JP2012194039A - In-vehicle radar system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accurately determine, in an in-vehicle radar system, whether a detected reflection object is an object over which a present vehicle can travel or an object having such a height as to avoid contact.SOLUTION: An in-vehicle radar system comprises: a measuring unit 104 which measures information relating to a relative distance, a relative speed, a horizontal angle and a reflection wave strength in a fixed time cycle; a time-series correlation unit 106 which inputs the information measured by the measuring unit over a plurality of cycles and regards reflection objects correlated in time-series as the same object; a height-based reflection wave strength prediction unit 111 which estimates a plurality of kinds of heights as a height of a reflection object, predicts a reflection wave strength in a present signal processing cycle with a reflection wave strength of a past signal processing cycle as a reference for each of the plurality of kinds of estimated heights, and calculates a height-based reflection wave strength prediction value; and an object height estimation unit 112 which sequentially calculates a degree of matching for each height between the reflection wave strength actually measured by the measuring unit and the height-based reflection wave strength prediction value, and estimates the height of the reflection object.

Description

  The present invention transmits a radio wave in front of the host vehicle and detects a radio wave reflected from a reflecting object existing in front of the host vehicle, thereby detecting the relative distance, relative speed, horizontal direction between the host vehicle and the reflecting object. This is related to in-vehicle radar devices that measure angles, reflected wave intensity, etc., and in particular, it can calculate the height information of reflective objects and identify whether or not the vehicle is at a height that can travel across the vehicle. The present invention relates to an on-vehicle radar device.

  Conventionally, the measurement results of the relative distance, relative speed, horizontal angle, reflected wave intensity, etc., measured by the in-vehicle radar device, when the vehicle collides with an obstacle ahead. It is used in vehicle applications to improve vehicle safety and comfort, such as a collision damage reduction brake system that reduces damage and an adaptive cruise control system that follows the vehicle ahead.

Among in-vehicle radar devices used in such applications, there are radar devices that measure the horizontal angle of a reflecting object but do not measure the vertical angle.
Such an on-vehicle radar device does not require an antenna for measuring the angle in the vertical direction or a mechanism for switching the beam direction, so the configuration is simplified compared to a radar device that also measures the angle in the vertical direction. It has the feature that it can be done and the cost can be reduced.

  However, since the on-vehicle radar device as described above does not measure the vertical angle of the reflecting object, information on the height of the reflecting object cannot be obtained. As a result, low-position objects such as empty cans and cat's eyes that can travel across the vehicle cannot be distinguished from non-low-position objects such as vehicles that cannot travel across the vehicle. However, there is a problem that the above-mentioned application operates and may give the driver a sense of incongruity.

  There exists Unexamined-Japanese-Patent No. 2010-162975 (patent document 1) as a prior art with respect to such a subject. This prior art includes target object image detection means for detecting a target object image from a captured image including a part of the search space, a received signal level exceeding a reference level, and the target object image is other than a lower region of the search space. A control means for controlling the operation of the actuator of the vehicle based on the detected distance when detected from the captured image, and the reflected wave intensity of the object detected by the in-vehicle radar device is stronger than the threshold value Even so, the image sensor determines whether or not the size of the image of the reflective object is greater than or equal to the reference value, and determines whether or not the reflective object is to be controlled by the vehicle control application.

  Another conventional technique is Japanese Patent No. 3966673 (Patent Document 2). In this prior art, a stationary object is extracted from objects detected by an on-vehicle radar device, an object on the expected course of the vehicle is extracted from the object, and a variation in reflected wave intensity is detected from the first threshold value. The following objects are extracted, and when the decrease rate of the reflected wave intensity at a short distance is less than or equal to the second threshold value and the reflected wave intensity is less than or equal to the third threshold value, the object is identified as a low position object, For position objects, the operation of vehicle control applications such as a collision damage reduction brake system mounted on the vehicle is stopped.

JP 2010-162975 A Japanese Patent No. 3966673

In the method of Patent Document 1, the strength of the reflected wave intensity itself is used as an index for determination. However, since the reflected wave intensity measured by the in-vehicle radar device originally differs depending on the material and shape of the reflecting object, the reflected wave intensity is reflected. It is not desirable to use only the strength of the wave intensity itself as an index for determination.
For example, a reflective object made of metal such as a manhole cover or a cat's eye has a high reflected wave intensity among low-position objects, and may have a reflected wave intensity equivalent to a non-low-position object such as a vehicle. Among low-position objects, reflection objects such as people and motorcycles have a weak reflected wave intensity, and may have the same reflected wave intensity as a low-position object such as an empty can.

Further, the reflected wave intensity can be changed by the road surface multipath. Road surface multipath is a phenomenon that occurs due to interference between radio waves transmitted from a radar device and received by the radar device by directly hitting a reflecting object, and radio waves received by the radar device after being reflected once on the road surface. If the relative distance between the radar device and the reflecting object changes, the reflected wave intensity changes greatly.
In general, the higher the height of the reflective object, the more easily the radio wave transmitted by the radar device is reflected on the road surface and then received by the radar device. Therefore, the higher the height of the reflective object, the greater the influence of road surface multipath. As the height of the reflecting object is lower, the influence of the road surface multipath is smaller.
For this reason, the reflected wave intensity of a reflective object having a high height such as a vehicle is more likely to fluctuate greatly, and depending on the relative distance of the reflected object, the reflected wave intensity of a non-low-position object such as a vehicle is It may be weaker than the reflected wave intensity of a low-position object such as an empty can.
Furthermore, the reflected wave intensity can change due to individual differences among radar devices.

  Since the intensity of the reflected wave intensity itself can change as described above, the method of Patent Document 1 determines not only the intensity of the reflected wave intensity itself but also the size of the reflected object by combining with the image sensor, Although it is determined whether or not to be a control target of the vehicle control application, this method requires a complicated configuration in which the radar apparatus and the image sensor are combined.

  On the other hand, the method of Patent Document 2 focuses on the fact that the reflected wave intensity varies greatly due to the influence of the road surface multipath instead of the reflected wave intensity itself, and the reflected wave intensity varies greatly. The relative characteristics of the reflected wave intensity are used to identify whether the object is a low-position object or a non-low-position object. Although it is possible to identify whether it is a low-position object, in order to calculate the variation in reflected wave intensity, the measurement results of the past signal processing cycle are sufficiently accumulated for each reflected object detected by the radar device. Therefore, there is a problem that the storage area such as a memory is compressed.

  The present invention has been made to solve the above-described problems, and among the measurement results of the reflecting object obtained by the radar device alone, the current signal processing cycle and the most recent signal processing cycle up to the past several times. It is an object of the present invention to provide an on-vehicle radar device that identifies whether a reflecting object is a low-position object or a non-low-position object from the relationship of the reflected wave intensity.

  An in-vehicle radar device according to the present invention is mounted on a vehicle and outputs at least a relative distance, a relative speed, a horizontal angle, and a reflected wave intensity as information on a plurality of reflective objects existing in the vicinity. The measurement unit measures the relative distance, relative speed, horizontal angle and reflected wave intensity information of multiple reflective objects, and the relative distance, relative speed and horizontal angle of multiple reflective objects measured by the measurement unit. Assuming multiple types of height as the height of the reflective object and the time-series correlator that inputs the related information over multiple cycles and considers the reflective object correlated in time series as the same object The reflected wave of the current signal processing cycle is based on the reflected wave intensity of the past signal processing cycle for each of a plurality of assumed heights assumed for the same reflected object in the time series correlation unit. Predict strength The reflected wave intensity prediction unit for each height that calculates the reflected wave intensity prediction value for each height, and the actual reflected wave intensity measured by the measurement unit and the height calculated by the reflected wave intensity prediction unit for each height. An object height estimator that calculates the degree of coincidence at each height for each assumed height from the reflected wave intensity prediction value and estimates the height of the reflecting object from the degree of coincidence at each height. It is provided.

  Since the present invention uses not the reflected wave intensity itself but the temporal change of the reflected wave intensity as an index, metal low-position objects such as manhole covers and cat's eyes are non-low-position objects such as vehicles and guardrails. Although the reflected wave intensities may be equal, the radar apparatus alone can reliably identify these reflected objects as low-position objects without combining with an image sensor or the like as in the method described in Patent Document 1. be able to.

  Furthermore, according to the present invention, in each signal processing cycle, whether the object is a low-position object only from the reflected wave intensity of the signal processing cycle up to the last several times (about 5 times at the maximum) in the latest signal processing cycle. Since it is identified whether the object is a non-low-position object, the measurement result of the past signal processing period is sufficiently obtained for each reflective object detected by the radar device, compared with the method described in Patent Document 2. There is no need for storage, and a storage area such as a memory can be saved.

1 is a block diagram of an in-vehicle radar device according to Embodiment 1 of the present invention. It is a graph showing the relationship between the angle of a perpendicular direction, and the reflected wave intensity coefficient by a beam pattern in the vehicle-mounted radar apparatus of Embodiment 1 of this invention. It is a graph showing the relationship between the relative distance in the vehicle-mounted radar apparatus of Embodiment 1 of this invention, and the reflected wave intensity coefficient by a beam pattern. It is a flowchart showing the processing content of the object height estimation part in the vehicle-mounted radar apparatus of Embodiment 1 of this invention. It is a graph showing the relationship between the relative distance and reflected wave intensity in the vehicle-mounted radar apparatus of Embodiment 1 of this invention. 5 is a graph showing the relationship between the relative distance and the degree of coincidence for each height when the reflecting object is a low-position object in the in-vehicle radar device according to the first embodiment of the present invention. 5 is a graph showing the relationship between the relative distance and the degree of coincidence for each height when the reflecting object is a vehicle in the in-vehicle radar device according to the first embodiment of the present invention. It is a flowchart showing the whole processing content in the vehicle-mounted radar apparatus of Embodiment 1 of this invention.

Embodiment 1 FIG.
Hereinafter, an in-vehicle radar device according to Embodiment 1 of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram of an in-vehicle radar device according to an embodiment of the present invention.
In FIG. 1, an on-vehicle radar device 1 is mounted on a vehicle and is present in the vicinity by transmitting a radio wave in front of the host vehicle and detecting a radio wave reflected from a reflecting object existing in front of the host vehicle. At least a relative distance, a relative speed, a horizontal angle, and a reflected wave intensity are output as information on the plurality of reflecting objects. Information about the reflecting object is output as a signal to a vehicle control unit 3 such as a collision damage reducing brake system or an adaptive cruise control system that follows a vehicle ahead. Further, the on-vehicle radar device 1 receives a signal from the traveling speed sensor 2 that detects the traveling speed of the vehicle on which the radar device is mounted, and serves as additional information when detecting information on the reflecting object.

  The on-vehicle radar device 1 includes a control unit 101, a transmission / reception unit 102, a transmission / reception unit 103, a measurement unit 104, a storage unit 105, a time series correlation unit 106, a stationary object determination unit 107, a reflected wave intensity prediction unit 111 for each height, An object height estimation unit 112 is provided. Hereinafter, these components will be described in detail.

The control unit 101 is configured by, for example, a dedicated logic circuit, a program in a general-purpose CPU (Central Processing Unit), or a combination of both, and the operation timing of each component of the in-vehicle radar device 1 described below is determined. Control.
In the transmission / reception unit 102 and the transmission / reception unit 103, the transmission signal generated by the transmission / reception unit 102 is radiated from the transmission / reception unit 103 to the space in front of the vehicle as a transmission electromagnetic wave under the control of the control unit 101. The electromagnetic wave reflected by the object or the like is received by the transmission / reception unit 103, and the transmission / reception unit 102 converts it into a reception signal.

The measurement unit 104 in which the input timing of the received signal and the output timing of the measurement result are controlled by the control unit 101 is configured by a dedicated logic circuit, a general-purpose CPU, a program in a DSP (Digital Signal Processor), or a combination thereof. , Perform measurement signal processing corresponding to the radar method and angle measurement method to be used, and at least for each reflective object,
Polar coordinate system relative distance: Dst
Polar coordinate system radial relative speed: Vlc
Polar coordinate system horizontal angle: Ang
Reflected wave intensity: Amp
Measure.

These measurement results are output to the storage unit 105 under the control of the control unit 101, and the storage unit 105 stores them in a predetermined storage area under the control of the control unit 101.
In order to measure the relative distance Dst, the relative velocity Vlc, and the reflected wave intensity Amp, the transmission / reception unit 102 is configured to realize an FMCW (Frequency Modulated Continuous Wave) method, a pulse Doppler method, and the like, which are known as radar methods. The transmission / reception timing is controlled by the control unit 101.

  Further, in order to measure the horizontal angle Ang, the wave transmitting / receiving unit 103 has a mechanism for changing the direction of transmitted / received electromagnetic waves in the azimuth direction of the polar coordinate system for a known monopulse angle measuring method, or a known array signal processing angle measuring method. For this purpose, a plurality of transmission / reception elements are provided, and the control unit 101 controls the direction of transmission / reception electromagnetic waves, the timing of transmission / reception of waves by the plurality of elements, and the like.

  The time series correlator 106 receives information related to the relative distances, relative velocities, and horizontal angles of a plurality of reflecting objects from the storage unit 105 after the measurement unit 104 finishes outputting to the storage unit 105 under the control of the control unit 101. Are input over a plurality of periods, and reflection objects correlated in time series are regarded as the same object and identified. That is, what is considered to be the same reflection object in the previous signal processing cycle and the current signal processing cycle is identified. This identification is performed every signal processing cycle (for example, every 100 msec), and the objects that are considered to be the same reflecting object are continuously identified.

Specifically, from the measurement result of each reflecting object in the current signal processing cycle of the measurement unit 104, a predicted value that predicts the polar coordinate system relative distance, polar coordinate system relative speed, horizontal angle, etc. of the next signal processing cycle is calculated. In the next signal processing cycle, the actual measurement result of each reflecting object of the measuring unit 104 is compared with the predicted value, and the reflecting object closest to the predicted value is determined to be the same reflecting object. Those that are determined to be objects are assigned the same reflective object number: ID in each signal processing cycle. The IDs identified as the same reflecting object are stored in the storage unit 105.
The processing performed by the time series correlator 106 may be another method as long as it can identify the same reflecting object in each signal processing cycle.

In the stationary object determination unit 107, after the processing of the time series correlation unit 106 is finished under the control of the control unit 101, the traveling speed of the radar-equipped vehicle obtained by the traveling speed sensor 2: Vself is measured by the measurement unit 104. Whether the reflecting object is a moving object or a stationary object is determined from the relative velocity Vlc or the like. Specifically, the ground speed: Vearth of the reflecting object is calculated from the traveling speed Vself and the relative speed Vlc, and an object whose absolute value of the ground speed Vearth is equal to or less than a threshold is extracted as a stationary object.
After finishing the processing of the stationary object determination unit 107, the control unit 101 sequentially controls the reflected wave intensity prediction unit 111 for each height and the object height estimation unit 112 to estimate the height of the reflection object.

  In the reflected wave intensity prediction unit 111 for each of the reflected objects assigned with the same ID, the reflected wave intensity of the reflected object extracted as a stationary object is inversely proportional to the square of the relative distance to the reflected object, and Reflected wave observed in current signal processing cycle from reflected wave intensity obtained in past signal processing cycle, assuming only multiple types of reflection object heights, using the proportionality to the beam pattern of radar device A reflected wave intensity prediction value for each height in which the intensity is predicted for each assumed height is calculated.

First, the operation principle of the reflected wave intensity prediction unit 111 for each height will be described.
Assume a radar apparatus in which the reflected wave intensity at other vertical angles with the vertical angle 0 deg as a reference is represented by the beam pattern of FIG. In addition, the intensity of another angle with respect to the vertical angle of 0 deg is referred to as a reflected wave intensity coefficient by the vertical beam pattern.

The angle in the vertical direction between the reflective object and the radar apparatus varies depending on the mounting height of the radar apparatus, the height of the reflective object, and the relative distance between the reflective object. When the mounting height of the radar apparatus is fixed, the reflected wave intensity coefficient due to the beam pattern changes depending on the relative distance between the height of the reflecting object and the reflecting object.
For example, assuming that the mounting height of the radar device is 600 mm assuming the vicinity of the front grill of the passenger car, and the assumed height of the reflecting object is 0 mm, 100 mm, and 200 mm, the horizontal axis represents the relative distance of the reflecting object, and the vertical axis represents When expressed as a reflected wave intensity coefficient by a beam pattern, it is as shown in FIG.

As apparent from FIG. 3, the reflected wave intensity coefficient varies depending on the beam pattern in the vertical direction relative to the relative distance depending on the height of the reflecting object. Different reflected wave intensity prediction values are obtained for each height.
The processing outline of the reflected wave intensity prediction unit 111 for each height will be described for the case where the assumed height is three types of 0 mm, 100 mm, and 200 mm.

In the reflected wave intensity prediction unit 111 for each height, among the reflected objects measured by the measuring unit 104 and output to the storage unit 105, reflected objects assigned the same ID in the previous signal processing cycle and the current signal processing cycle are displayed. For reflective objects extracted and further extracted as stationary objects,
Measured reflected wave intensity in the previous signal processing cycle: Amp_pre,
Measurement relative distance in the signal processing cycle: Dst_now,
Measurement relative distance in the previous signal processing cycle: From Dst_pre
Predicted value of reflected wave intensity at each height when the assumed height is 0 mm: Amp_est0mm
Predicted value of reflected wave intensity at each height when the assumed height is 100 mm: Amp_est100mm
Predicted value of reflected wave intensity per height when the assumed height is 200 mm: Amp_est200mm
The three kinds of reflected wave intensity prediction values for each height are calculated.

Assuming that the influence of noise and other factors is sufficiently small, the predicted reflected wave intensity at each height is calculated using the following formula because the reflected wave intensity is inversely proportional to the square of the relative distance and proportional to the beam pattern in the vertical direction. expressed.
Amp_est0mm =
Amp_pre / (Beam0mm [Dst_pre] / Dst_pre ^ 2) * (Beam0mm [Dst_now] / Dst_now ^ 2)
Amp_est100mm =
Amp_pre / (Beam100mm [Dst_pre] / Dst_pre ^ 2) * (Beam100mm [Dst_now] / Dst_now ^ 2)
Amp_est200mm =
Amp_pre / (Beam200mm [Dst_pre] / Dst_pre ^ 2) * (Beam200mm [Dst_now] / Dst_now ^ 2)

Here, Beam0mm [Dst_pre], Beam100mm [Dst_pre], Beam200mm [Dst_pre], Beam0mm [Dst_now], Beam100mm [Dst_now], and Beam200mm [Dst_now] are the relative distances Dst_pre, This is a reflected wave intensity coefficient according to a vertical beam pattern in Dst_now, and this value corresponds to the value of the graph shown in FIG.
This value may be stored in the radar device as a table of the reflected wave intensity coefficient by the beam pattern in the vertical direction with respect to the relative distance for each assumed height, or in the vertical direction with respect to the vertical angle shown in FIG. By approximating the reflected wave intensity coefficient due to the beam pattern with a quadratic equation, the approximate expression is held in the radar device, and the mounting height of the radar device, the relative distance of the reflecting object, and the reflection at each signal processing cycle The angle in the vertical direction of the reflecting object may be calculated from the assumed height of the object, and the reflected wave intensity coefficient based on the vertical beam pattern may be calculated from the approximate expression.

Here, the reflected wave intensity prediction value for each height was calculated using the fact that the reflected wave intensity is inversely proportional to the square of the relative distance and proportional to the vertical beam pattern of the radar device. Thus, the fact that the reflected wave intensity is proportional to the horizontal beam pattern may be utilized by using the horizontal angle measured by the radar device.
Thereby, even if the beam width in the horizontal direction is narrow, the reflected wave intensity prediction value for each height can be calculated with high accuracy.

  Next, in the object height estimation unit 112, the reflected wave intensity actually measured by the measurement unit 104 and the reflected wave intensity prediction value for each height predicted by the reflected wave intensity prediction unit 111 for each height are used for each time. The degree of coincidence for each height is calculated for each height assumed in the signal processing cycle, and the most likely reflective object height is estimated from the degree of coincidence for each height.

First, the operation principle of the object height estimation unit 112 will be described.
The reflected wave intensity is a) a phenomenon proportional to the reflected wave intensity coefficient due to the beam pattern in the vertical direction, b) a phenomenon inversely proportional to the square of the relative distance of the reflecting object, c) an influence of noise, d) an influence due to road multipath, It can be obtained as a composite.
On the other hand, the predicted value of reflected wave intensity at each height takes into account only a) a phenomenon proportional to the reflected wave intensity coefficient by the vertical beam pattern and b) a phenomenon inversely proportional to the square of the relative distance of the reflecting object. Yes.
For this reason, an error occurs between the actual reflected wave intensity measured by the measurement unit 104 and the predicted reflected wave intensity value for each height predicted by the reflected wave intensity prediction unit 111 for each height.

The closer the actual height and the assumed height of the reflecting object are, the more the reflected wave intensity actually measured by the measuring unit 104 and the reflected wave intensity prediction by the height predicted by the reflected wave intensity predicting unit 111 for each height. The error that occurs between the values is small. However, the above-mentioned error is a combination of the influence of noise and the influence of road multipath, and the error is too large as it is, and it is difficult to estimate the height of the reflecting object.
Therefore, in the present invention, the accumulated error obtained by accumulating the error in each signal processing cycle is calculated for each assumed height, and the actual height of the reflecting object is calculated from the relationship between the accumulated errors for each assumed height obtained as a result. Is estimated.

The processing outline of the reflected wave intensity prediction unit 111 for each height will be described for the case where the assumed height is three types of 0 mm, 100 mm, and 200 mm.
First, an error between the reflected wave intensity actually measured by the measuring unit 104 and the predicted reflected wave intensity value for each height predicted by the reflected wave intensity predicting unit 111 for each height is calculated by the following formula.
ΔAmp0mm = Amp_est0mm-Amp
ΔAmp100mm ＝ Amp_est100mm-Amp
ΔAmp200mm ＝ Amp_est200mm-Amp
This error is defined as an instantaneous error at every height.

Next, calculation of instantaneous error for each height is performed in each signal processing cycle, and is an accumulated value of instantaneous error for each height in each signal processing cycle for each reflected wave intensity prediction value,
Accumulated errors Σ (ΔAmp0mm), Σ (ΔAmp100mm), and Σ (ΔAmp200mm) for each height are calculated to smooth the influence of noise and the like.

The cumulative error for each height is closer to zero as the actual height of the reflecting object is closer to the assumed height.
For this reason, the cumulative error for each height is squared in each signal processing cycle and the sum is calculated. The closer the actual height and the assumed height of the reflecting object are, the closer the value is to zero.
In this embodiment, the degree of coincidence for each height is defined by squaring the accumulated error for each height in each signal processing cycle and calculating the sum thereof.

That is, the degree of coincidence for each height is expressed by the following formula.
Agreement 0mm ＝ Σ ((Σ (ΔAmp0mm)) ^ 2)
100mm coincidence = Σ ((Σ (ΔAmp100mm)) ^ 2)
Matching degree 200mm ＝ Σ ((Σ (ΔAmp200mm)) ^ 2)

Note that the effect of noise increases as the signal-to-noise power ratio decreases, such as when the relative distance of the reflecting object is long. Therefore, for the purpose of reducing the influence of noise included in the degree of coincidence at each height, the signal pair of the reflecting object is reduced. A configuration may be adopted in which calculation of coincidence for each height (estimation of height) is started at a short distance from a relative distance at which the noise power ratio (S / N ratio) is equal to or greater than a predetermined SN ratio threshold T01. Conversely, the calculation of the degree of coincidence for each height (estimation of the height) is not performed at a distance farther than the relative distance that is equal to or greater than the SN ratio threshold T01.
In addition to calculating the degree of coincidence at each height (estimating the height), the reflected wave intensity at each height is calculated from the relative distance at which the signal-to-noise power ratio (S / N ratio) is equal to or greater than a predetermined SN ratio threshold T01. The prediction unit 111 may start calculating the reflected wave intensity predicted value for each height, and may not calculate the predicted reflected wave intensity value for each height at a relative distance farther than the SN ratio threshold T01.

As is clear from FIG. 3, the longer the relative distance is, the smaller the difference in the reflected wave intensity coefficient due to the beam pattern at each assumed height, so that the relative distance from the reflecting object is reduced for the purpose of reducing the amount of calculation. A configuration may be adopted in which calculation of the degree of coincidence for each height (estimation of height) is started from the relative distance where the distance is equal to or less than a predetermined threshold. Conversely, the calculation of the degree of coincidence for each height (estimation of the height) is not performed at a relative distance farther than the threshold value.
In addition to calculating the degree of coincidence for each height (estimating the height), the reflected wave intensity prediction value for each height by the reflected wave intensity prediction unit 111 for each height from the relative distance at which the relative distance is equal to or less than a predetermined threshold value. The calculation of the reflected wave intensity predicted value for each height may not be performed at a relative distance farther than the threshold value.

As described above, the degree of coincidence at each height derived in this way approaches zero as the actual height of the reflective object is closer to the assumed height, and farther away from the actual height of the reflective object and the assumed height. The larger the value.
In other words,
When the degree of coincidence is the smallest, it is around the actual height of the reflective object,
When the matching degree is 100 mm, the actual height of the reflecting object is around 100 mm.
When the degree of coincidence is 200 mm, the actual height of the reflective object is around 200 mm or more than 200 mm.
It is possible to estimate the actual height of the reflective object by estimating the closest height to any one of the assumed heights among the plurality of types of assumed heights.

Conversely, when the degree of coincidence 0 mm is larger than the degree of coincidence at other heights, it can be estimated that the actual height of the reflecting object is not at least near 0 mm. That is, it can be estimated that it is not at least one assumed height among a plurality of assumed heights.
Here, the assumed height is three types of 0 mm, 100 mm, and 200 mm. However, in the present invention, the number of the assumed height types is not limited to these three types, and the interval between the assumed heights is set. It is not limited to 100 mm intervals.
For example, if the assumed height interval is narrowed and the number of assumed height types is increased, the estimation accuracy of the height can be improved accordingly.

On the other hand, the greater the number of types of assumed heights, the greater the amount of computation. In order to reduce the amount of calculation, it is necessary to reduce the number of types of assumed heights. However, if the number of types of assumed heights is simply reduced, the accuracy of height estimation is deteriorated.
Therefore, when the number of types of assumed heights is reduced to some extent, it is desirable to apply the following three ideas so as to ensure the height estimation accuracy.

The first is a method of devising the conditions for estimating the height so as to use the relationship with the degree of coincidence at other heights.
For example, when the degree of coincidence 100 mm is smaller than the degree of coincidence 0 mm, the degree of coincidence 0 mm is smaller than a predetermined threshold, and the degree of coincidence 100 mm is larger than the predetermined threshold, the actual height is around 0 mm to 25 mm. Thus, even if the number of types of assumed heights is small, the accuracy of height estimation can be improved by devising conditions for estimating the height.
Thus, it can also be estimated that it exists between two assumed heights among multiple types of assumed heights.

The second is a method of setting the assumed heights to be prepared, for example, at a height of 0 mm, 100 mm, 110 mm, 120 mm, and 200 mm, rather than at equal intervals.
In this method, the estimated height interval is narrowed in the height range where the height estimation accuracy is desired to be improved, and the estimated height interval is set wider in the other height ranges. While suppressing the increase, it is possible to improve the estimation accuracy of the height of the reflecting object only in a portion where the assumed height interval is narrowed.
This method is effective when, for example, it is desired to improve the estimation accuracy of the height in the vicinity of the height boundary for identifying whether the object is a low position object or a non-low position object.

Third, as is clear from FIG. 3, the farther the relative distance from the reflecting object is, the smaller the difference in the reflected wave intensity coefficient due to the beam pattern at each assumed height is. This is a method of decreasing the number of types of assumed heights to be prepared as the distance is longer, and increasing the number of types of assumed heights to be prepared as the relative distance to the reflecting object is closer.
By this method, the amount of calculation can be reduced at a long distance, and the accuracy of height estimation can be improved at a short distance. Therefore, the total amount of calculation of the reflected wave intensity prediction unit 111 for each height and the object height estimation unit 112 can be reduced. However, it is possible to obtain high height estimation accuracy at a short distance.
All of these three methods do not necessarily need to be used, and may be used in appropriate combination as necessary.

Since the vertical beam pattern shown in FIG. 2 is a vertical target, for example, when the mounting height of the radar apparatus is 600 mm and the height of the reflecting object is 200 mm, the mounting height of the radar apparatus is used as a reference. Upside down, the coincidence degree 200 mm and the coincidence degree 1000 mm are completely the same value.
Similarly, when the height of the reflecting object is 1000 mm, the coincidence degree is 200 mm, and the coincidence degree 1000 mm is completely the same value.
In this embodiment, it is impossible to identify whether the height of the reflecting object is 200 mm or 1000 mm. Therefore, in this embodiment, the vertical object countermeasure process is performed.

In general, an object having a higher height is more affected by the road surface multipath, and therefore, there is a relationship in which an object having a higher degree of coincidence at a height also has a higher height.
Therefore, in the vertical target countermeasure processing, when the degree of coincidence at each height is equal to or higher than a predetermined vertical target countermeasure threshold, the actual height of the reflecting object is estimated to be higher than the mounting height of the radar device, and the degree of coincidence at each height is If it is less than the vertical target countermeasure threshold, it is determined that the actual height of the reflecting object is lower than the mounting height of the radar apparatus.
In particular,
If the degree of coincidence is 200 mm = the degree of coincidence is 1000 mm <the upper and lower target countermeasure threshold, it is determined that the height of the reflecting object is 200 mm, which is lower than 600 mm, which is the mounting height of the radar device,
If the degree of coincidence is 200 mm = the degree of coincidence is 1000 mm ≧ the upper / lower target countermeasure threshold, it is determined that the height of the reflecting object is 1000 mm, which is higher than 600 mm which is the mounting height of the radar apparatus.

However, in the present embodiment, since the degree of coincidence for each height is a value that increases as the relative distance approaches, the vertical target countermeasure threshold value is set to increase as the relative distance approaches.
In the present embodiment, when calculating the cumulative error for each height, the degree of coincidence for each height increases depending on how many times the instantaneous error for each signal processing cycle has been added. The countermeasure threshold must be set according to the number of times.

As described above, the method for estimating the height of the object from the degree of coincidence at each height has been described. However, in order to estimate the height of the object from the degree of coincidence at each height, several conditions shown below are necessary.
Hereinafter, the conditions until the height of the object is estimated from the degree of coincidence for each height will be described using the flowchart of FIG.
First, in step S201, the process proceeds to step S202 only when the number of instantaneous errors added for each height is equal to or greater than a predetermined error addition number threshold. If the number of error additions is small, the influence of errors is large, and sufficient height estimation accuracy cannot be obtained. Step S201 is a necessary condition for ensuring the height estimation accuracy.

In step S202, the process proceeds to step S203 only when the moving distance of the own vehicle moves more than a predetermined own vehicle moving distance threshold from the signal processing cycle calculated for the first time with respect to the instantaneous error for a certain ID. If the distance traveled by the host vehicle is not sufficient, the change in the reflected wave intensity with respect to the relative distance cannot be measured sufficiently and there is a risk of erroneous determination. Therefore, step S202 is a determination condition necessary to prevent such a determination error. It is.
Note that the distance traveled by the vehicle is calculated based on the traveling speed of the radar-equipped vehicle obtained by the traveling speed sensor 2 and the elapsed time from the signal processing cycle calculated for the first time for each height.

  In step S203, the process proceeds to step 204 only when the relative distance of the reflecting object is equal to or less than a predetermined relative distance threshold (height estimation distance threshold) when starting the height estimation. As is apparent from FIG. 3, when the relative distance is far and the difference in coincidence at each height is too small, the coincidence at each height becomes almost the same value regardless of the assumed height. This is because it is difficult to estimate the actual height of the reflecting object from the degree of coincidence. Therefore, height estimation is started when the relative distance of the reflecting object is equal to or smaller than a predetermined threshold.

In addition, the predetermined relative distance threshold (height estimation distance threshold) for starting the height estimation in step S203 is larger as the influence of noise is larger and the influence of road surface multipath is larger. It is desirable to set to. This is because if the influence of noise or the influence of road surface multipath is great, the judgment accuracy of the coincidence at each height deteriorates, and it is difficult to estimate the actual height of the reflecting object from the coincidence at every height. It is.
In this embodiment, it is considered that the greater the influence due to noise and the greater the influence due to road multipath, the greater the coincidence at each height. The estimation distance threshold is set to be small (the relative distance at the start of height estimation is set to a short distance).

Although not described in the flowchart of FIG. 4, since the signal-to-noise power ratio (S / N ratio) of the reflecting object is small if the influence of noise is large, the signal-to-noise power ratio of the reflecting object is the SN ratio threshold value. Preventing the process from proceeding to step S204 in the case of T02 or less is also an effective means for improving the estimation accuracy of the height of the reflecting object.
Here, the SN ratio threshold T02 is a threshold set to be equal to or higher than the above-described SN ratio threshold T01.

  Finally, in step 204, the height of the reflecting object is estimated from the degree of coincidence for each height by the method described above. In step 204, even if the height of the reflective object is estimated to be around 100 mm at a relative distance equal to or less than the height estimation distance threshold in a certain signal processing cycle, it is reestimated to be around 0 mm in the next signal processing cycle, It may be reestimated to be around 200 mm or 200 mm or more.

  In this embodiment, when the relative distance of the reflective object is far from the height estimation distance threshold, the signal-to-noise power ratio of the reflective object is equal to or smaller than the SN ratio threshold T02, and the relative distance of the reflective object is If the estimated height of the reflective object fluctuates greatly even though it is closer than the height estimation distance threshold, it is determined that the influence of noise or the influence of the road surface multipath is large. Since the height cannot be estimated, the estimated value of the height is not calculated.

In addition, various thresholds such as the up / down target countermeasure threshold value, the error addition frequency threshold value, the own vehicle moving distance threshold value, the height estimation distance threshold value, the SN ratio threshold value T01, and the SN ratio threshold value T02 mentioned in this embodiment are This value is set as appropriate based on the detection performance.
Specifically, an example of processing when the traveling speed of the radar-equipped vehicle is 36 km / h (10 m / sec) and the signal processing cycle is 100 msec is shown.
FIG. 5 shows the change in reflected wave intensity with respect to the relative distance in the case where the height of the radar device is 600 mm and the height of an empty can is assumed and the case where a high object such as a vehicle is assumed. Show.
In this processing example, it is assumed that the empty can is a reflective object having a height of 100 mm and is not affected by the road surface multipath, and the vehicle is a reflective object having a height of 600 mm and is affected by the road surface multipath. A simple model that ignores the spread of reflection points and changes in reflection points will be explained.

In the present invention, since the actual height of the reflecting object is estimated only from the difference in relative reflected wave intensity between the previous signal processing period and the current signal processing period, in FIG. 5, the reflected wave intensity at a relative distance of 50 m is used. Is normalized as 1.
In this processing example, the assumed heights for determining the degree of coincidence for each height are seven types of 0 mm, 100 mm, 200 mm, 300 mm, 400 mm, 500 mm, and 600 mm.

FIG. 6 shows a change in the degree of coincidence for each height when the target reflective object is an empty can.
In this processing example, the error addition frequency threshold is 10 times, the own vehicle movement distance threshold is 10 m, and the height estimation distance threshold is 20 m.
FIG. 6 shows that when the target reflective object is an empty can, the degree of coincidence 100 mm is the smallest in the region where the relative distance of the reflective object is 20 m or less. As a result, the height of the reflecting object is estimated to be around 100 mm.

On the other hand, FIG. 7 shows a change in the degree of coincidence for each height with respect to the relative distance when the target reflective object is a vehicle.
In this processing example, the height estimation distance threshold is 10 m, and the vertical target countermeasure threshold at a distance of 10 m is 100.
In FIG. 7, since the minimum value of the degree of coincidence at each height is larger than the upper and lower target countermeasure thresholds, the degree of coincidence at a low height such as the degree of coincidence of 0 mm and the degree of coincidence of 100 mm is the degree of coincidence of 1200 mm, the degree of coincidence of 1100 mm, etc. The height of the reflecting object is estimated by substituting the degree of coincidence of the assumed heights upside down with respect to the mounting height of the radar device.

FIG. 7 shows that when the target reflective object is a vehicle, the degree of coincidence 600 mm is the smallest at a relative distance of 10 m between the reflective objects. As a result, the height of the reflecting object is estimated to be around 600 mm.
In this processing example, the matching degree 600 mm is the smallest, but the relation with the matching degree 500 mm may be reversed depending on the influence of noise or the like.
When the difference between the coincidences at each height is small as described above, it is estimated that the height of the reflective object is not at least 0 mm to 300 mm but not 900 mm to 1200 mm by utilizing the fact that the coincidence at other heights is large. However, the height of the reflecting object is estimated to be between 400 mm and 800 mm, so that the height of the reflecting object is given a certain margin from the relationship with the degree of coincidence at other heights. Also good.

  Finally, the vehicle control unit 3 identifies whether the detected reflecting object is a low position object or a non-low position object based on the estimated actual height information of the reflecting object, and the reflecting object is a low position object. When it is identified that the vehicle is in contact with the object, an alarm for avoiding contact with the object and a process for stopping the automatic deceleration are performed.

The flowchart of FIG. 8 summarizes the embodiments of the present invention so far.
In this flowchart, first, in step S301, the reflection object in the current signal processing cycle is measured by the processing of the transmission / reception unit 102, the transmission / reception unit 103, and the measurement unit 104.
Next, in step S302, the measurement result of the reflecting object in the previous signal processing cycle and the current signal processing cycle is compared by the processing of the time series correlation unit 106, and it is considered that there is a correlation in time series (same in the previous time and this time). The same ID is assigned to the reflected object.

Next, in step S303, it is determined by the processing of the stationary object determination unit 107 whether or not the reflective object is a stationary object.
Next, when it is determined in step S304 that the reflecting object is a stationary object, in step S305, a plurality of heights of the reflecting object are assumed by the processing of the reflected wave intensity prediction unit 111 for each height, and the previous time and the current time. From the reflected wave intensity in the previous signal processing cycle of the reflective object with the same ID, the reflected wave intensity prediction value for each height in the current signal processing cycle is calculated.

Next, in step S306, the reflected wave intensity actually obtained by the measurement unit 104 in the current signal processing cycle and the high predicted by the reflected wave intensity prediction unit 111 for each height are processed by the object height estimation unit 112. The reflected wave intensity predicted values for each height are compared to calculate the degree of coincidence for each height, and the height of the reflecting object is estimated from the degree of coincidence for each height.
Finally, when it is determined in step S307 that the reflective object is a low-position object from the estimated height of the reflective object, in step S308, the vehicle control unit 3 avoids the vehicle from contacting the reflective object. For example, a process for stopping the operation of the vehicle control application is executed.

As mentioned above, although embodiment of this invention was described, this invention can add a various design change in the range which does not deviate from the summary.
For example, in the above-described embodiment, the mounting height of the radar apparatus is 600 mm. However, the present invention does not limit the mounting height of the radar apparatus to this range, and the mounting height of the radar apparatus is known. Other heights may be used.

In the above-described embodiment, the radar apparatus having the beam pattern shown in FIG. 2 is targeted. However, the present invention can be applied to any beam pattern as long as the beam pattern can be stored in the radar apparatus in advance. In particular, when the beam pattern has a vertically asymmetric shape, it is not necessary to carry out the vertical object countermeasure process shown in the above embodiment.
In the embodiment described above, the sum of squares of the accumulated error per height obtained by accumulating the error between the measured actual reflected wave intensity and the reflected wave intensity predicted value per height is defined as the degree of coincidence per height. If the method can determine whether the measured actual reflected wave intensity and the reflected wave intensity predicted value per height are similar, for example, the measured actual reflected wave intensity and the reflected wave intensity predicted value per height Other indices such as the sum of squares or the mean square of instantaneous errors at each height, which are errors between the two, may be defined as the degree of coincidence at each height.

  In the above-described embodiment, the predicted reflected wave intensity for each height of the current signal processing cycle is calculated from the reflected wave intensity actually measured in the previous signal processing cycle. The predicted value of reflected wave intensity may be calculated for each height, or from the measurement results up to the past several times (preferably 2 to 4 times, up to about 5 times) before the storage area such as the memory is compressed. A reflected wave intensity prediction value may be calculated.

1: vehicle-mounted radar device, 2: traveling speed sensor,
3: Vehicle control unit,
101: Control unit, 102: Transmission / reception unit,
103: Transmission / reception unit, 104: Measurement unit,
105: Storage unit, 106: Time series correlation unit,
107: Stationary object determination unit 111: Reflected wave intensity prediction unit for each height,
112: Object height estimation unit.

Claims (16)

In a radar device that is mounted on a vehicle and outputs at least a relative distance, a relative speed, a horizontal angle, and a reflected wave intensity as information on a plurality of reflective objects existing in the vicinity,
A measurement unit that measures information related to the relative distance, relative speed, horizontal angle, and reflected wave intensity of the plurality of reflective objects at a constant time period;
Information related to the relative distance, relative speed, and horizontal angle of multiple reflective objects measured by the measurement unit is input over multiple periods, and reflective objects correlated in time series are regarded as the same object. A time series correlator,
Assuming multiple types of height as the height of the reflective object, the past signal processing cycle for each of the multiple types of assumed heights for the object considered as the same reflective object in the time series correlation unit The reflected wave intensity prediction unit for calculating the reflected wave intensity prediction value for each height by predicting the reflected wave intensity of the current signal processing cycle with reference to the reflected wave intensity of
From the actual reflected wave intensity measured by the measurement unit and the reflected wave intensity prediction value for each height calculated by the reflected wave intensity prediction unit for each height, for each estimated height at each signal processing cycle An in-vehicle radar device comprising: an object height estimating unit that calculates a degree of coincidence and estimates the height of the reflecting object from the degree of coincidence for each height.   A stationary object judging means for judging whether the reflecting object is a moving object or a stationary object from a ground speed calculated from a traveling speed of a vehicle equipped with a radar device and a relative speed of the reflecting object; The in-vehicle radar device according to claim 1, wherein the height of the reflecting object is estimated only when the determination unit determines that the object is a stationary object.   The reflected wave intensity prediction value calculated by the reflected wave intensity prediction unit for each height is inversely proportional to the square of the relative distance to the reflecting object and proportional to the beam pattern in the vertical direction of the radar apparatus. The in-vehicle radar device according to claim 1, wherein the on-vehicle radar device is calculated using the above.   The reflected wave intensity prediction value calculated by the above reflected wave intensity prediction unit for each height is inversely proportional to the square of the relative distance to the reflecting object, and the vertical beam pattern and the horizontal direction of the radar apparatus. The in-vehicle radar device according to claim 1, wherein the on-vehicle radar device is calculated using the fact that it is proportional to the beam pattern.   The degree of coincidence for each height calculated by the object height estimation unit is calculated based on the actual reflected wave intensity measured by the measurement unit and the reflected wave intensity prediction for each height calculated by the reflected wave intensity prediction unit for each height. The error between each value is defined as the instantaneous error at each height, and this instantaneous error at each height is calculated at each signal processing cycle to obtain the accumulated error at each height. The on-vehicle radar device according to claim 1, wherein the vehicle-mounted radar device is obtained by squaring at a processing cycle and calculating a sum thereof.   The height of the reflective object estimated by the object height estimation unit is estimated to be the closest to any one of the plurality of types of assumed heights. The on-vehicle radar device according to any one of 1 to 5.   The height of the reflection object estimated by the object height estimation unit is estimated to be between two assumed heights among the plurality of assumed heights. The in-vehicle radar device according to any one of?   The height of the reflective object estimated by the object height estimation unit is estimated to be not at least one assumed height among the plurality of types of assumed heights. The on-vehicle radar device according to any one of 1 to 5.   The height assumed by the reflected wave intensity prediction unit for each height increases the number of types of assumed heights as the relative distance to the reflecting object is closer. The on-vehicle radar device according to the item.   The object height estimation unit starts height estimation when the relative distance to the reflective object is equal to or less than a predetermined threshold, and does not perform height estimation at a relative distance farther than the threshold. Item 10. The on-vehicle radar device according to any one of Items 1 to 9.   The reflected wave intensity prediction unit for each height and the object height estimation unit start calculating the reflected wave intensity prediction value for each height and the matching degree for each height when the relative distance from the reflecting object is equal to or less than a predetermined threshold. 11. The on-vehicle radar device according to claim 10, wherein calculation of the reflected wave intensity predicted value per height and the matching degree per height is not performed at a relative distance farther than the threshold value.   The object height estimation unit starts estimating the height at a short distance from the relative distance where the signal-to-noise power ratio from the reflective object is equal to or greater than a predetermined threshold, and estimates the height at a distance farther than the relative distance. The on-vehicle radar device according to claim 1, wherein the on-vehicle radar device is not implemented.   The reflected wave intensity prediction unit for each height and the object height estimation unit are configured to calculate the reflected wave intensity prediction value for each height at a short distance from the relative distance at which the signal-to-noise power ratio from the reflected object is equal to or greater than a predetermined threshold. 13. The vehicle-mounted radar device according to claim 12, wherein calculation of the degree of coincidence is started, and the reflected wave intensity predicted value for each height and the degree of coincidence for each height are not calculated at a distance farther than the relative distance.   The object height estimation unit calculates the degree of coincidence at each height so that the degree of coincidence at each height decreases as the assumed height and the actual height of the reflecting object are closer, and the higher the degree of coincidence at each height, the higher the height. The on-vehicle radar device according to claim 1, wherein the relative distance of the reflecting object when estimating the distance is a short distance.   The object height estimation unit calculates the degree of coincidence for each height so that the degree of coincidence for each height decreases as the assumed height and the actual height of the reflecting object are closer, and the degree of coincidence for each height is equal to or greater than a predetermined threshold. In this case, it is estimated that the actual height of the reflecting object is higher than the mounting height of the radar device, and when the coincidence for each height is less than the above threshold, the actual height of the reflecting object is lower than the mounting height of the radar device. The on-vehicle radar device according to claim 1, wherein the on-vehicle radar device is estimated.   16. The on-vehicle radar device according to claim 1, wherein a vertical beam pattern of the radar device is vertically asymmetric with respect to a road surface.

Images