JP2000028713A - Vehicle rear side monitoring method and vehicle rear side monitoring device - Google Patents

Abstract

(57) [Summary] [PROBLEMS] To accurately monitor a parallel traveling vehicle by removing a non-parallel vehicle such as a stationary object on a road shoulder. SOLUTION: The output of a laser radar 2 for monitoring a predetermined monitoring area on the rear side of a truck 1 and the outputs of a gyro 4 and a vehicle speed sensor 5 for calculating a running locus of the truck 1 are represented by E.
When received by the CU 3, the ECU 3 removes, as noise, the plurality of parallel vehicle candidates in the coordinate system near the transition traveling locus, and estimates the current position of the parallel vehicle candidate after noise removal. In addition, a function for calculating the speed and the moving direction of the parallel traveling vehicle is specified by specifying the position of the parallel traveling vehicle at this time, and the stationary object on the road shoulder is efficiently removed, and the speed and the traveling direction of the parallel traveling vehicle are accurately recognized.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a rear-side monitoring device and a rear-side monitoring method for monitoring the status of a parallel vehicle with respect to the rear side of the vehicle.

[0002]

2. Description of the Related Art When making a left turn or a lane change in a vehicle including a large vehicle such as a truck, a driver checks the presence of a parallel vehicle by checking the rear side with a mirror. At this time, the driver can overlook the parallel vehicle by checking with the mirror,
When it is difficult to confirm the rear side by the mirror at night or in rainy weather, the recognition of the parallel traveling vehicle may be insufficient.

[0003]

Therefore, there is a demand for a technique capable of monitoring the status of a parallel traveling vehicle with respect to the rear side of the vehicle without looking at the driver. As a device for detecting objects around the vehicle, for example, Japanese Patent Application Laid-Open No. Hei 7-229961 discloses a method in which the distance and direction of an obstacle is detected by a radar when the vehicle moves backward, and the obstacle and the vehicle are displayed on a display. There is disclosed a technology for clearly recognizing the positional relationship between an obstacle and a host vehicle. However, when such a technique is directly applied to monitoring the situation of a parallel vehicle with respect to the rear side when the vehicle is traveling forward, it is not possible to distinguish between a stationary object on the roadside and the parallel vehicle, and the parallel vehicle is not accurately determined. Is impossible to monitor.

The present invention has been made in view of the above circumstances, and a rear side monitoring device and a rear side monitoring method capable of accurately monitoring a parallel traveling vehicle by removing a non-parallel vehicle such as a stationary object on a road shoulder. The purpose is to provide.

[0005]

According to the present invention, there is provided a rear side monitoring method for monitoring a predetermined monitoring area on a rear side of a vehicle. By converting to a coordinate system consisting of directional axes and detecting parallel traveling vehicle candidates, the traveling locus of the own vehicle is transited in the vehicle width direction axial direction in the coordinate system, and multiple concurrent traveling vehicle candidates in the coordinate system are transition traveling locus In the case of being in the vicinity of the vehicle, those parallel vehicle candidates are removed as noise, the current position of the parallel vehicle candidate after noise removal is estimated from the position, moving direction, and speed of the parallel vehicle detected last time. The current parallel vehicle position is specified based on the moving direction from the position to the position of the candidate parallel vehicle after noise removal and the previously detected moving direction, and the estimated current position and the position of the candidate parallel vehicle after noise removal. Of the parallel running vehicle Calculate the velocity and direction of movement of the adjacent parallel vehicle based on the position of the adjacent parallel vehicle detected location and time, the shoulder of the stationary object such as to efficiently remove accurately recognize the speed and direction of movement of the adjacent parallel vehicle.

In order to achieve the above object, a rear side monitoring apparatus according to the present invention comprises: an output of a radar apparatus for monitoring a predetermined monitoring area on a rear side of a vehicle; and a travel path calculation for calculating a travel path of the vehicle. When the output of the means is received by the control means and the parallel traveling vehicle on the rear side of the vehicle is monitored, the output of the radar device is converted into a coordinate system consisting of the vehicle longitudinal axis and the vehicle width direction axis by the function of the control means. In addition to detecting the parallel traveling vehicle candidate, the traveling locus of the own vehicle is transited in the vehicle width direction in the coordinate system, and when a plurality of parallel traveling vehicle candidates in the coordinate system are in the vicinity of the transition traveling locus, the parallel traveling is performed. The vehicle candidate is removed as noise, the current position of the parallel vehicle candidate after noise removal is estimated from the position, moving direction and speed of the parallel vehicle detected last time, and the parallel vehicle after noise removal from the position of the parallel vehicle detected last time To the position of the candidate Based on the moving direction of the vehicle and the previously detected moving direction, and the estimated current position and the position of the parallel vehicle candidate after noise removal, the current parallel vehicle position is specified, and the position of the previously detected parallel vehicle and the current parallel vehicle detected By calculating the speed and moving direction of the parallel traveling vehicle based on the position, the stationary object on the road shoulder is efficiently removed, and the speed and the traveling direction of the parallel traveling vehicle are accurately recognized.

[0007]

FIG. 1 is a block diagram showing a schematic configuration of a vehicle rear side monitoring device according to an embodiment of the present invention. FIG. 2 is a schematic diagram showing a detection status of a laser radar.
FIG. 3 shows a process description of a vehicle rear side monitoring method according to an embodiment of the present invention for recognizing a parallel vehicle.

As shown in FIG. 2, a laser radar 2 as a radar device for irradiating a laser beam toward a predetermined monitoring area behind the vehicle is provided on a side of a driver's seat of a truck 1 as a vehicle. 2 scans, for example, in a horizontal 12-degree range, irradiates a laser beam toward a monitoring area in a horizontal 12-degree range and a vertical 4-degree range, and detects a candidate for a parallel vehicle. ing.

As shown in FIG. 1, detection information of the laser radar 2 is inputted to an ECU 3 as a control means of the truck 1 and a gyro 4 as a travel locus calculation means.
And information of the vehicle speed sensor 5 is input. The ECU 3 also has a function of calculating the traveling locus of the truck 1 based on information from the gyro 4 and the vehicle speed sensor 5, and constitutes a traveling locus calculating means together with the gyro 4 and the vehicle speed sensor 5.

The ECU 3 detects candidate parallel vehicles based on the detection information of the laser radar 2 and the input information of the gyro 4 and the vehicle speed sensor 5, and also uses a reflective object representing a road shape such as a reflective tape of a guard rail or a reflector from the candidate parallel vehicles. Is removed, and the speed and moving direction of the parallel traveling vehicle are calculated.

[0011] The information of the turn signal 6 is input to the ECU 3, and when the intention of the course change is determined based on the information of the turn signal 6, an alarm 7 is issued when a parallel vehicle exists at the course change destination. It has become. In the illustrated example, the alarm 7 is issued based on the information of the turn signal 6. However, the warning 7 is issued by determining the intention to change the course in accordance with information such as the deceleration of the truck 1, the steering wheel angle or the lateral acceleration. You may make it emit. Further, instead of the alarm 7, it is possible to display and recognize the presence of the parallel vehicle by a display means such as a lamp or an image, or to recognize the presence of the parallel vehicle by voice.

As shown in FIG. 3, the ECU 3 converts the output of the laser radar 2 into a coordinate system (xy coordinate system) composed of a vehicle longitudinal axis (x axis) and a vehicle width direction axis (y axis). Xy coordinate conversion (S1: parallel vehicle candidate detection step) for detecting parallel vehicle candidates, and the detected detection points are divided into a plurality of groups (parallel vehicles) by adjacent detection points in the xy coordinate system. (S2). Next, the trajectory of the own vehicle is transited in the vehicle width direction axial direction in the xy coordinate system, and when a plurality of groups in the xy coordinate system are near the transition traveling trajectory, the plurality of groups are changed. It is removed as a non-parallel running vehicle (noise) (S3: noise removal step).

After removing the non-parallel vehicles, the current position of the parallel vehicles (the center of gravity of the group of parallel vehicles) is estimated from the position, moving direction, and speed of the parallel vehicles detected last time (S4: current position estimation step). ). After estimating your current location,
A parallel vehicle that is the same as the previously detected parallel vehicle is determined from the parallel vehicle candidates after noise removal (S5). That is, the angle difference between the moving direction from the position of the parallel vehicle detected last time to the position of the candidate parallel vehicle after noise removal and the previously detected moving direction, and the position of the candidate parallel vehicle after noise removal from the estimated current position. Based on the distance to the vehicle, the same vehicle as the previously detected parallel vehicle is specified from among the parallel vehicle candidates after noise removal, and is set as the current parallel vehicle detection result (parallel vehicle detection step).

After detecting the traveling vehicle, the traveling direction and traveling speed of the traveling vehicle are calculated from the position of the traveling vehicle (center of gravity) detected last time and the position of the traveling vehicle (center of gravity) specified and detected this time. (S6). The traveling speed of the parallel traveling vehicle is calculated based on the distance since the detection cycle of the laser radar 2 is constant.

By executing the above-mentioned steps S1 to S6, a candidate for a parallel vehicle is detected, and information on a non-parallel vehicle that is a reflector representing a road shape such as a reflective tape or a reflector of a guardrail is efficiently removed from the candidate parallel vehicle. Then, the speed and moving direction of the traveling vehicle are accurately calculated. Details of the processing of S1 to S6 will be described based on FIGS.

The xy coordinate conversion (S1) will be described with reference to FIG. FIG. 4 shows the concept of xy coordinate conversion.
As shown in the figure, the xy coordinates (px, py) of the vehicle longitudinal direction axis (x axis) and the vehicle width direction axis (y axis) of the parallel vehicle candidate 11 (detection point) detected by the laser radar 2 are: , Laser radar 2
Px = Lcos θ, py
= L sin θ.

A method of grouping a plurality of parallel traveling vehicle candidates (S2) will be described with reference to FIGS. FIG.
FIG. 6 shows a flowchart for executing detection point grouping, FIG. 6 shows a setting state of the grouping range, and FIG. 7 shows a setting state of the grouping range update.

As shown in FIG. 5, in step S11, it is determined whether the grouping flag of the first detection point (for example, P1) is 0 or not. If the grouping flag is 1, it means that grouping has been performed with another detection point, so the grouping flag is 0 at the first detection point, and step S1 is performed.
In step 1, the grouping flag is determined to be 0, and in step S12, the grouping flag of the detection point (for example, P1) is set to 1
Set to.

After setting the grouping flag of the detection point P1 to 1 in step S12, the initial state of the grouping range of the first detection point is set in step S13. That is, FIG.
As shown in the figure, a rectangle (2dx × 2dy) is set in the range of dx (for example, 2 m) in the x-axis direction and dy (for example, 2 m) in the y-axis direction around the detection point (for example, P1). The grouping range 21 in P1 is set. After setting the initial grouping range 21, the process moves to the next detection point (for example, P2) in step S14.

In step S15, the next detection point (for example,
It is determined whether the grouping flag of P2) is 0 or not. Since the grouping is not determined at the next detection point P2, the grouping flag is determined to be 0. If it is determined in step S15 that the grouping flag is 0, the process proceeds to step S15.
At 16, it is determined whether the next detection point (for example, P2) is within the grouping range 21 at the detection point P1. If it is determined in step S16 that the detection point P2 is within the grouping range 21 at the detection point P1,
At 17, the grouping range 21 is updated to the grouping range 22 at the detection points P1 and P2.

That is, as shown in FIG. 7, when the detection point P2 exists within the grouping range 21 at the detection point P1,
A square (2dx × 2dy) with a range of dx in the x-axis direction and a range of dy in the y-axis direction around the detection point P2 is set as a grouping range 21a of the detection point P2, and a grouping range 21a of the detection point P2. And the grouping range 21 of the detection point P1 are superimposed and updated as a new grouping range 22 in step S17.

In step S17, a new grouping range 2
When 2 is updated, the grouping flag of the detection point P2 is set to 1 in step S18, and the center-of-gravity coordinates P0 of the grouping range 22 is calculated and updated in step S19. Next, in step S20, it is determined whether or not the detection point P2 is the final detection point. Since there are a large number (for example, n) of normal detection points, the process proceeds to the next detection point (for example, P3) in step S14. S
The processing from step 15 to step S20 is repeated.

In step S15, the next detection point (for example, Pn)
If the grouping flag is not 0, that is, if it is determined that the grouping has already been performed with another detection point, the process proceeds to step S20. If it is determined in step S16 that the next detection point is not within the set grouping range, the process proceeds to step S20.

If it is determined in step S20 that the detected point is the final detected point (for example, Pn), the process proceeds to the first detected point P1 in step S20, and proceeds to the first detected point P1 in step S11.
It is determined whether or not the grouping flag is 0. At this time,
Since the grouping flag of the detection point P1 is 1, the process proceeds to step S22 to determine whether or not the grouping flags of all the detection points P1 to Pn are 1, that is, the grouping is performed at all the detection points P1 to Pn. Is determined. Step S
If it is determined at 22 that grouping has been determined for all of the detection points P1 to Pn, the grouping process ends.

In step S22, the grouping flags of all the detection points P1 to Pn are not 1, ie, all the detection points P1 to Pn
If it is determined that grouping has not been determined in Pn, the process proceeds to the next detection point in step S23, and step S1
The process proceeds to 1 and the grouping is determined for detection points that are not grouped based on the next detection point, and the above-described processing is repeated until the grouping flag is set to 1 at all the detection points P1 to Pn. In this way, the detection points P1 to Pn are grouped as the parallel vehicle candidates of the plurality of groups GP1 to GPn.

For example, when the detection points P1 to P5 exist, the grouping flag of the detection point P1 is first set to 1, and the detection points P2 to P5 are sequentially determined to be grouped based on the detection point P1. The grouping flag is set to 1 for the grouped detection points. Detection point
After the processing based on P1 is completed, the process returns to the first detection point. However, the detection point P1 proceeds to the next detection point P2 because the grouping flag is already set to 1. If the grouping flag of the detection point P2 is 0, the grouping flag of the detection point P2 is set to 1 and the determination of grouping is performed on the remaining detection points whose grouping flag is 0 based on the detection point P2. Is performed.

After the grouping determination process based on the detection point P2 is completed, or the grouping flag of the detection point P2 is set to 1 in the grouping determination process based on the detection point P1. If it is, the grouping flags of the detection points P1 and P2 are set to 1, and it is determined whether the grouping flag of the detection point P3 is set to 1 and the grouping flag of the detection point P3 is determined. If is not set to 1, the grouping flag of the detection point P3 is set to 1 and a process of determining grouping based on the detection point P3 is performed. Such processing is repeated, and all the detection points P1
When the grouping flags of .about.P5 are set to 1, the grouping process ends.

Next, with respect to the xy coordinate system of the plurality of groups GP1 to GPn set in the above grouping process,
The trajectory of the own vehicle is shifted in the vehicle width direction in the xy coordinate system, and groups GP1 to GP
If there are a plurality of groups in n, the groups should be recognized as non-parallel vehicles (noise) that are reflective objects that represent the road shape, such as reflective tapes and reflectors on guardrails, so as not to be mistaken as parallel vehicles. Remove.

A method of removing a non-parallel running vehicle as noise (S3) will be described with reference to FIGS. FIG.
FIG. 9 shows a concept for explaining a method of removing a non-parallel running vehicle. In the illustrated example, 4 is set in the grouping process.
An example is shown in which non-parallel vehicles are determined and removed from the groups GP1 to GP4.

As shown in the figure, the traveling locus of the truck 1 is calculated based on the information of the gyro 4 and the vehicle speed sensor 5,
For example, it is assumed that a trajectory for the past 40 m is a trajectory S obtained by performing a three-dimensional curve approximation. The locus S approximated to the curve is transited in the vehicle width direction axial direction (y-axis direction: arrow direction in the drawing) in the xy coordinate system to be a transition locus T, and the transition locus T is divided into four groups GP1 to GP.
Overlay one of the four. The distance between the center of gravity of each of the groups GP1 to GP4 and the transition trajectory T is calculated. If there are three or more groups GP1 to GP4 in which the distance between the center of gravity to the transition trajectory T is within a threshold value (for example, 0.5 m), Are removed as non-parallel vehicles.

That is, as shown in FIG. 8, when the transition locus T is adjusted to the center of gravity of the group GP1, the other group GP2
The center of gravity of ~ GP4 is far from the transition locus T, and
There is no group whose distance to the transition locus T is within the threshold. As a result, it is recognized that the group GP1 is not a non-parallel vehicle, that is, a group vehicle.

As shown in FIG. 9, when the transition trajectory T is set to the center of gravity of the group GP2, the groups GP3 and GP4
Is within the threshold of the transition locus T, and the group whose distance to the transition locus T is within the threshold is group GP2.
And three in total. As a result, the groups GP2 to GP
4 is presumed to be a reflector or the like existing substantially along the traveling locus of the truck 1, and the groups GP2 to GP4 are recognized as non-parallel vehicles and removed.

As described above, the groups GP2 to GP4 are recognized and removed from the plurality of groups GP1 to GP4 as non-parallel vehicles, and the group GP1 is recognized as a parallel vehicle. After removing groups GP2 to GP4 as non-parallel vehicles, the group, position, direction and speed of the
The current position (centroid position) of GP1 is estimated (S4). That is, the center of gravity of the current position of the current parallel vehicle (group GP1) is estimated from the center of gravity, the moving direction, and the moving speed of each group GP recognized as the parallel vehicle last time.

After estimating the current position of the center of gravity of the parallel vehicle this time, it is determined that the parallel vehicle detected last time and the parallel vehicle estimated this time are the same parallel vehicle (S5). FIG.
A method for determining the same parallel vehicle based on 0 will be described. FIG. 10 shows a concept for explaining a method of determining the same parallel traveling vehicle.

As shown in the figure, each group (for example, groups GPa, GPa) set this time in the grouping process
b), the difference from the estimated position of the center of gravity (d-dist-a, d-dist
-b) and the difference in movement direction (d-dir-a, d-dir-b),
(iden-rate) is calculated by the following equation (1), and it is determined whether the estimated group is the same as the previous group.
That is, the object having the highest degree of identity is specified as the same object. (iden-rate) = (dist-rate) * k + (dir-rate) ・ ・ ・ (1) where k is, for example, a weighting coefficient set to about 1.5, and (dist-rate) is 1.0- In (d-dist) / (dist-th), (d
ist-th) is, for example, 2m, (dir-rate) is 1.0- (d-di
r) / 180.

The reason why the difference in the moving direction (d-dir) in addition to the difference (d-dist) from the position of the center of gravity is added to the determination as to whether or not the objects are the same will be described with reference to FIG. FIG. 11 shows a concept in the case where the same parallel traveling vehicle is determined only by the difference from the position of the center of gravity.

As shown in the figure, each of the current groups (eg, groups GPa, GPa) set in the grouping process
If the movement directions of b) are opposite to each other, the group
This group GPa, GP with respect to the center of gravity of the estimated position of GPa
Looking at the difference (d-dist-a, d-dist-b) in the position of the center of gravity of b, the difference (d-dist-b) from the group GPb is the difference (d-dist-b) from the group GPa.
-dist-a). In such a case, if the same parallel vehicle is determined based only on the difference (d-dist-a, d-dist-b) of the center of gravity, the current group GPb becomes the previous group GPb.
It is determined that they are the same as a. Movement direction difference (dd
By adding ir), it is determined that the current group GPa is the same as the previous group GPa.

Therefore, by taking into account the difference (d-dir) in the moving direction, the position of the same parallel traveling vehicle can be specified with high accuracy.

After determining the same target as the previous group GP for each of the current group GPs, the current group GP (the same target as the same target) is determined from the center of gravity position of the previous group GP and the center of gravity of the current group GP. Then, the moving direction and the moving speed of the parallel traveling vehicle are calculated (S6). The traveling speed of the traveling vehicle is
Since the detection cycle of the laser radar 2 is constant, the speed is calculated based on the distance.

With the above processing, the information of the actual parallel vehicle can be specified accurately from the parallel vehicle candidates detected by the laser radar 2 without erroneous recognition. The alarm 7 can be reliably issued when is present.

In the vehicle rear side monitoring device and the vehicle rear side monitoring method described above, the detection information of the laser radar 2 for monitoring a predetermined monitoring area on the rear side of the truck 1 is converted into an xy coordinate system. To determine the traveling locus S of the own vehicle from the output information of the gyro 4 and the vehicle speed sensor 5,
The traveling locus is shifted in the y-axis direction in the xy
When a plurality of parallel vehicle candidates in the y coordinate system are near the transition trajectory T, these parallel vehicle candidates are removed as noise. Objects and the like can be efficiently removed.

In the above-described vehicle rear side monitoring device and the vehicle rear side monitoring method, the moving direction from the position of the preceding parallel vehicle to the position of the candidate for the parallel vehicle after removing noise and the previous parallel running Since the current position of the parallel vehicle is specified based on the angle difference from the moving direction of the vehicle and the distance from the estimated current position of the parallel vehicle to the position of the parallel vehicle candidate after noise removal, Thus, it is possible to accurately specify the current position of the same parallel vehicle as the previous time, and at the same time, it is possible to accurately calculate the speed and the moving direction.

[0043]

According to the method for monitoring the rear side of a vehicle and the rear side monitoring device of the vehicle according to the present invention, the output of a radar device for monitoring a predetermined monitoring area on the rear side of the vehicle is controlled by using a vehicle front-rear axis and a vehicle width direction. Convert to a coordinate system consisting of axes and detect parallel vehicle candidates,
Since the trajectory of the own vehicle is shifted in the vehicle width direction in the coordinate system and non-co-running vehicles are removed as noise according to the transition trajectory, stationary road shoulders arranged substantially corresponding to the trajectory of the own vehicle Objects and the like can be efficiently removed.

Further, the current parallel vehicle position is specified based on the moving direction to the parallel vehicle candidate after noise removal and the previously detected moving direction, and the estimated current position and the position of the parallel vehicle candidate after noise removal. Since the speed and moving direction of the vehicle are calculated, the current position of the same parallel vehicle as the previous one can be accurately specified from among the candidate vehicles, and the speed and moving direction of the parallel vehicle can be accurately recognized. can do.

As a result, it is possible to accurately monitor the parallel running vehicle by removing the non-parallel running vehicle such as a stationary object on the road shoulder.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a schematic configuration of a vehicle rear side monitoring device according to an embodiment of the present invention.

FIG. 2 is a schematic diagram showing a detection situation of a laser radar.

FIG. 3 is a process explanatory view of a method for monitoring a rear side of a vehicle according to an embodiment of the present invention for recognizing a parallel traveling vehicle;

FIG. 4 is a conceptual diagram of xy coordinate conversion.

FIG. 5 is a flowchart for executing detection point grouping.

FIG. 6 is an explanatory diagram of a setting status of a grouping range.

FIG. 7 is an explanatory diagram of a setting status of grouping range update.

FIG. 8 is a conceptual diagram illustrating a method for removing a non-parallel running vehicle.

FIG. 9 is a conceptual diagram illustrating a method of removing a non-parallel running vehicle.

FIG. 10 is a conceptual diagram illustrating a method of determining the same parallel traveling vehicle.

FIG. 11 is a conceptual diagram in the case where the same parallel running vehicle is determined only based on the difference from the position of the center of gravity.

[Explanation of symbols]

 Reference Signs List 1 track 2 laser radar 3 ECU 4 gyro 5 vehicle speed sensor 6 turn signal 7 alarm 11 parallel running candidate S running locus T transition locus

Claims (2)

[Claims] 1. A parallel vehicle candidate detection system for converting an output of a radar device for monitoring a predetermined monitoring area on the rear side of a vehicle into a coordinate system including a vehicle longitudinal direction axis and a vehicle width direction axis to detect a parallel vehicle candidate. Steps, the traveling locus of the own vehicle is transited in the vehicle width direction axial direction in the coordinate system, and when a plurality of parallel traveling vehicle candidates in the coordinate system are near the transition traveling locus, those parallel traveling vehicle candidates are regarded as noise. A noise removal process for removing, a current position estimation process for estimating a current position of a candidate for a parallel vehicle after noise removal from a position, a moving direction, and a speed of the parallel vehicle detected last time; The moving direction up to the position of the parallel vehicle candidate and the moving direction detected last time,
And a parallel vehicle detection step of specifying the current parallel vehicle position based on the estimated current position and the position of the parallel vehicle candidate after noise removal, based on the position of the parallel vehicle detected previously and the position of the parallel vehicle detected this time. A parallel vehicle status calculating step of calculating a speed and a moving direction of the parallel vehicle, and a rear side monitoring method of the vehicle, comprising: 2. A radar device for monitoring a predetermined monitoring area on a rear side of a vehicle, a traveling locus calculating means for calculating a traveling locus of the vehicle, and an output of the radar device and an output of the traveling locus calculating means. Control means for receiving and monitoring a parallel vehicle on the rear side of the vehicle, wherein the control means converts the output of the radar device into a coordinate system consisting of a vehicle longitudinal axis and a vehicle width direction axis. In addition to detecting concurrently traveling vehicle candidates, the traveling locus of the own vehicle is shifted in the vehicle width direction in the coordinate system, and when a plurality of concurrently traveling vehicle candidates in the coordinate system are near the transition traveling locus, those concurrently traveling vehicle candidates are Is removed as noise, the current position of the parallel vehicle candidate after noise removal is estimated from the position, moving direction, and speed of the parallel vehicle detected last time, and the parallel vehicle candidate after noise removal is estimated from the position of the parallel vehicle detected last time. Up to position Based on the moving direction and the previously detected moving direction, and the estimated current position and the position of the parallel vehicle candidate after the noise removal, the current parallel vehicle position is specified, and the position of the parallel vehicle detected previously and the position of the parallel vehicle detected this time are determined. A rear side monitoring device for a vehicle, comprising a function of calculating a speed and a moving direction of a parallel traveling vehicle based on a position.