* NOTICE*
`
`JPO and INPIT are not responsible for any damages caused bythe useofthis translation.
`
`1. This document has been translated by computer. So the translation may notreflect the original precisely.
`
`2.“ shows a word which cannotbetranslated.
`
`3.
`
`In the drawings, any words are nottranslated.
`
`Patent Number
`
`JP2007257536A
`
`Bibliography
`
`(19) [Publication country] JP
`
`(12) [Kind of official gazette] A
`
`(11) [Publication number] 2007257536
`
`(43) [Date of publication of application] 20071004
`
`(54) [Title of the invention] ROAD TRAFFIC MONITORING DEVICE BY MEANS OF
`
`MILLIMETER-WAVE RADAR
`
`(51) [International Patent Classification]
`
`GO08G
`
`1/01
`
`(2006.01)
`
`GO1S
`
`13/91
`
`(2006.01)
`
`[Fl]
`
`GO08G
`
`1/01
`
`GO1S
`
`13/91
`
`K
`
`Z
`
`(21) [Application number] 2006083934
`
`(22) [Filing date] 20060324
`
`(71) [Applicant]
`
`[Name] FUJITSU LTD
`
`(72) [Inventor]
`
`[Full name] KAJIKI JUNKO
`
`[Full name] OTONARI SHOE!
`
`[Full name] OKAMOTO SHINYA
`
`[Full name] SENKAWA YUKA
`
`Abstract
`
`(57) [Overview]
`
`PROBLEM TO BE SOLVED: Toprovide a road traffic monitoring device which monitors
`
`vehicle traffic while automatically judging the number of lanes, width of each lane and
`
`

`

`width of a median strip, etc. by measuring a running vehicle by a millimeter-wave radar.
`
`SOLUTION: Millimeter-wave beams 7 are emitted by the millimeter-wave radar 2 and a
`
`plurality of running vehicles are detected by reception of reflected waves. Distance from
`
`the millimeter-wave radar 2, a direction angle and speed, etc. are calculated from
`
`relation among an irradiation angle, time to reflection and Doppler shift, a running
`
`trajectory of a vehicle is calculated, relation between the number of running vehicles and
`
`the vehicle width direction is made into a graph, grouped, the number of lanes is
`
`extracted from a mountain of the graph andthe lane width is extracted from the mountain
`
`and valleys of both sides of the graph. The width of the median strip at a boundary part
`
`where the speed component direction of the running trajectory switches from positive to
`
`negative is calculated. The vehicle traffic is measured and monitored based on road
`
`arrangement obtained by the above.
`
`Claim
`
`[Patent Claims]
`
`[Claim 1]
`
`A travel locus acquisition means for acquiring a travel locus of a vehicle passing through
`
`a road by a millimeter wave radar ;
`
`A road lane number determining means for determining the numberof road lanes based
`
`on the traveling locus ;
`
`A road traffic monitoring device comprising :.
`
`[Claim 2]
`
`A travel locus acquisition means for acquiring a travel locus of a vehicle passing through
`
`a road by a millimeter wave radar ;
`
`A road width determining means for determining a width of a road lane based on the
`
`traveling locus ;
`
`A road traffic monitoring device comprising :.
`
`[Claim 3]
`
`A road traffic monitoring device according to claim 1 and 2, further comprising road
`
`arrangement determining means for determining a road arrangement based on the
`
`number of road lanes and the width of the road lane.
`
`[Claim 4]
`
`The road traffic monitoring device according to claim 3, further comprising a measuring
`
`unit that measures the number of traveling of the vehicle per road lane based on the
`
`road arrangement.
`
`[Claim 5]
`
`

`

`The road traffic monitoring device according to claim 1, wherein the travel
`
`locus
`
`acquisition unit acquires a travel locus of a vehicle passing through a road periodically
`
`using a millimeter wave radar.
`
`Description
`
`[Detailed description of the invention]
`
`[Technical field]
`
`[0001]
`
`The present invention relates to a road traffic monitoring apparatus for monitoring a
`
`vehicle traffic volume by automatically determining the number of lanes and each lane
`
`width by means of a millimeter wave radar.
`
`[Background of the Invention]
`
`[0002]
`
`As the number of vehicles traveling on a road increases, various vehicle monitoring
`
`devices and monitoring methods for monitoring a traveling vehicle have been proposed.
`
`For example, a vehicle sensing apparatus has been proposed in which, for a vehicle
`
`entering an intersection, a vehicle is determined to be a right turn, a left turn, and a
`
`straight vehicle based on a vehicle trajectory extracted from a video image, and each
`
`traffic volume is measured. (For example, see Patent document 1).
`
`There has also been proposed a road traffic flow measuring apparatus and a
`
`measurement method for measuring traffic volume by accurately detecting a vehicle
`
`passing through a road on the basis of locus data obtained by tracking the vehicle
`
`position with time from an image captured at a predetermined cycle without providing a
`
`specific vehicle discovery region. (For example, see Patent document2).
`
`Further, there has been proposed a traveling vehicle detection method using a millimeter
`
`wave radar which reliably performs tracking detection of a traveling vehicle underall
`
`weather conditions and prevents rear-end collision and rear-end collision at an
`
`intersection of a line-of-sight defect. (For example, see Patent document3).
`
`However, even if the vehicle is detected as described above, it is necessary to know
`
`which lane the vehicle is traveling in order to perform the monitoring of the traffic flow
`
`using the result.
`
`[0003]
`
`For this reason, when a traveling vehicle detection device such as a radar is installed, it
`
`is necessary to measure details of the road arrangementat the installation location and
`
`to provide data on the road arrangement such as the number oflanes, the lane width,
`
`and the central separation band width obtained by the measurement to a computer
`
`

`

`connected to the vehicle detection device in advance as a parameter.
`
`[0004]
`
`Conventionally, a computer associated with a traveling vehicle detection device
`
`determines a lane on which a vehicle is traveling by comparing the parameter provided
`
`as described above with a measurementresult by the traveling vehicle detection device.
`
`[Patent document 1]JP 2000 090389 (Abstract, FIG. 1). </S>.
`
`[Patent document 2]JP 2002 358595 (Abstract, FIG. 1). </S>.
`
`[Patent document 3]JP 2002 - 099986A (a summary, Fig.1, Fig.3)
`
`[Disclosure of invention]
`
`[Problem to be solved by the invention]
`
`[0005]
`
`The technique of Patent Document 1 is to determine the direction of travel of the vehicle
`from the track of the vehicle and measurethe traffic volume in each direction of travel of
`
`the vehicle, and does not recognizethe lane itself.
`
`Further, the technique of Patent Document 2 is premised on not providing a specific
`
`vehicle discovery region from the beginning, and does not include an idea of recognizing
`the lane itself.
`
`[0006]
`
`In any event, these devices or methods are provided. To solve the problem that vehicle
`
`detection cannot be accurately performed in a dark state such as a dark fog, a rain, a
`
`snowfall or the like, or in a dark state such as a night state, when an image sensor such
`
`as an image sensor suchasvisible light or infrared light is used.
`
`[0007]
`
`On the other hand, although the technique of Patent Document 3 surely performs
`
`tracking detection of a traveling vehicle under all weather conditions by a millimeter wave
`
`radar that is hardly affected by bad weather or the like, it
`
`is intended to prevent an
`
`accident such as a collision, and does not include the idea of recognizing the lane itself.
`
`[0008]
`
`Further, a road arrangement information such as the number of lanes of a road which
`
`becomes a measurement range, a lane width, a width of a center separation band, and
`
`a width of a road shoulder is given to the traveling vehicle detection device as a
`
`parameter in advance. In the case where there is no detailed map information (survey
`
`data), it takes time and labor to temporarily close the road in order to acquire the road
`
`arrangementinformation, I. e., to actually measure the road arrangement.
`
`[0009]
`
`

`

`As described above,
`
`it
`
`is necessary to provide the data of the measured road
`
`arrangement as a parameter to a computer connectedto the traveling vehicle detection
`
`device. Thus, there was a problem that a great man-hour and expensestarted.
`
`[0010]
`
`In addition,
`
`it is difficult to confirm a white line on a road during bad weather in which
`
`poorvisibility is caused by snow, fog, or the like. Further,
`
`if there is a snowy snow
`
`removed from a road surface on a central separation band or a road shoulder, there is a
`
`case where a lane becomesnarrow. In addition, depending on road construction or the
`
`like, a lane different from a normal lane is often temporarily provided. In any event, the
`
`vehicle will not be able to run correctly and accordingly on a defined lane.
`
`[0011]
`
`Whensuch a vehicle cannot travel along a predetermined road, there is a problem that
`
`an accurate measurementof a traveling vehicle cannot be performed, |. e., an accurate
`
`road monitoring cannot be performed because an actual parameter such as a lane width
`
`fluctuates from a predetermined parameter.
`
`[0012]
`
`In view of the above circumstances, it is an object of the present invention to provide a
`
`road traffic monitoring apparatus which monitors a traveling vehicle by a millimeter wave
`
`radar and monitors the traffic volume of a vehicle while automatically determining the
`
`number of lanes, the width of each lane, and the like.
`
`[Means for solving the problem]
`
`[0013]
`
`First, a road traffic monitoring apparatus according to a millimeter wave radar of the 1
`
`aspect of the present invention includes a travel locus acquisition unit which acquires a
`
`travel locus of a vehicle passing through a road by a millimeter wave radar, and a road
`lane number determination unit which determines the number of road lanes based on
`
`the travel locus.
`
`[0014]
`
`Next, a road traffic monitoring device according to a 2 aspect of the present invention is
`
`configured to include a travel locus acquisition unit that acquires a travel locus of a
`
`vehicle passing through a road by a millimeter wave radar, and a road width
`determination unit that determines a width of a road lane based on the travel locus.
`
`[0015]
`
`Further, a road traffic monitoring device using the millimeter wave radar according to the
`
`3 aspectof the present invention is characterized in that the roadtraffic monitoring device
`
`

`

`obtains a travel locus of a vehicle passing through a road on a millimeter wave radar.
`
`The vehicle is provided with a road lane number determining means for determining the
`
`numberof road lanes based on the traveling locus, a road lane width determining means
`
`for determining the width of the road lane based on the traveling locus, and a road
`
`arrangement determining means for determining the road arrangement on the basis of
`the number of road lanes and the width of the road lane.
`
`[0016]
`
`The running-locus acquisition means from which the road traffic monitoring instrument
`
`by the millimeter wave radar of the 4th invention acquires the running locus of vehicles
`
`which passes through a road with a millimeter wave radar and a number determination
`means of road lanes to determine the number of road lanes based on the above-
`
`mentioned running locus, The vehicle is provided with a road lane width determining
`
`means for determining a width of a road lane based on the traveling locus, a road
`
`arrangement determining means for determining a road arrangement on the basis of the
`
`numberof the road lane and the width of the road lane, and a measuring means for
`
`measuring the traveling number of the vehicle for each road lane based on the road
`
`arrangement.
`
`[0017]
`
`And the running-locus acquisition means from which the road traffic monitoring
`
`instrument by the millimeter wave radar offifth invention acquires the running locus of
`
`vehicles which passes through a road with a millimeter wave radar periodically and a
`
`number determination means of road lanes to determine the number of road lanes based
`
`on the above-mentioned running locus, The vehicle is provided with a road lane width
`
`determining means for determining a width of a road lane based on the traveling locus,
`
`a road arrangement determining means for determining a road arrangement on the basis
`
`of the number of the road lane and the width of the road lane, and a measuring means
`
`for measuring the traveling number of the vehicle for each road lane based on the road
`
`arrangement.
`
`[Effect of the Invention]
`
`[0018]
`
`According to the present
`
`invention,
`
`it
`
`is possible to automatically recognize road
`
`arrangement information such as the number of lanes, the lane width, and the central
`
`separation band of a road on the side of a road using the detected travel locus of a
`
`vehicle only by installing a road traffic monitoring device. Thus, it becomes unnecessary
`
`to provide road arrangementinformation as a parameter as a parameter in advancefor
`
`the road traffic monitoring device, so that a large number of man-hours such as the
`
`

`

`survey of the road arrangement when the roadtraffic monitoring device is installed on
`
`the site can be reduced and the economical efficiency can be improved.
`
`[0019]
`
`Further, since the road arrangement information such as the number of lanes of the road,
`
`the lane width, and the center separation band is automatically recognized by using the
`
`travel
`
`locus of the vehicle on a regular basis. Even if the vehicle cannot travel
`
`accordance with the prescribed lane due to weather, construction, or the like,
`
`it
`
`in
`
`is
`
`possible to measure an accurate numberof passing vehicles for each of the temporarily
`
`set lanes, and it is possible to greatly contribute to the monitoring of the road traffic under
`various environments.
`
`[Best modefor carrying out the invention]
`
`[0020]
`
`Hereinafter, embodiments of the present invention will be described with reference to the
`
`accompanying drawings.
`
`[Example 1]
`
`[0021]
`
`FIG. 1
`
`is a block diagram showing a configuration of a road traffic monitoring apparatus
`
`using a millimeter wave radar in an embodiment. As shown in FIG. 1, a road traffic
`
`monitoring device 1 using a millimeter wave radar includes a millimeter wave radar 2
`
`corresponding to a millimeter wave transmission / reception device and a computer 3 for
`
`measuring a number of vehicles connected to the millimeter wave radar 2.
`
`[0022]
`
`FIG. 2 is a diagram showing an example of a road arrangement of a road on which the
`
`millimeter wave radar 2 of the road traffic monitoring apparatus 1 according to the
`
`millimeter wave radar is installed in one embodiment.
`
`As shownin FIG. 4, the millimeter wave radar 2 is installed at a position in which, for
`
`example, the entire road in the vicinity of the road 4 can be seen in advance. This
`
`millimeter wave radar 2 emits a narrow beam 6 of millimeter wave on a road 5 to be
`
`monitored, while scanning the beam in a range of a scanning angle 7 indicated by a
`
`broken line, and receives the reflected wave.
`
`[0023]
`
`In this example, although it depends on the function of a motor (not shown) that drives
`
`the millimeter wave radar 2 to oscillate, the scanning angle 7 is 18 ° and the scanning
`
`period is 400. That is, scanning is performed at a rate of about 2.5 times / second.
`
`[0024]
`
`

`

`FIG. 3 is a flowchart of the detection of the traveling vehicle, the number of lanes and
`
`the detection of the lane width performed by the road traffic monitoring device 1 using
`the millimeter wave radar.
`
`In FIG. 3, first, a plurality of traveling vehicles are detected (S 1). In this processing, the
`
`distance, the direction angle, the speed, and the like from the millimeter wave radar 2
`
`are calculated from the time until the reception of the reflected wave of the millimeter
`
`wave beam7irradiated by the millimeter wave radar 2 and the Doppler shift, and the
`
`vehicle detection is performed from the result.
`
`[0025]
`
`For example, a method disclosed in Patent Document 3 is used for detecting a traveling
`
`vehicle. In the detection of the traveling vehicle, traffic information such as the number
`
`of traveling vehicles and presence / absence of a stopped vehicle is also detected.
`
`Further, there is no upper limit on the number of vehicles to be detected, and detection
`
`of the vehicle is performed every measurementperiod by the millimeter wave beam 7.
`
`[0026]
`
`Next, the travel locus of the vehicle is obtained using the detection result of the traveling
`
`vehicle (S 2). In this process, the detection result of the traveling vehicle is plotted on an
`
`x-y road coordinate axis whoseorigin is an installation position of the millimeter wave
`
`radar 2.
`
`[0027]
`
`FIG. 4 is a diagram showing the number of traveling vehicles plotted on the road
`
`coordinate system based on the detection result of the traveling vehicle performed by
`
`the road traffic monitoring device using the millimeter wave radar as described above
`
`and the traveling locus thereof.
`
`[0028]
`
`FIG. 4 shows x-y road coordinates with an installation position of the millimeter wave
`
`radar 2 as an origin 0, and shows a vehicle width direction on a horizontal axis x and a
`
`vehicle travel direction on a vertical axis y. A plot 8 of black circles showsthe position of
`
`the detected vehicle obtained by 4 periods of scanning.
`
`[0029]
`
`A line 9 connecting a plot 8 of 4 blackcircles in the vehicle running direction of the vertical
`
`axis y indicates the travel locus of the individual traveling vehicle.
`
`Generally, since the vehicle travels within a predetermined lane range although there is
`
`somevariation in the driving position within the road range, it is possible to roughly divide
`
`the density of the vehicle into a portion where the density of the traveling locus 9 is
`
`increased and a portion where the density is reduced, as shownin FIG. 4.
`
`

`

`[0030]
`
`Returning to FIG. 3, following extraction of the travel locus, grouping is performed (S 3).
`
`In this process, a relationship between the number of traveling vehicles and the lane
`
`width direction is graphed to extract a traveling point (lane).
`
`FIG. 5 is a graph showing the relationship between the number of traveling vehicles and
`
`the lane width direction on the basis of the result of detection of the traveling vehicle
`
`performed by the road traffic monitoring apparatus using the millimeter wave radar
`described above.
`
`[0031]
`
`In this figure, a horizontal axis x indicates a lane width direction, and a vertical axis y
`
`indicates a vehicle running number. In addition, the graph shown in FIG. 4 shows an
`
`average value of the vehicle passing through the same point in the lane width direction
`of the horizontal axis x.
`
`In other words, a value in the x-axis direction may be determined by taking an average
`
`value of the detected traveling locus, since the traveling locus may not be a constant
`
`value due to a change in lane or the like.
`
`[0032]
`
`For example, assuming that 3 points [(a 1, b 1), (a2, b 2), (a 3, b 3)] are detected as the
`
`running plot point of the 1 vehicle, the value of the x-axis at that time is calculated as "x
`
`=(a1a2a3)/3".
`
`[0033]
`
`Referring back to FIG. 3, based on the result of the grouping, the lane number and the
`
`lane width are detected (S 4). First, as shown in FIG. 5, the obtained graph is similar to
`that shown in FIG. 4.
`
`It is divided into the section when the numberof traveling vehicles increases, and little
`
`section. Thatis, it becomes a graph whichrises and falls in a mountain and a valley.
`
`[0034]
`
`From this result, the numberof lanes is detected by detecting the number of portions 11
`
`hitting the apex of the mountain where the number of traveling vehicles increases.
`
`Further, according to the above graph, the width of each lane is estimated by determining
`
`that one apex portion is included and a width up to the valley portion is one lane.
`
`[0035]
`
`FIG. 6 shows the same graph as FIG. 5 with changing the view point
`
`in order to
`
`specifically illustrate the estimation method of the lane width.
`
`In other words,
`
`it
`
`is
`
`determined that the portion of the number of traveling vehicles shownin FIG. 6 becomes
`
`

`

`a valley 12.
`
`[0036]
`
`In other words, it is determined that the width of the valley portion 12 and the trough
`
`portion 12, which includes one portion 11 of the mountain portion where the number of
`
`traveling vehicles is large and that sandwiches the apex 11, is one lane. Thus, the width
`of each lane is estimated.
`
`[0037]
`
`FIG. 6 shows a 1 lane width, a 2 lane width, a 3 lane width, and a 4 lane width estimated
`
`by this method. In FIG. 6, only the 3 lane width is extremely wider than the other vehicle
`
`width. From this, it is possible to estimate the width of the central separation band from
`
`the difference from the other lane widths, since the actual 3 lane width and the width of
`
`the central separation zone are included in FIG. 6 and the actual 3 lane width should be
`
`equivalent to the other lane widths of the 3 lane.
`
`[0038]
`
`In the automatic determination of the road arrangement in this embodiment, since the
`
`presence of the shoulder is not important, the distance between the millimeter wave
`
`radar 2 and the road edge is set as the shoulder. In this way, all road arrangements are
`known.
`
`[0039]
`
`In addition, in this example, the detection process of the number of lanes and the lane
`
`width is performed at a rate of, for example, 1 5 minutes. Then, in accordance with the
`
`detected number of lanes and the lane width, the detection of the traveling vehicle shown
`
`in Step S 1 of FIG. 3 is repeated to perform the traffic flow measurement.
`
`[0040]
`
`As described above, in the road traffic monitoring apparatus 1 using the millimeter wave
`
`radar, since the road arrangementis recognized by the apparatus itself and the traffic
`
`monitoring and the traffic flow measurement are performed, it is not necessary to input
`
`the parametersindicating the road arrangement in advance, and it is easy to install the
`
`road traffic monitoring apparatus.
`
`[0041]
`
`Further, since the apparatus itself recognizes the road arrangement,it is possible to cope
`
`with the situation in which the prescribed lane and the position actually running are
`
`forcibly deviated by periodically performing the recognition processing.
`
`[0042]
`
`Accordingly, it is possible to provide a roadtraffic monitoring device which is superior in
`
`the processing capability because it can easily cope with changes in the road
`
`

`

`environment as compared with the case wheretraffic flow measurement is performed on
`
`the basis of data of a fixed road arrangement.
`
`[0043]
`
`Here, a description will be further given of the criteria for the calculation and the
`
`determination in each process described above.
`<How to carve a lane>
`
`First, an array X [] for counting the numberof vehicles is prepared by dividing the vehicle
`
`lane width x-axis direction at regular intervals. Then, on the basis of the calculated x
`
`coordinate of the travel locus, the numberof times is counted in the array X [].
`
`[0044]
`
`For example, when an array with a constant interval of 20 cm and a lane width of 10 m
`
`is prepared, X [50] is obtained. When the travel locus x coordinate is 4.5 m, 1 number of
`
`times are countedin X [22].
`
`FIG. 7 is a diagram illustrating a specific example of lane division based on the
`
`relationship between the lane width direction and the number of vehicles to be traveled.
`
`In this figure, a horizontal axis x indicates a lane width direction, and a vertical axis y
`
`indicates a number count. The figures are 4 lane ways.
`
`In this case, the road data is reproduced, and the number of vehicles is counted for each
`
`lane by using the traveling locus.
`
`[0045]
`
`A graph 13 plotted with black squares indicates a number of counts, and a graph 14
`
`plotted with a white triangle is smoothed by taking a moving average (before division) of
`
`the graph 13.
`
`In view of this figure,
`
`it
`
`is understood that there are 4 peaks in the graph, and it
`
`is
`
`understood that there are 4 lanes, but what is considered here is that, since the number
`
`of running lanesis different for each lane, the driving lane cannot be determined to be a
`
`driving lane when the threshold value is simply prepared and the number count X []
`exceeds it.
`
`[0046]
`
`For this reason, it is judged whether or not the vehicle is a driving lane by the degree of
`
`weight of the data indicating the "horizontal position" as compared with the neighborhood
`data.
`
`To calculate the weight of the data, 2 moving averagesare obtained using the array X [].
`
`[0047]
`
`(1) moving average Xave 1
`
`[] of data to be 1 m including before and after.
`
`

`

`(2) moving average Xave2 [] of data to be 3 m including before and after.
`
`Here, "1 m" is a value for extracting lateral position information in a range narrower than
`
`an actual lateral width of an actual vehicle in order to obtain a peak value of the number
`
`of running as a feature point, and "3 m"is taken into consideration in consideration of a
`
`width of 2.5 m of a large vehicle and is taken into consideration by adding a margin of
`about 0.5 m.
`
`[0048]
`
`Under this setting, in the weight calculation of data, in the range of Xave 2 [], a ratio of
`
`Xave 1
`
`[] is calculated as follows.
`
`Ratio [] = Xave 1
`
`[] / Xave 2 [] x 100 (%)
`
`A portion where the weight (ratio []) calculated by the above calculation exceeds the
`
`threshold value (%)
`
`is determined as 1
`
`lanes. However, when the position of the
`
`threshold value (%) or more is within 2.5 m, it is determined that the vehicle is the same
`
`lane. For example, a case wherea set threshold value = 50% is satisfied.
`
`[0049]
`
`And the ratio[] greatest by the ratio group judged to be one lane is defined as the center
`
`of the lane, and let even middle with the center of the ratio group of adjacent lanes be 1
`lane width.
`
`In addition, while the above description uses the moving average as a 1 way offiltering
`
`for extracting feature points, the present invention is not limited thereto.
`
`[0050]
`
`FIG. 8 is a diagram showing a result of weight calculation performed by an arithmetic
`
`program on the relationship between the vehicle lane width direction and the vehicle lane
`
`width direction.
`
`In this figure, the horizontal axis indicates the lane width direction, and
`
`the vertical axis indicates the weight (%) after 50% subtraction.
`
`[0051]
`
`In FIG. 8,
`
`it is possible to divide the lane by dividing the vertex from the vertex of the
`
`group to the vertex of the adjacent group. Further,
`
`in consideration of calculation
`
`calculation of the center separation band and the road shoulder width, which will be
`
`described later, the road shape can be further recognized.
`
`[0052]
`
`<Median strip>
`
`A boundary portion where the direction of the speed component of the travel locus
`
`switches from positive to negative becomesa central separation band.
`
`The following 2 points are searched to determine a central separation band width.
`
`[0053]
`
`

`

`(1) the x-coordinate xPlus, which is the rightmost in the x-axis direction of the traveling
`
`locus of the component;
`
`(2) the x coordinate of the x coordinate xmin, whichis the leftmost in the x-axis direction,
`
`of the traveling locus of the component;
`
`Under this setting, the central separation band width Cswis obtained as follows.
`
`[0054]
`Csw=xMinus-xPlus
`
`<Road-shoulder width>
`
`Since lane distinction was not completed unless the positional relationship of a millimeter
`
`wave radar and a road was found when carrying out by fixing a road arrangement
`
`parameter like a prior art, it is the distance of a millimeter wave radar and a lane end.
`
`This was necessary. For example, when a millimeter wave radaris installed on a road
`
`shoulder, a road shoulder width parameter is required. As used herein, the "shoulder
`
`width parameter" refers to a width corresponding to a left end of a lane from a millimeter
`
`waveradar position as shownin FIG. 2, rather than an actual shoulder width. Also, when
`
`the millimeter wave radar is installed on the lane, a width corresponding to the left end
`
`of the lane is required from the position of the millimeter wave radar.
`
`[0055]
`
`However,
`
`in the case of recognizing the lane position from the traveling locus and
`
`performing processing as in the present invention, since the positional relationship
`
`between the millimeter wave radar 2 and the road does not need to be conscious, a width
`
`corresponding to the left end of the lane from the position of the millimeter wave radar
`described above does not need to be obtained.
`
`[0056]
`
`On the other hand, if it is determined to be strong, it can be judged that the distance from
`
`the x-coordinate, whichis the leftmost in the x-axis direction, of the traveling locus of the
`
`componentto the origin of the coordinate is the shoulder width.
`
`[Area Setting and Coordinate Transformation]
`
`In the road traffic monitoring device 1 of the millimeter wave radar of this embodiment,
`
`in the course of detecting the detection of the traveling vehicle, a plurality of coordinate
`
`systems are utilized according to the situation and the position calculation is performed.
`
`The reason whya plurality of coordinate systems is used in this wayis as follows.
`
`[0057]
`
`1, the millimeter wave radar 2 performs 1 axis scanning. Accordingly, the output from the
`
`millimeter wave radar is a polar coordinate system (millimeter wave radar polar
`
`

`

`coordinate) output of the distance from the origin position and the scanning angle, taking
`
`the millimeter wave radar installation position at an origin.
`
`[0058]
`
`2, the vehicle number measuring computer 3 has an orthogonal coordinate system
`
`(referred to as road coordinate) matching with a roadline. Thus, it is necessary to convert
`
`the polar coordinate system into road coordinates.
`
`In converting this polar coordinate system into road coordinates,
`
`in this example, the
`
`coordinates are internally converted in the order of "millimeter wave radar polar
`
`coordinates" — "millimeter wave local coordinates" — "millimeter wave world
`
`coordinates" — "road coordinates”. Hereinafter,
`
`these coordinate systems will be
`
`described individually.
`
`[0059]
`
`(a) Millimeter-wave radar polar coordinates
`
`FIG. 9 is a diagram showing a method of taking a millimeter wave radar polar coordinate
`
`and coordinates. Data from the millimeter wave radar 2 is output in a polar coordinate
`
`system as shown in FIG. 9.
`
`[0060]
`
`As an example, the sector center position indicated by the broken line 15 is a direction
`
`in which the millimeter wave radar is oriented when the scanning angle is 0 degrees
`
`(origin). Further, the right side of the broken line from the broken line 15 toward the
`
`millimeter wave radar is a plus direction of an angle, and the left side is a minus direction
`
`of an angle.
`
`[0061]
`
`The coordinate representation at the millimeter wave radar polar coordinates is a
`
`coordinate (distance d, angle 6). This represents the coordinates of the millimeter wave
`
`reflection point, but each point further has information on the speed and the reflection
`
`intensity.
`
`In the computer 3 for measuring the numberof vehicles, processing is performed using
`
`data of a polar coordinate system, and conversion processing of an output format is
`
`performed on the coordinate system of (b) to (d) below just before outputting to the road
`
`coordinate.
`
`[0062]
`
`(b) millimeter wave local coordinates;
`
`As a first step of coordinate transformation to road coordinates, a coordinate
`
`transformation is performed to a millimeter wave local coordinate.
`
`FIG. 10 is a diagram showing a method of taking a millimeter wave local coordinate and
`
`

`

`coordinates. As shown in FIG. 4,
`
`the coordinate conversion is converted into an
`
`orthogonal coordinate with the broken line portion of the polar coordinate of (a) as an
`axis.
`
`[0063]
`
`The transformation at this time,
`
`X= distance dxcostheta
`
`Y= distance dxsintheta
`
`It becomes.
`
`[0064]
`
`(c) Millimeter-wave world coordinates
`
`Since there is a deviation between the broken line 15 (the origin of the millimeter wave
`
`radar angle) shown in FIG. 9 and the traveling direction vector of the road,
`
`this is
`
`converted so as to match the traveling direction vector of the road. The coordinate
`
`system created hereis called a millimeter wave world coordinate.
`
`[0065]
`
`FIG. 11 is a diagram showing a method of taking a millimeter world world coordinate and
`
`coordinates. As shownin FIG. 4, when an angle formed between a traveling direction
`
`vector of a road and a millimeter wave local coordinate is a, (x 1, y 1) of a millimeter
`
`wave local coordinate can be converted into a position (Xw, Yw) in a millimeter wave
`
`world coordinate using the following equation (1).
`
`[0066]
`
`[Mathematical formula 1]
`X,
`QO}
`xX,
`cos@ sing
`
`
`1 0 0
`1)
`4)
`
`Y, (=| 7sin@ cos@ Olly,|cerees (1)
`
`[0067]
`
`(d) Road coordinates
`
`A computer 3 for measuring the number of vehicles has an orthogonal coordinate system
`
`and road coordinates matched with a road line. The relationship between the road
`
`coordinate and the millimeter wave world coordinate is different for each installation
`
`con