JP3397732B2 - FM-CW radar device - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is used for preventing collision of a moving body such as an automobile, and by transmitting and receiving a laser wave, a distance and a relative speed between another moving body and an obstacle outside the moving body. The present invention relates to an FM-CW radar device that detects

[0002]

2. Description of the Related Art A vehicle-mounted radar device mounted on a moving body such as an automobile monitors another automobile traveling in the front or the rear and simultaneously measures a distance and a relative speed to the other automobile. can do. In particular, the monitoring of the vehicle running ahead prevents the rear-end collision or collision of the vehicle.
It is expected to be very useful in the future.

The on-vehicle radar includes a millimeter-wave radar device that applies millimeter-wave band radio waves, and an FM-CW using a continuous wave of an FM-modulated wave by a triangular wave is also used in the millimeter-wave radar device method.
There is a radar system.

The FM-CW radar device transmits a continuous wave FM-modulated by a triangular wave as a transmission wave toward an obstacle such as an automobile located in front of the transmission wave. Then, the reflected wave reflected by the obstacle and returned is taken in as a received wave, and the beat signal obtained by mixing the transmitted wave and the received wave at that time is processed by a frequency analysis method such as FFT. The target observation position and relative velocity are calculated by.

When a target (another moving body or obstacle) is observed by a radar, a random noise component is superimposed on the original trajectory due to the effects of random noise and multipath due to road surface conditions. It is known that it will be in the form of being. A filter is generally used to remove a random noise component and extract a trajectory of an actual vehicle traveling. By filtering, it is possible to obtain the observed position and relative velocity (smooth position and smooth velocity) of the target that is on a curved line close to the actual trajectory of the target and exhibits a smooth trajectory.

In order to remove noise components and obtain a smooth vehicle trajectory, it is necessary to make a strong use of the effect of the filter. For example, FIG. 12 shows an observation position (marked with X) of a target traveling on an ideal straight line, but in the same figure, two curves (and) show a trajectory after filtering. The curve of is a locus when the coefficient that determines the effect of the filter is sufficiently used. The curve of
The trajectory is drawn with a slight error with respect to the ideal straight line.

As described above, the coefficient that determines the effect of the filter is generally set so that the smooth position of the target becomes extremely smooth according to the motion state when traveling on a straight road or a gentle curve.

On a straight road, a locus close to the actual running locus can be obtained by a sufficient filter effect, but when a road approaching a curved road with a small radius is reached, or when the target as shown in FIG. 13 changes lanes from adjacent lanes. ,
When the vehicle makes a sudden motion, such as when the vehicle is suddenly braked, on the contrary, the correct trajectory cannot be obtained due to the filter effect. Further, due to the response delay due to the response filter effect as shown in FIG. 14, when an interrupting vehicle occurs due to a sudden lane change, recognition of the vehicle is delayed.

[0009]

As described above, in the conventional FM-CW radar device, since the filter coefficient is set so as to exert a sufficient effect on the straight road, the target has a sharp curve. When the vehicle approaches a road, makes a lane change from an adjacent lane, or makes a sudden motion such as sudden braking, the trajectory drawn by the smooth position of the target deviates from the true trajectory of the target. In addition, there will be a delay in response. In particular, when an interrupted vehicle occurs, there is a problem that a response delay occurs due to a filter effect, a target detection is delayed, and an alarm such as a danger alarm cannot be output in a timely manner.

The present invention is intended to solve the above problems, and detects a smooth position of a target close to a true trajectory without detection delay even when the target makes a sudden movement. It is an object of the present invention to provide an FM-CW radar device that can be used.

[0011]

The FM-CW according to the present invention
The radar device is based on a conversion means for converting beat signal data into a frequency spectrum, a target recognition means for calculating an observation position of a target based on the frequency spectrum obtained by the conversion means, and an observation position calculated by the target recognition means. A target tracking unit that calculates a smooth position and a smooth velocity that are a position and a relative velocity on a smooth movement locus of the target, and the target tracking unit is, for example, a pre-calculation timing (pre-cycle).
Current calculation timing from the smooth position and speed at
Estimate and estimate the position of the target in the (current cycle)
Target estimated position calculation means for outputting fixed position information,
Target position calculated by estimated position information and target recognition means
Of the target's motion based on the information of the observation position of the target
Target motion state detecting means for detecting
Filter effect based on the detection result of motion state detection means
Set the coefficient that determines the
Processing to calculate smooth position and smooth velocity.
Including target smooth position calculation means, target motion state
For example, the target estimated position calculation means calculates the detection means.
The estimated position information and the target calculated by the target recognition means
The cumulative value of the position error from the target
Predetermined threshold and ratio for motion discrimination
It is characterized in that the motion state of the target is detected by comparison .

[0012]

[0013]

The target smooth position calculation means is, for example,
When the target motion state detecting means determines that the cumulative value of the position error is larger than the threshold value, a coefficient determined in advance corresponding to the rapid motion change of the target is set as a coefficient for determining the filter effect. Is configured as follows.

The target smooth position calculating means lowers the threshold value used by the target motion state detecting means by a predetermined value when the target motion state detecting means determines that the cumulative value of the position error is larger than the threshold value. It may be configured as follows.

Further, the target smoothed position calculation means corresponds to the movement of the target on a straight road or a gently curved road when the target motion state detection means determines that the accumulated value of the position error is less than or equal to the threshold value. Is set in advance and the threshold value used by the target motion state detecting means is returned to the original value.

As described above, by changing the coefficient for determining the filter effect according to the motion of the vehicle, a smooth trajectory can be obtained by the filter effect on a straight road or a gently curved road, and the recognition of the preceding vehicle is stabilized. To be executed. on the other hand,
When driving on a curved road with a small radius, lane change, or sudden braking, the response of vehicle tracking detection is improved by changing the coefficient that determines the effect of the filter by temporarily detecting the motion state of the vehicle. Even in such a case, the recognition delay can be minimized. Therefore, it is possible to obtain a locus of vehicle detection that is close to actual traveling even in various target traveling states, and it is possible to accurately perform risk determination.

The target motion state detecting means may be configured to detect the motion state of the target by comparing the cumulative value of the position error with a plurality of threshold values.

Further, the target smoothed position calculation means, when the target motion state detection means determines that the cumulative value of the position error is equal to or less than the threshold value of the smallest value, the target smoothed position calculation means detects the target on a straight road or a gently curved road. It may be configured to set a predetermined coefficient corresponding to the exercise and return all the threshold values used by the target exercise state detecting means to the original values.

[0020]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings. 1 is a block diagram showing an example of the configuration of an FM-CW radar device according to the present invention, FIG. 2 is a block diagram showing the internal block configuration of the target follower shown in FIG. 1, and FIG. 3 is the microprocessor shown in FIG. 4 is a flowchart for explaining the target estimated position calculation operation of the target follower by the microprocessor shown in FIG. 2, and FIGS. 5 and 9 are for the microprocessor shown in FIG. 7 and 10 are flowcharts for explaining the target motion state detection operation of the target follower, and FIGS. 7 and 10 are flowcharts for explaining the target smoothed position calculation operation of the target follower by the microprocessor shown in FIG.

In the configuration shown in FIG. 1, in the FM-CW radar device for detecting the distance to the target and the relative speed to the target by using the radio signal, the radar head 1 is provided with a transmitting antenna (not shown). , The FM wave frequency-modulated by the triangular wave is transmitted from the transmitting antenna to FM-C.
It is transmitted to the outside of the moving body in which the W radar device is installed.

The FM wave is reflected by an obstacle outside the moving body, returns as a reflected wave, and is received by the radar head 1 via a receiving antenna (not shown). The radar head 1 outputs a beat signal obtained by mixing the transmitted wave and the received wave to the low pass filter 2. The low pass filter 2 outputs the beat signal from which the unnecessary wave component is removed to the amplifier 3. The amplifier 3 amplifies the beat signal and outputs it to the AD converter 4. Then, the AD converter 4 outputs the digitized beat signal to the microprocessor 5.

The microprocessor 5 realizes, for example, a frequency analysis processor 51, a target recognizer 52, a target follower 53 and a danger discriminator 54. The frequency analysis processor 51 extracts a spectrum from the discrete value data of the beat signal sampled by the AD converter 4 at regular intervals using a frequency component analysis method such as FFT. The target recognizer 52 calculates the distance to the obstacle and the relative speed based on the spectrum calculated by the frequency analysis processor 51. The target follower 53 calculates a smooth position and a smooth velocity from the observed position and estimated position of the target.

Then, the danger discriminator 54 discriminates the degree of danger such as a collision between an obstacle and a moving body from the smoothed position and the smoothed velocity calculated by the target follower 53, and outputs a discrimination signal to the alarm device 6. When the degree of danger according to the discrimination signal is high, the alarm device 6 notifies the driver of the danger by displaying an alarm or issuing an alarm sound.

FIG. 2 is a block diagram showing an internal block configuration of the target follower 53. The estimated target position calculator 531 calculates an estimated position in the current cycle from the smoothed position of the target obtained in the previous cycle and velocity information. The target motion state detector 532 detects the motion state of the vehicle from the observed position of the target obtained by the target recognizer 52 and the estimated position of the target obtained by the target estimated position calculator 531. The target smooth position calculator 533 calculates the smooth position of the target according to the motion state obtained by the target motion state detector 532. Target estimated position calculator 531, target motion state detector 532, and target smooth position calculator 53
3 is realized by a program executed by the microprocessor 5.

FIG. 3 is a main routine showing the operation of the microprocessor 5 shown in FIG. In addition to the main routine, the microprocessor 5 also executes various operating programs such as frequency analysis processing.

In the main routine, when the operator (driver) operates the ignition key to operate the engine of the vehicle and the power is turned on, and the reset of the microprocessor 5 is released after a certain time has elapsed. Execution is started. That is, the microprocessor 5 is activated upon reset release, and the program is executed from, for example, address zero.

In the main routine, first, step S
At 1, initialization is performed and various variables are initialized.

The transmitted wave transmitted from the radar head 1 is reflected by the obstacle and returns as a received wave, and a beat signal is generated from the transmitted / received wave. At the same time that the frequency-modulated radio wave of the triangular wave is output, the AD converter 4 starts the AD conversion of the beat signal. The discrete-valued data of the beat signal is taken into the microprocessor 5 one by one and accumulated therein.

Further, the A-D conversion is executed at constant sampling time. Therefore, a timer interrupt or the like is generally used for monitoring the sampling time. When the digitization of the beat signal by the AD converter 4 is executed for one cycle of the triangular wave, discrete value data of the beat signal corresponding to the rising section and the falling section of the triangular wave is accumulated in the microprocessor 5 at that time. It will be.

The frequency analysis processor 51 of the microprocessor 5 obtains the discrete spectrum of each section from the discrete value data of the beat signal corresponding to the rising section and the falling section of the triangular wave by the frequency analysis method represented by FFT. (Step S2).

In step S3, the target recognizer 52 by the microprocessor 5 finds a combination of peaks in the up section and the down section based on the discrete spectrum obtained in step S2, and determines the target observation position and the target. Detect the relative speed between.

In step S4, the target follower 53 by the microprocessor 5 performs the target follow-up processing according to the vehicle motion state based on the target observation position obtained in step S3, and the target smooth position and smooth speed are obtained. Ask for.

Then, in step S5, the risk discriminator 54 by the microprocessor 5 discriminates the degree of danger to the vehicle based on the smoothed position and the smoothed speed obtained in step S4. If it is determined that the risk is high,
Alert the driver by displaying an alarm.

As described above, in this embodiment, the frequency analysis processor 51 for extracting the spectrum related to the target by processing the radio wave signal, and the peak frequency detected from the spectrum and relative to the target observation position. A target recognizer 52 that calculates the speed, a target estimated position calculator 521 that obtains an estimated position of the current cycle from the smoothed position and the smoothed velocity of the previous cycle, and a motion state of the vehicle is detected from the observed position and the estimated position of the current cycle. An FM- having a target motion state detector 522 and a target smooth position calculator 523 for obtaining a smooth position and a smooth velocity in the current cycle from the observed position and estimated position in the current cycle.
A CW radar device is used.

Next, the target follower 53, which is a feature of the present invention, will be described in detail. Target recognizer 52
Is the observation position (Rxc (k), Ryc (k)) of the target and the relative velocity R in the current cycle (k cycle).
Output Vyc (k). In addition, the target follower 5
3 is the smoothing position (PMx (k−1), PM) of the target in the previous cycle (this is referred to as k−1 cycle).
y (k-1)), smoothing speed (VSx (k-1), VSy
(K-1)) has already been output.

First, the operation of the estimated target position calculator 531 will be described with reference to the flowchart of FIG. In step S411, the target estimated position calculator 531 determines the x-direction smoothed position PMx in the previous cycle (k-1).
From (k) and the smoothing velocity VSx (k-1) in the x direction, the estimated position in the x direction PSx (k) of the current cycle (k) is calculated by the following equation.

X-direction estimated position PSx (k) = previous cycle smoothing position PMx (k-1) + previous cycle smoothing speed VSx
(K-1) * Tcyc

Here, Tcyc is a cycle time indicating the repetition period of the entire program.

In step S412, the target estimated position calculator 531 uses the y-direction smooth position PMy (k) and the y-direction smooth velocity VSy (k-1) in the previous cycle (k-1) in the same manner as in the x direction. The y-direction estimated position PSy (k) of the current cycle (k) is calculated by the following equation.

Y-direction estimated position PSy (k) = previous cycle smoothing position PMy (k-1) + previous cycle smoothing speed VSy
(K-1) * Tcyc

Then, the estimated target position calculator 531
To the target estimated position (PSx
(K), PSy (k)) are output.

Next, the operation of the target motion state detector 532 will be described with reference to the flowchart of FIG. 5 and the explanatory view of FIG. FIG. 6 is an explanatory diagram showing how the vehicle motion is detected. In step S421, the target motion state detector 532 determines the position of the target estimated position (PSx (k), PSy (k)) and the target observation position (Rxc (k), Ryc (k)) in the current cycle (k). The error ΔPn is calculated by the equation (1).

[0044]

[Equation 1]

As shown in FIG. 6, the position error ΔPn obtained by the equation (1) occurs when the motion state of the target changes abruptly. Its magnitude is proportional to the degree of change in the state of motion.

The target motion state detector 532 determines, in step S422, the past N including the current cycle (k).
The cumulative value ΣPn of the position error ΔPn in the cycle is calculated by the equation (2).

[0047]

[Equation 2]

In this example, the position error ΔPn of 5 cycles
The cumulative value of is used. Using the cumulative value of the position error is a very effective means to prevent the behavior of the vehicle from being erroneously detected due to the influence of sudden noise. By using the cumulative value of the position error, the motion state of the vehicle can be accurately detected.

In step S432, the target motion state detector 532 determines whether the motion state of the vehicle has changed abruptly. For example, the cumulative value of position error ΣΔPn
Is compared with a vehicle behavior determination threshold value VSMth.

As a result of the discrimination in step S432, the cumulative value ΣΔPn of the position errors is the threshold VSM for discriminating the behavior of the vehicle.
If it is determined that it is larger than th, step S424.
Proceed to. In step S424, the target motion state detector 532 determines the target rapid motion flag THMFLG.
Turn on.

On the other hand, if the cumulative value ΣΔPn of the position errors is equal to or less than the vehicle behavior determination threshold value VSMth, in step S425, the target rapid motion flag THM.
Turn off FLG. After the above processing is completed, the routine is ended (return).

Next, the operation of the target smoothed position calculator 533 will be described with reference to the flowchart of FIG. 7 and the explanatory view of FIG. FIG. 8 is an explanatory diagram for explaining the change of the threshold (changing the holding of the threshold). Step S431
At, the target smooth position calculator 533 checks ON / OFF of the target rapid motion flag THMFLG. If the target rapid motion flag THMFLG is ON, the process proceeds to step S432.

In step S432, the target smoothed position calculator 533 sets predetermined coefficients αTHM and βTHM to the coefficients αth and βth that determine the filter effect in order to improve the response to the rapid movement of the vehicle. . Further, in step S433, in order to provide the threshold value VSMth for vehicle behavior determination with a hysteresis function, VSMth is reduced by a certain value from VSMth.
Set −. That is, as shown in FIG. 8, when the motion state of the vehicle suddenly changes, the vehicle behavior determination threshold value VSMth is lowered by a certain value in order to ensure the stability of the vehicle trajectory.

Here, in general, β = α ² / (2-α), and αth and βth <1. ΑTHM and βTHM are set to values close to 1 in order to cope with the rapid movement of the target. Also, as a constant value of the threshold decrease,
For example, a value corresponding to 1/2 lane width is used to cope with a sudden movement such as a lane change with the target.

In step S431, if the target rapid motion flag THMFLG is OFF, the process proceeds to step S437. In step S437, the target smooth position calculator 533 determines predetermined coefficients αNOR and βN for the coefficients αth and βth that determine the filter effect in order to obtain a sufficiently smooth trajectory on a straight road or a gently curved road.
Set OR.

Here, αNOR and βNOR are set to values close to 0 so that a sufficiently smooth locus can be obtained on a straight road or a gently curved road.

Further, in step S438, as shown in FIG. 8, the vehicle behavior determination threshold value VSMth is set to
Set VSMth + for canceling the hysteresis function from VSMth- which is reduced by a fixed value.

Then, in step S440, the target smooth position calculator 533 determines the coefficients αth and βth that determine the filter effect determined by the motion state of the vehicle.
The smooth position and smooth velocity of the target based on (3)
~ Calculated by the equation (6). After the above processing is completed, the routine is ended (return).

[0059]

[Equation 3]

As described above, in this embodiment, the rapid motion of the target is detected from the estimated position in the current cycle based on the observed position of the target in the current cycle and the smoothed position of the target in the previous cycle. . When such a rapid movement change is detected, the filter effect can be weakened and the responsiveness can be improved by temporarily changing the coefficient for determining the strength of the filter effect. As a result, the delay time can be minimized even when the target rapidly moves, and the smooth position can be brought close to the actual target trajectory.

Here, examples of abrupt movements of the vehicle include a case where a vehicle is approaching a curved road with a small radius, a case where a lane change is made from an adjacent lane as shown in FIG. 13, and a case where a sudden braking is applied. The vehicle motion state is different in each case. Therefore, by dividing the exercise state into several groups and setting the coefficient for determining the filter effect for each group, it is possible to obtain the optimum filter effect corresponding to the exercise state.

FIGS. 9 and 10 show an example in which the motion state is divided into three groups and the coefficient for determining the filter effect is set to obtain the smooth position and the smooth velocity of the target.

First, the operation of the target motion state detector 532 will be described with reference to FIG. In step S421, the target motion state detector 532 determines the target estimated position (PSx (k), PS in the current cycle (k).
y (k)) and target observation position (Rxc (k), Ry
The position error ΔPn of c (k)) is calculated by the equation (1).

As already described (see FIG. 6), (1)
The position error ΔPn obtained by the expression occurs when the motion state of the target changes abruptly. Its magnitude is proportional to the degree of change in the state of motion.

Next, the target motion state detector 532
Is the cumulative value ΣP of the position error ΔPn in the past N cycles including the current cycle (k) in step S422.
n is calculated by the equation (2).

As described above, using the cumulative value of the position error is a very effective means for preventing the vehicle behavior from being erroneously detected due to the influence of sudden noise. Thereby, the motion state of the vehicle can be accurately detected.

The target motion state detector 532 determines in step S423A whether the motion state of the vehicle has changed abruptly. That is, the cumulative value ΣΔPn of the position error is compared with the vehicle behavior determination threshold value VSMth3. As a result of the comparison, when it is determined that the cumulative value ΣΔPn of the position error is larger than the vehicle behavior determination threshold value VSMth3, the process proceeds to step S424A. In step S424A, the target rapid motion flag THMFL is set.
Turn on G3.

On the other hand, if the cumulative value ΣΔPn of the position error is less than or equal to the threshold VSMth3 for determining the behavior of the vehicle, the process proceeds to step S423B. Target motion state detector 532
Is the cumulative value Σ of the position error in step S423B.
ΔPn is compared with the threshold VSMth2 for determining the behavior of the vehicle. As a result of the comparison, when it is determined that the cumulative value ΣΔPn of the position error is larger than the vehicle behavior determination threshold value VSMth2, the target rapid motion flag THMFLG2 is turned on in step S424B.

If the cumulative value ΣΔPn of the position error is less than or equal to the threshold VSMth2 for determining the behavior of the vehicle, step S4
Proceed to 23C. The target motion state detector 532 determines, in step S423C, the cumulative value ΣΔPn of the position errors.
And the threshold VSMth1 for determining the behavior of the vehicle are compared. As a result of the comparison, when it is determined that the cumulative value ΣΔPn of the position error is larger than the vehicle behavior determination threshold value VSMth1, the target rapid motion flag THMFLG1 is turned on in step S424C.

If the cumulative value ΣΔPn of the position error is less than or equal to the threshold VSMth1 for determining the behavior of the vehicle, step S
425A, the target rapid motion flag THMF
Turn off LG1, THMFLG2, and THMFLG3. After the above processing is completed, the routine is ended (return).

Here, VSMth1 is set to about 1/3 of VSMth shown in FIG. 8, and VSMth2 is set.
Is set to about 2/3 of VSMth, for example.

Next, the operation of the target smooth position calculator 533 will be described with reference to the flowchart of FIG. 10 and the explanatory view of FIG. FIG. 11 is an explanatory diagram for explaining the change of the threshold (changing the holding of the threshold). As shown in FIG. 11, the threshold value for detecting the motion state of the vehicle is set stepwise according to the motion state.

The target smooth position calculator 533 determines in step S431A the target rapid motion flag T.
Check ON / OFF of HMFLG3. If the target rapid motion flag THMFLG3 is ON, the coefficients αth and βt that determine the filter effect are determined in step S432A.
Predetermined coefficients αTHM3 and βTHM3 are set to h in order to improve the responsiveness to the rapid movement of the vehicle.

Further, in step S433A, as shown in FIG.
As shown in FIG. 1, the threshold VSMth for determining the behavior of the vehicle
In order to provide a hysteresis function, VSMth3 is set to VSMth3 with a threshold value lowered from VSMth3 by a certain value. Similarly, for VSMth2 and VSMth1, VS
Mth2- and VSMth1- are set. With such settings, if the vehicle's motion state changes suddenly,
In order to ensure the stability of the trajectory of the vehicle, the vehicle behavior determination threshold value VSMth is lowered by a certain value. Further, in step S435A, the target rapid motion flag THMFLG3 is turned off.

If the target rapid motion flag THMFLG3 is OFF in step S431A, step S4
Proceed to 31B. In step S431B, the target smooth position calculator 533 checks ON / OFF of the target rapid motion flag THMFLG2. When the target rapid motion flag THMFLG2 is ON, the process proceeds to step S432B. In step S432B, the target smoothed position calculator 533 determines the coefficients αth and βth, which determine the filter effect, by the predetermined coefficients αTHM2 and βTHM for improving the responsiveness to the rapid movement of the vehicle.
Set 2.

Further, in step S433B, as shown in FIG.
As shown in FIG. 1, the threshold VSMth for determining the behavior of the vehicle
2 is set to VSMth2- in which the threshold value is lowered from VSMth2 by a certain value in order to have a hysteresis function. Similarly, VSMth1- is set for VSMth1. With such a setting, when the vehicle motion state suddenly changes once, the vehicle behavior determination threshold value VSMth is lowered by a certain value in order to ensure the stability of the vehicle trajectory.

Further, in step S434B, the threshold value VSMth3- is set to VSMth3 + in order to cancel the hysteresis function of the vehicle behavior determination threshold value VSMth3.
Return to. Further, in step S435B, the target rapid motion flag HMFLG2 is turned off.

If the target rapid motion flag THMFLG2 is OFF in step S431B, step S4
Proceed to 31C. In step S431C, the target smooth position calculator 533 checks ON / OFF of the target rapid motion flag THMFLG1. If the target rapid motion flag THMFLG1 is ON, the coefficients αth and β that determine the filter effect are determined in step S432C.
Predetermined coefficients αTHM1 and βTHM1 are set to th in order to improve the responsiveness to the rapid movement of the vehicle.

Further, in step S433C, as shown in FIG. 11, the threshold VSMth1 for determining the behavior of the vehicle is set to
In order to provide a hysteresis function, VSMth1- is set by lowering the threshold value from VSMth1 by a certain value. With such a setting, when the vehicle motion state suddenly changes once, the vehicle behavior determination threshold value VSMth is lowered by a certain value in order to ensure the stability of the vehicle trajectory.

Further, in step S434C, the threshold value VSMth3- is set to VSMth3 + in order to cancel the hysteresis function of the threshold value VSMth3 for determining the behavior of the vehicle.
Return to. Similarly, the hysteresis is released for the threshold value VSMth2. Further, in step S435C, the target rapid motion flag THMFLG1 is turned off.

In the above processing, αTHM1 and βTH
M1 is, for example, αTH used in the previous embodiment.
It is set to a value of about 1/3 of M and βTHM.

If the target rapid motion flag THMFLG1 is OFF in step S431C, step S43 is executed.
Proceed to 7. In step S437, the target smooth position calculator 533 determines the coefficient αt that determines the filter effect.
Predetermined coefficients αNOR and βNOR are set to h and βth in order to remove noise components and obtain a smooth trajectory.

In step S439, the threshold value VSMth3- is returned to VSMth3 + in order to cancel the hysteresis function of the vehicle behavior determination threshold value VSMth3. Similarly, the hysteresis is released for the threshold values VSMth1 and VSMth2.

Then, in step S440, the target smoothed position calculator 533 determines the coefficients αth and βth which determine the filter effect determined by the motion state of the vehicle.
The smooth position and smooth velocity of the target based on (3)
~ Calculated by the equation (6). After the above processing is completed, the routine is ended (return).

As described above, the target motion state is divided into several groups and the threshold values are switched in stages, so that the optimum filter effect according to the target motion state can be obtained.

[0086]

According to the present invention, in the FM-CW radar device, when the target tracking means calculates the smooth position and the smooth velocity by using the filter processing, the coefficient for determining the filter effect is used as the target motion state. As the target vehicle is approaching a sharp curve, lane change from an adjacent lane, sudden braking, etc. By changing the coefficient for determining the filter effect in the means for detecting such movement and obtaining the smoothed position of the target, the response characteristic of the target detection can be improved.

Further, by setting the threshold value stepwise according to the motion state of the vehicle, it is possible to set the coefficient that determines the filter effect most suitable for each motion state. Can be calculated. Therefore, even if an interrupted vehicle or the like occurs, the detection delay can be suppressed to the minimum, and the danger warning based on the inter-vehicle distance can be quickly executed.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration example of an FM-CW radar device according to the present invention.

FIG. 2 is a block diagram showing an internal block configuration of a target follower.

FIG. 3 is a flow chart for explaining the operation of the microprocessor.

FIG. 4 is a flowchart for explaining a target estimated position calculation operation of a target follower by a microprocessor.

FIG. 5 is a flowchart for explaining a target motion state detection operation of the target follower by the microprocessor.

FIG. 6 is an explanatory diagram showing how vehicle motion is detected.

FIG. 7 is a flowchart for explaining a target smoothed position calculation operation of the target follower by the microprocessor.

FIG. 8 is an explanatory diagram for explaining a change of a threshold (change of holding of a threshold).

FIG. 9 is a flowchart for explaining a target motion state detection operation of the target follower by the microprocessor.

FIG. 10 is a flowchart for explaining a target smoothed position calculation operation of a target follower by a microprocessor.

FIG. 11 is an explanatory diagram for explaining a change of a threshold value (change of holding threshold value).

FIG. 12 is an explanatory diagram showing an observation position of a target traveling on an ideal straight line.

FIG. 13 is an explanatory diagram showing a case where a target has changed lanes from an adjacent lane.

FIG. 14 is an explanatory diagram for explaining a response delay due to a filter effect.

[Explanation of symbols]

1 radar head 2 Low pass filter 3 amplifier 4 A-D converter 5 microprocessors 6 alarm 51 Frequency analysis processor 52 Target recognizer 53 Target follower 54 Risk classifier 531 Target Estimated Position Calculator 532 Target motion state detector 533 Target smooth position calculator

─────────────────────────────────────────────────── ─── Continuation of the front page (56) Reference JP-A-10-206531 (JP, A) JP-A-11-72557 (JP, A) JP-A-9-90027 (JP, A) JP-A-8- 278361 (JP, A) JP 64-73274 (JP, A) JP 11-14748 (JP, A) JP 9-90026 (JP, A) JP 5-150038 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) G01S ⁷ /00-7/42 G01S 13/00-13/95 B60R 21/00 G08G 1/16

Claims (6)

(57) [Claims] 1. A continuous wave that is frequency-modulated with a triangular wave is transmitted.
The waves that are transmitted as waves are generated from the transmitted and received waves.
External objects based on the spectrum of the signal
FM-CW radar device that detects the distance between and
In the Conversion to convert beat signal data to frequency spectrum
Means and Based on the frequency spectrum obtained by the conversion means
Target recognition means to calculate the observation position of the target
When, Based on the observation position calculated by the target recognition means
Position and phase in a smooth moving trajectory of the target
A target that calculates smooth position and smooth velocity, which is the velocity
And a tracking unit, The target tracking means isFrom the smooth position and smooth velocity at the previous calculation timing
Estimate the target position at the current calculation timing
Target estimated position calculating means for outputting estimated position information
When, The estimated position information and the target calculated by the target recognition means.
Target movement based on the information of the observed position of the get
Target motion state detection means for detecting the state, Based on the detection result of the target motion state detection means
Set the coefficient that determines the filter effect by
Smoothing position and smoothing velocity by filtering using
And a target smooth position calculation means for calculating The target motion state detecting means is a target estimation position.
Estimated position information calculated by the position calculation means and the target recognition hand
Accumulation of position error calculated by Dan from the target's observation position
The value is set in advance to determine the vehicle behavior.
Target motion by comparing with a threshold
Detect state FM-CW radar device characterized by
Place 2. The target smoothed position calculating means determines the rapid motion change of the target as a coefficient for determining the filter effect when the target motion state detecting means determines that the cumulative value of the position error is larger than the threshold value. F according to claim 1, wherein setting the coefficients corresponding to that predetermined
M-CW radar device. 3. The target smoothed position calculating means sets the threshold value used by the target motion state detecting means to a predetermined value when the target motion state detecting means determines that the cumulative value of the position error is larger than the threshold value. The FM-CW radar device according to claim 2, wherein the FM-CW radar device is lowered only by a certain amount. 4. The target smoothed position calculating means corresponds to the movement of the target on a straight road or a gently curved road when the target motion state detecting means determines that the cumulative value of the position error is less than or equal to a threshold value. And a threshold value used by the target motion state detecting means is returned to the original value.
The FM-CW radar device according to item 3 . 5. A target motion state detecting means, claims 1 to 3 for detecting the motion state of the target by comparing the cumulative value of the position error between a plurality of threshold
The described FM-CW radar device. 6. The target smoothed position calculating means, when the target motion state detecting means determines that the cumulative value of the position error is equal to or smaller than the minimum threshold value, the target smooth position calculating means In addition to setting a predetermined coefficient corresponding to exercise,
The FM-CW radar device according to claim 5, wherein all the threshold values used by the target motion state detecting means are returned to their original values.