal detection unit 27 detects the target 99 using the hypothesis having the highest likelihood calculated by the likelihood calculation unit 26 in the hypothesis likelihood calculation step 127.
The observation signal storage unit 21 returns the processing to the observation signal accumulation step 121 and stores the next observation signal 101.
The initial hypothesis generation process 122 for the second and subsequent times is executed after the observation signal storage unit 21 newly stores the observation signal 101 for the observation time K × Δt. That is, the sensor signal detection device 12 divides the observation signal 101 output from the sensor 11 for each observation time K × Δt and processes each separately.
The initial hypothesis generation process 122 for the second and subsequent times may be executed after the observation signal storage unit 21 newly stores the observation signal 101 for the pulse repetition period Δt. In that case, the signal compensation unit 24 excludes the observation signal 101 corresponding to the oldest pulse repetition period Δt from among the observation signals 101 included in the observation signal sequence 105 processed last time, and the observation signal 101 corresponding to (K−1) × Δt. Then, the observation signal sequence 105 composed of the observation signals 101 for K × Δt, which is composed of the observation signals 101 for Δt newly stored in the observation signal storage unit 21, is processed.
Further, in the initial hypothesis generation step 122 after the second time, the initial hypothesis generation unit 22 may generate the hypothesis group 103 based on the hypothesis adopted by the signal detection unit 27 in the previous signal detection step 130. Good. For example, the initial hypothesis generation unit 22, by increasing the probability of selecting as the acceleration closer to an acceleration a in the hypothesis that the signal detector 27 is employed to select the N number of acceleration a _{n} randomly.
FIG. 12 is a diagram showing a specific example of the signal detection processing 120 in this embodiment.
The horizontal axis indicates the acceleration of the target 99 assumed in the hypothesis.
The vertical axis represents the probability that the initial hypothesis generator 22 and the hypothesis regenerator 28 select the acceleration as a hypothesis, or the signaltonoise ratio of the integrated observation signal sequence 107 calculated by the signal integrator 25.
In the initial hypothesis generation step 122, the probability that the initial hypothesis generation unit 22 selects acceleration is uniform within a range from the minimum value a _{min} represented by the broken line 191 to the maximum value a _{max} represented by the broken line 192.
Here, it is assumed that the initial hypothesis generation unit 22 has selected N accelerations (hypotheses) represented by arrows (N = 10 in this example).
For each of the N accelerations selected by the initial hypothesis generation unit 22, in the hypothesis selection step 123 to the hypothesis likelihood calculation step 127, the signal compensation amount calculation unit 23 calculates the signal compensation amount 104, and the signal compensation unit 24 has already been compensated. The observation signal sequence 106 is calculated, the signal integration unit 25 calculates the integrated observation signal sequence 107, and the likelihood calculation unit 26 calculates the likelihood.
Since the maximum value of the calculated likelihood (signaltonoise ratio) is smaller than the threshold value, the signal detection unit 27 advances the process to the hypothesis regeneration step 129 in the search end determination step 128.
In the hypothesis regeneration step 129, the hypothesis regeneration unit 28 extracts M (M = 5 in this example) acceleration from the N accelerations in descending order of likelihood. The hypothesis regenerating unit 28 selects a range of −Δa or more and + Δa or less around the extracted accelerations as a selection range. As in this example, the M selection ranges may overlap each other. The selection probability of each selection range is proportional to the likelihood of the central acceleration. The selection probability of acceleration in each selection range is equal. When the selection ranges overlap, the acceleration selection probabilities within the ranges are the sum of the acceleration selection probabilities in each of the overlapping selection ranges.
Here, it is assumed that the hypothesis regeneration unit 28 has selected N accelerations represented by arrows.
For each of the N accelerations selected by the hypothesis regeneration unit 28, in the hypothesis selection step 123 to the hypothesis likelihood calculation step 127, the signal compensation amount calculation unit 23 calculates the signal compensation amount 104, and the signal compensation unit 24 has already been compensated. The observation signal sequence 106 is calculated, the signal integration unit 25 calculates the integrated observation signal sequence 107, and the likelihood calculation unit 26 calculates the likelihood.
Since the calculated maximum likelihood value is smaller than the threshold value, in the search end determination step 128, the signal detection unit 27 proceeds to the hypothesis regeneration step 129 again.
In the hypothesis regeneration step 129, the hypothesis regeneration unit 28 extracts M accelerations from the N accelerations, determines an acceleration selection probability, and selects new N accelerations.
For each of the N accelerations selected by the hypothesis regeneration unit 28, in the hypothesis selection step 123 to the hypothesis likelihood calculation step 127, the signal compensation amount calculation unit 23 calculates the signal compensation amount 104, and the signal compensation unit 24 has already been compensated. The observation signal sequence 106 is calculated, the signal integration unit 25 calculates the integrated observation signal sequence 107, and the likelihood calculation unit 26 calculates the likelihood.
Since the calculated maximum value of the likelihood is larger than the threshold value, in the search end determination step 128, the signal detection unit 27 advances the process to the signal detection step 130 again.
In the signal detection step 130, the signal detection unit 27 detects the target 99 using the acceleration with the maximum likelihood as the estimated acceleration of the target 99. In this example, there are two accelerations whose likelihood exceeds the threshold, but the signal detection unit 27 sets the acceleration with the higher likelihood as the estimated acceleration of the target 99.
In this manner, hypothesis regeneration is repeated while changing the acceleration selection probability. As shown in this figure, although there are irregularities in fine parts, as a general rule, the closer the assumed acceleration is to the acceleration at which the signaltonoise ratio is maximized, the larger the signaltonoise ratio of the integrated observation signal sequence 107 becomes. . For this reason, as the hypothesis regeneration is repeated, the closer to the acceleration at which the signaltonoise ratio is maximized, the higher the probability of being selected.
As a result, it is possible to find a hypothesis in which the signaltonoise ratio of the integrated observation signal sequence 107 exceeds the threshold with a small amount of calculation, and the target can be detected.
If it is known in advance that the target 99 is approaching while accelerating, or if it is only necessary to detect the target 99 approaching while accelerating, the initial hypothesis generating unit 22 and the hypothesis regenerating unit 28, for example, A hypothesis is generated by setting a _{min} <a _{max} = 0.
Conversely, when it is known in advance that the target 99 is approaching while decelerating, or when it is only necessary to detect the target 99 approaching while decelerating, the initial hypothesis generating unit 22 and hypothesis regenerating unit 28, for example, , A _{max} > a _{min} = 0, generating a hypothesis.
If there is no such premise and it is desired to detect both the acceleration target and the deceleration target, the initial hypothesis generation unit 22 and the hypothesis regeneration unit 28 generate hypotheses, for example, as a _{min} <0 <a _{max} .
Further, the hypotheses generated by the initial hypothesis generator 22 and the hypothesis regenerator 28 do not have to assume that the target 99 is in a uniform acceleration motion. For example, the hypotheses generated by the initial hypothesis generator 22 and the hypothesis regenerator 28 may assume that the target 99 is performing an equal jerk motion. In that case, the initial hypothesis generator 22 and the hypothesis regenerator 28 assume the acceleration 99 and the jerk b of the target 99, and assume a set of the assumed acceleration a and jerk b as a hypothesis.
The signal compensation amount calculating section 23, instead of ignoring the third or higher order term of t _{2,} may be configured to calculate the signal compensation value 104. For example, signal compensation amount calculation unit 23, into account up to the third order terms of t _{2,} ignoring fourthorder or higherorder term of t _{2,} and calculates a signal compensation value 104.
Further, the signal compensation amount calculation unit 23 may be configured to calculate the signal compensation amount 104 as follows.
For example, the state vector of the target 99 is defined by the following equation.
Assuming that the target 99 is accelerating at an acceleration a, a motion model of the target 99 for the kth hit is defined by the following equation.
Based on the target distance R _{k} of k hit th defines a model of the baseband signal by the following equation.
Signal compensation amount calculating section 23, and the three equations, assumed on the basis of the acceleration a _{n,} calculates a signal compensation value [delta] _{k, n} of k hits th. Signal compensation amount [delta] _{k, n,} from the temporal relationship between the acceleration a _{n} and the signal which is assumed in an amount to cancel the effects of acceleration on the distance and speed. The signal compensation unit 24 calculates the compensated observation signal z hat _{k, n} by the following equation.
The likelihood calculating unit 26 may be configured to calculate the likelihood based on other index values instead of the signaltonoise ratio SNR calculated by the calculation method described above. For example, the likelihood calculating unit 26, a maximum amplitude value S _{n} of integration already observed signal sequence 107, the average amplitude value N _{n} of maximum amplitude S _{n} other signals, the amplitude standard of the amplitude maximum value S _{n} than the signal Deviation σ _{n} is calculated. The likelihood calculating unit 26 calculates a difference obtained by subtracting the average amplitude value N _{n} from the maximum amplitude value S _{n} . The likelihood calculating unit 26 calculates a quotient obtained by dividing the calculated difference by the amplitude standard deviation σ _{n} , and uses it as an index value instead of the signaltonoise ratio SNR.
Moreover, the hypothesis regenerating unit 28, an acceleration a _{n} in the new hypothesis, the minimum value a _{min} or more, scheme fall within the scope of the following maximum value a _{max} may be, for example, a method as follows.
The hypothesis regenerating unit 28 selects the larger one of the difference between the acceleration obtained by subtracting Δa from the acceleration of each of the N hypotheses selected with duplication from the extracted M hypotheses and the minimum value a _{min} as a ′. _{Let} it be _{min} . The hypothesis regenerating unit 28 sets a smaller one of the sum of the acceleration in each of the N hypotheses plus Δa and the maximum value a _{max} as a ′ _{max} . Hypothesis regenerating unit 28, from among the following real a _{'min} or a' _{max,} select the real randomly, and acceleration _{a n} in the new hypotheses.
Further, the fluctuation added to the hypothesis by the hypothesis regenerating unit 28 may not be selected with a uniform probability from a predetermined range. For example, the hypothesis regeneration unit 28 selects a real number to be added to the acceleration in each of the selected N hypotheses by lowering the selection probability for a real number having a smaller absolute value and lowering the selection probability for a real number having a larger absolute value. To do. For example, for a real number a in the range of more than Δa and less than 0, the selection probability is proportional to Δa + a, and for a real number a in the range of more than 0 and less than Δa, the selection probability is proportional to Δaa. For example, the hypothesis regenerating unit 28 selects a real number uniformly from among real numbers of 0 or more and less than Δa ^{2} , calculates a square root of the selected real number, and calculates a difference obtained by subtracting the calculated square root from Δa. . The hypothesis regenerating unit 28 selects an integer uniformly and randomly from two integers of −1 or 1, calculates a product obtained by multiplying the selected integer by the calculated difference, and adds a real number to the acceleration. To do.
Note that the method for finding a hypothesis in which the signaltonoise ratio of the integrated observation signal sequence 107 exceeds the threshold may be another optimization method such as a genetic algorithm.
Further, the signal detection unit 27 may be configured to determine that the search is ended when the relative error of the integral SN ratio with respect to the ideal integral SN ratio is equal to or less than a predetermined threshold.
The signal detection information 102 the signal detection unit 27 outputs the signal integration result, acceleration a _{n} which _{employs,} may be configured, including integration SN ratio.
The sensor signal detection device (12) in this embodiment detects a weak target signal in the sensor (11).
The sensor signal detection apparatus includes an observation signal accumulation unit (observation signal storage unit 21), an initial hypothesis generation unit (22), a signal compensation amount calculation unit (23), a signal compensation unit (24), and a signal integration unit ( 25), a likelihood calculating unit (26), a signal detecting unit (27), and a hypothesis regenerating unit (28).
The observation signal accumulating unit accumulates a plurality of sampling observation signals (101).
The initial hypothesis generation unit generates a plurality of hypothesized target motion parameter (acceleration) candidates (hypotheses).
The signal compensation amount calculation unit calculates a signal compensation amount (104) for the observation signal for each hypothesis from the temporal relationship between the target motion parameter and the observation signal.
For each hypothesis, the signal compensation unit converts the observation signal sequence (105) accumulated by the observation signal accumulation unit into a compensated observation signal sequence (106) based on the signal compensation amount.
The signal integration unit converts the compensated observation signal sequence into an integrated observation signal sequence (107) for each hypothesis.
The likelihood calculation unit calculates a likelihood representing a signal integration effect with respect to an ideal value based on the integrated observation signal sequence for each hypothesis.
The signal detection unit determines whether or not the search for the target motion parameter is completed based on the calculation result of the likelihood for each hypothesis, and calculates the likelihood for each hypothesis when it is determined that the search is ended. Based on the result, the signal integration result is output.
When the signal detection unit determines that the search has not been completed, the hypothesis regenerating unit regenerates the hypothesis based on the likelihood calculation result for each hypothesis.
The initial hypothesis generator (22) generates the hypothesis using uniform random numbers within a predetermined search range.
The initial hypothesis generation unit (22) generates the hypothesis by dividing a predetermined search range at equal intervals.
The likelihood calculation unit (26) calculates the likelihood based on the maximum amplitude value of the integrated observation signal sequence (107), the average amplitude value other than the maximum value, and the standard deviation.
The signal detection unit (27) determines that the search is finished when the maximum value of the likelihood is equal to or greater than a predetermined threshold value or when the number of searches reaches a predetermined upper limit value.
The hypothesis regenerating unit (28) selects only hypotheses with the highest likelihood, restores and extracts hypotheses with a probability proportional to the likelihood, and regenerates the hypotheses by adding random numbers.
Set up multiple acceleration hypotheses to compensate the observed signal, select acceleration hypotheses with a large signal integration effect, and regenerate acceleration hypotheses based on those acceleration hypotheses to maximize the signal integration effect By adopting the acceleration hypothesis (that is, the integration path), the weak target signal can be detected even when it is difficult to determine the integration path by sequential processing.
In addition, by evaluating the signal integration effect for multiple acceleration hypotheses as a likelihood and searching for the acceleration hypothesis around the acceleration hypothesis with a high likelihood, even if the true target acceleration is unknown, it is weak with a small calculation load. A target signal can be detected.
The configuration described above is an example, and other configurations may be used. For example, a configuration in which a configuration of a nonessential part is replaced with another configuration may be used.
The detection device (sensor signal detection device 12) described above detects the target based on the signal intensity received from the reflected wave reflected by the radiated electromagnetic wave hitting the target (99).
The detection apparatus includes an acceleration assumption unit (initial hypothesis generation unit 22, hypothesis regeneration unit 28), compensation amount calculation unit (signal compensation amount calculation unit 23), compensation unit (signal compensation unit 24), and Fourier transform unit ( A signal integration unit 25), an acceleration estimation unit (signal detection unit 27), and a detection unit (signal detection unit 27).
The acceleration assumption unit assumes a plurality of target accelerations.
The compensation amount calculation unit calculates a signal compensation amount (104) for compensating the phase of the reflected wave for each acceleration assumed by the acceleration assumption unit.
The compensation unit compensates the phase of the reflected wave with the signal compensation amount for each acceleration assumed by the acceleration assumption unit.
The Fourier transform unit Fouriertransforms the signal intensity (106) of the reflected wave whose phase is compensated by the compensation unit for each acceleration assumed by the acceleration assumption unit.
The acceleration estimation unit estimates the target acceleration from the acceleration assumed by the acceleration assumption unit based on the signal intensity (107) Fouriertransformed by the Fourier transform unit.
The detection unit detects the target with respect to the acceleration estimated by the acceleration estimation unit based on the signal intensity Fouriertransformed by the Fourier transform unit.
Thereby, even when the acceleration of the target is large, the target can be detected.
The detection device (12) further includes a likelihood calculating unit (likelihood calculating unit 26).
The likelihood calculation unit determines that the target acceleration is the assumed acceleration based on the signal intensity Fouriertransformed by the Fourier transform unit (25) for each acceleration assumed by the acceleration assumption unit (22, 28). A certain likelihood is calculated.
The acceleration assumption unit (28) calculates a selection probability of the assumed acceleration based on the likelihood calculated by the likelihood calculation unit, and randomly adds a plurality of target accelerations based on the calculated selection probability. Assume.
As a result, the target can be detected with a small amount of calculation.
The acceleration assumption unit (28) extracts a predetermined number of accelerations in descending order of the likelihood calculated by the likelihood calculation unit (27) from the assumed accelerations, and calculates the acceleration for each of the extracted accelerations. A selection range as a center is set, and for each set selection range, a selection probability proportional to the likelihood calculated by the likelihood calculation unit for the acceleration is set.
Thereby, acceleration with high likelihood can be found with a small amount of calculation.
The likelihood calculating unit (27) is based on a signaltonoise ratio obtained by dividing the signal strength at the frequency at which the signal strength is maximum by the average value of the signal strength at frequencies other than the frequency at which the signal strength is maximum. The likelihood is calculated.
Thereby, an acceleration with a high signaltonoise ratio can be found.
The acceleration estimator (27) uses the signal intensity at the frequency at which the signal intensity is maximum for the acceleration assumed by the acceleration assumption section (22, 28), and the signal at a frequency other than the frequency at which the signal intensity is maximum. When the signaltonoise ratio divided by the average value of the intensity is equal to or greater than a predetermined threshold, the acceleration is set as the estimated target acceleration.
As a result, an acceleration with a high signaltonoise ratio can be found with a small amount of calculation.
DESCRIPTION OF SYMBOLS 10 Sensor signal detection system, 11 Sensor, 12 Sensor signal detection apparatus, 21 Observation signal storage part, 22 Initial hypothesis generation part, 23 Signal compensation amount calculation part, 24 Signal compensation part, 25 Signal integration part, 26 Likelihood calculation part, 27 Signal detection unit, 28 Hypothesis regenerating unit, 91 Control device, 92 Input device, 93 Output device, 94 Storage device, 95 Arithmetic unit, 99 Target, 101 Observation signal, 102 Signal detection information, 103, 109 Hypothesis group, 104 Signal compensation amount, 105 observation signal sequence, 106 compensated observation signal sequence, 107 integrated observation signal sequence, 108 likelihood information, 120 signal detection process, 121 observation signal accumulation step, 122 initial hypothesis generation step, 123 hypothesis selection step, 124 signal compensation amount calculation step, 125 signal compensation step, 126 signal integration step, 127 hypothesis likelihood calculation step 128 search end determination step, 129 hypothesis regeneration step, 130 signal detection step, 140 pulse repetition period, 141145, 161165 time, 148, 149 line, 151155 reflection position, 170 frequency, 171175 electromagnetic wave, 180 noise, 181 to 184, 189 reflected wave, 190 signal to noise ratio, 191 to 193 broken line.
Claims (7)
 In the detection device for detecting the target based on the signal intensity received by the reflected wave reflected by the radiated electromagnetic wave hitting the target,
An acceleration assumption unit, a compensation amount calculation unit, a compensation unit, a Fourier transform unit, an acceleration estimation unit, a detection unit, and a likelihood calculation unit ,
The acceleration assumption unit assumes a plurality of target accelerations,
The compensation amount calculation unit calculates a signal compensation amount for compensating the phase of the reflected wave for each acceleration assumed by the acceleration assumption unit,
The compensation unit compensates the phase of the reflected wave with the signal compensation amount for each acceleration assumed by the acceleration assumption unit,
The Fourier transform unit, for each acceleration assumed by the acceleration assumption unit, Fourier transform the signal intensity of the reflected wave compensated for the phase by the compensation unit,
The acceleration estimation unit estimates the target acceleration from the acceleration assumed by the acceleration assumption unit based on the signal intensity Fouriertransformed by the Fourier transformation unit,
The detection unit is a detection device that detects the target based on the signal intensity Fouriertransformed by the Fourier transform unit with respect to the acceleration estimated by the acceleration estimation unit ,
The likelihood calculation unit calculates the likelihood that the target acceleration is the assumed acceleration based on the signal intensity Fouriertransformed by the Fourier transform unit for each acceleration assumed by the acceleration assumption unit,
The acceleration assumption unit, based on the likelihood that the likelihood calculating unit has calculated, to calculate the probability of choosing assumed acceleration, based on the calculated selection probabilities, a plurality assumed randomly newly acceleration of the target detection equipment shall be the characterized in that.  The acceleration assumption unit extracts a predetermined number of accelerations in descending order of the likelihood calculated by the likelihood calculation unit from the assumed accelerations, and selects a selection range centered on the acceleration for each of the extracted accelerations. The detection apparatus according to claim 1 , wherein a selection probability proportional to the likelihood calculated by the likelihood calculation unit for the acceleration is set for each of the set selection ranges.
 The likelihood calculating unit is configured to calculate the likelihood based on a signaltonoise ratio obtained by dividing the signal strength at the frequency at which the signal strength is maximum by the average value of the signal strength at a frequency other than the frequency at which the signal strength is maximum. detection device according to claim 1 or claim 2, characterized in that calculated.
 In the detection device for detecting the target based on the signal intensity received by the reflected wave reflected by the radiated electromagnetic wave hitting the target,
An acceleration assumption unit, a compensation amount calculation unit, a compensation unit, a Fourier transform unit, an acceleration estimation unit, and a detection unit;
The acceleration assumption unit assumes a plurality of target accelerations,
The compensation amount calculation unit calculates a signal compensation amount for compensating the phase of the reflected wave for each acceleration assumed by the acceleration assumption unit,
The compensation unit compensates the phase of the reflected wave with the signal compensation amount for each acceleration assumed by the acceleration assumption unit,
The Fourier transform unit, for each acceleration assumed by the acceleration assumption unit, Fourier transform the signal intensity of the reflected wave compensated for the phase by the compensation unit,
The acceleration estimation unit estimates the target acceleration from the acceleration assumed by the acceleration assumption unit based on the signal intensity Fouriertransformed by the Fourier transformation unit,
The detection unit is a detection device that detects the target based on the signal intensity Fouriertransformed by the Fourier transform unit with respect to the acceleration estimated by the acceleration estimation unit ,
For the acceleration assumed by the acceleration assumption unit, the acceleration estimation unit divides the signal intensity at the frequency at which the signal intensity is maximum by the average value of the signal intensity at a frequency other than the frequency at which the signal intensity is maximum. noise ratio, if a predetermined threshold value or more, detection device you characterized in that the acceleration, the acceleration of the target estimated.  A computer program that, when executed by a computer, causes the computer to function as the detection device according to any one of claims 1 to 4 .
 Radiates electromagnetic waves,
Receive the reflected wave reflected when the electromagnetic wave hits the target,
Assuming multiple accelerations of the above target,
For each assumed acceleration, calculate the signal compensation amount to compensate the phase of the reflected wave,
For each assumed acceleration, compensate the phase of the reflected wave with the signal compensation amount,
For each assumed acceleration, Fourier transform the signal intensity of the reflected wave with phase compensation,
Based on the signal strength obtained by Fourier transform, the target acceleration is estimated from the assumed acceleration,
About the estimated acceleration, a detection method for detecting the target based on Fouriertransformed signal intensity ,
For each hypothesized acceleration, the likelihood that the target acceleration is the hypothesized acceleration is calculated based on the Fourier transformed signal intensity,
A detection method, wherein a selection probability of an assumed acceleration is calculated based on the calculated likelihood, and a plurality of new target accelerations are randomly assumed based on the calculated selection probability .  Radiates electromagnetic waves,
Receive the reflected wave reflected when the electromagnetic wave hits the target,
Assuming multiple accelerations of the above target,
For each assumed acceleration, calculate the signal compensation amount to compensate the phase of the reflected wave,
For each assumed acceleration, compensate the phase of the reflected wave with the signal compensation amount,
For each assumed acceleration, Fourier transform the signal intensity of the reflected wave with phase compensation,
Based on the signal strength obtained by Fourier transform, the target acceleration is estimated from the assumed acceleration,
About the estimated acceleration, a detection method for detecting the target based on Fouriertransformed signal intensity ,
For the assumed acceleration, when the signaltonoise ratio obtained by dividing the signal strength at the frequency at which the signal strength is the maximum by the average value of the signal strength at a frequency other than the frequency at which the signal strength is the maximum is greater than or equal to a predetermined threshold In addition, the acceleration is the estimated acceleration of the target .
Priority Applications (1)
Application Number  Priority Date  Filing Date  Title 

JP2012220975A JP6053436B2 (en)  20121003  20121003  Detection apparatus, computer program, and detection method 
Applications Claiming Priority (1)
Application Number  Priority Date  Filing Date  Title 

JP2012220975A JP6053436B2 (en)  20121003  20121003  Detection apparatus, computer program, and detection method 
Publications (2)
Publication Number  Publication Date 

JP2014074602A JP2014074602A (en)  20140424 
JP6053436B2 true JP6053436B2 (en)  20161227 
Family
ID=50748847
Family Applications (1)
Application Number  Title  Priority Date  Filing Date 

JP2012220975A Active JP6053436B2 (en)  20121003  20121003  Detection apparatus, computer program, and detection method 
Country Status (1)
Country  Link 

JP (1)  JP6053436B2 (en) 
Families Citing this family (5)
Publication number  Priority date  Publication date  Assignee  Title 

US20170146638A1 (en) *  20140513  20170525  Mitsubishi Electric Corporation  Radar device 
JP2016017929A (en) *  20140711  20160201  三菱電機株式会社  Radar device 
JP2016017928A (en) *  20140711  20160201  三菱電機株式会社  Radar device 
JP5987204B1 (en) *  20150612  20160907  三菱電機株式会社  Radar equipment 
WO2018167873A1 (en) *  20170315  20180920  三菱電機株式会社  Radar device 
Family Cites Families (6)
Publication number  Priority date  Publication date  Assignee  Title 

JP3080589B2 (en) *  19960828  20000828  三菱電機株式会社  Radar equipment 
JP3645133B2 (en) *  19990902  20050511  三菱電機株式会社  Radar equipment 
US7212149B2 (en) *  20040617  20070501  The Boeing Company  System, method and computer program product for detecting and tracking a moving ground target having a single phase center antenna 
JP4963240B2 (en) *  20060612  20120627  三菱電機株式会社  Radar equipment 
JP5025403B2 (en) *  20070928  20120912  株式会社東芝  Target detection device 
JP5465148B2 (en) *  20101019  20140409  三菱電機株式会社  Radar equipment 

2012
 20121003 JP JP2012220975A patent/JP6053436B2/en active Active
Also Published As
Publication number  Publication date 

JP2014074602A (en)  20140424 
Similar Documents
Publication  Publication Date  Title 

JP6406601B2 (en)  Radar apparatus and object detection method  
CN106796282B (en)  Radar system and method for determining range, relative velocity and bearing of object using continuous wave and chirp signals  
Berizzi et al.  Autofocusing of inverse synthetic aperture radar images using contrast optimization  
US8681585B2 (en)  Multirange object location estimation  
JP5407272B2 (en)  Radar equipment  
JP5623338B2 (en)  Car radar system and method for determining a sweep pattern transmitted by an automobile radar as an output radar waveform  
Kershaw et al.  Optimal waveform selection for tracking systems  
US7626535B2 (en)  Track quality based multitarget tracker  
CN103477243B (en)  Radar installations  
CN102313884B (en)  Target trackbeforedetect (TBD) method based on multiframe coherent integration  
CN102073043B (en)  Multiframe phase coherence accumulation target trackingbeforedetecting method  
JP5554688B2 (en)  Radar equipment  
EP2581758B1 (en)  Methods for resolving radar ambiguities using multiple hypothesis tracking  
Yardim et al.  An overview of sequential Bayesian filtering in ocean acoustics  
Houssineau et al.  PHD filter with diffuse spatial prior on the birth process with applications to GMPHD filter  
JP6150474B2 (en)  Signal processing apparatus and signal processing method in wind profiler  
US9851444B2 (en)  Laser radar device and method of calculating speed of object to be measured  
CN103885057B (en)  Adaptive strain sliding window multiobject tracking method  
JP2009508199A (en)  Vehicle driving assistance method and improved related apparatus  
JP5503961B2 (en)  Observation signal processor  
Habtemariam et al.  Multiple detection probabilistic data association filter for multistatic target tracking  
WO2015095739A1 (en)  Single laser lidar system  
JP4220821B2 (en)  Sensor signal processing system  
CN104280728B (en)  Radar sensor and the method for running radar sensor  
Frehlich et al.  Coherent Doppler lidar measurements of winds in the weak signal regime 
Legal Events
Date  Code  Title  Description 

A621  Written request for application examination 
Free format text: JAPANESE INTERMEDIATE CODE: A621 Effective date: 20150911 

A977  Report on retrieval 
Free format text: JAPANESE INTERMEDIATE CODE: A971007 Effective date: 20160714 

A131  Notification of reasons for refusal 
Free format text: JAPANESE INTERMEDIATE CODE: A131 Effective date: 20160816 

A521  Written amendment 
Free format text: JAPANESE INTERMEDIATE CODE: A523 Effective date: 20160928 

TRDD  Decision of grant or rejection written  
A01  Written decision to grant a patent or to grant a registration (utility model) 
Free format text: JAPANESE INTERMEDIATE CODE: A01 Effective date: 20161101 

A61  First payment of annual fees (during grant procedure) 
Free format text: JAPANESE INTERMEDIATE CODE: A61 Effective date: 20161129 

R150  Certificate of patent or registration of utility model 
Ref document number: 6053436 Country of ref document: JP Free format text: JAPANESE INTERMEDIATE CODE: R150 

R250  Receipt of annual fees 
Free format text: JAPANESE INTERMEDIATE CODE: R250 