WO1997031301A1

WO1997031301A1 - Obstacle-detection mapping method for unmanned vehicle and obstacle detection apparatus

Info

Publication number: WO1997031301A1
Application number: PCT/JP1997/000448
Authority: WO
Inventors: Tsugio Sudo; Akiharu Nishijima; Yukio Okawa
Original assignee: Komatsu Ltd.
Priority date: 1996-02-21
Filing date: 1997-02-19
Publication date: 1997-08-28
Also published as: JPH09230936A; AU1810697A

Abstract

A mapping method by which an obstacle is detected without an increase of the computing load of a controller and deterioration of precision. A region including a travel course (6) is divided into a large number of areas with straight lines parallel to the coordinate axes x, y at predetermined intervals. The areas through which the travel course (6) passes are stored. When the area where a detected obstacle (2) is positioned coincides with at least one of the stored passage areas at the time of automatic travelling, the obstacle is judged to be present on the travel course (6).

Description

Description Mapping method and obstacle detection device for obstacle detection of unmanned vehicles

Technical field

The present invention relates to a mapping method and a detection device for detecting an obstacle on a traveling course of an unmanned vehicle. Background technology

Conventionally, there have been many proposals for an unmanned vehicle traveling system capable of unmanned traveling of a vehicle along a predetermined traveling course. At quarry sites over a wide area, an unmanned dump truck operation system that transports sediment using unmanned vehicles, for example, unmanned dump trucks, is well known. This unmanned dump truck operation system is, for example, as follows. The traveling course of the unmanned dump truck is taught by a predetermined method in advance, and coordinate data for each predetermined distance or time on the traveling course is stored in the storage device as traveling course data. At the time of automatic traveling, the unmanned dump truck detects and confirms the current position where the vehicle is actually traveling by the position detecting means at every predetermined sampling time, and checks the actual traveling position and the previously stored traveling course data. Is calculated. Then, in order to reduce the deviation, steering control, unmanned dumping such as steering, vehicle speed, starting and stopping, and the like are performed, and the unmanned dumping is controlled so as to travel along a traveling course stored in advance. is there.

Usually, an obstacle detecting means is provided at the front end of such an unmanned dump truck to detect the presence or absence of an obstacle on the traveling course. As the obstacle detection means, for example, a laser radar is well known. This laser radar measures the reflection propagation time until the laser beam transmitted to the obstacle is reflected from the obstacle.

It detects the distance to obstacles. Fig. 8 shows the conventional method of detecting obstacles on an unmanned dumping course using such obstacle detection means. You. Each point on the traveling course 6 of the unmanned dump 1 is stored in the storage unit (not shown) of the unmanned dump 1 as coordinates on the traveling course coordinate system X-y. In addition, the controller (not shown) of the unmanned dump truck 1 uses its position detecting means (not shown) to determine its own current position as coordinate data on the traveling course coordinate system X-y. To detect . Based on the current position data, the traveling direction D of the vehicle, and the detected obstacle distance L, the controller calculates coordinate data of the obstacle 2 on the traveling course coordinate system X-y. Calculate. Based on the calculated coordinate data of the obstacle 2 and the coordinate data of each point on the traveling course 6, the distance between the two points is sequentially calculated. When the calculated distance is within the predetermined value, it is determined that the obstacle 2 is in the vicinity of the traveling course 6, and a predetermined process such as a stop or a detour is performed to evacuate the obstacle 2. ing.

However, in the above obstacle detection device, the distance between the obstacle 2 and each point on the traveling course 6 is sequentially calculated, and it is determined whether or not this distance is within an allowable value. The computational load on the rollers increases significantly. In addition, it is necessary to perform this calculation simultaneously with other calculations for the automatic cruise control, which further increases the calculation load. Therefore, in order to perform automatic cruise control and obstacle detection without any problem, these calculations must be processed within a predetermined time. This requires the use of a high-speed computer, which makes the controller expensive. In addition, since the distance between the obstacle 2 and each point on the traveling course 6 is sequentially calculated, there is substantially a lot of dead time until the final determination of the presence or absence of the obstacle 2. That is, there is a problem that it is not efficient to process the obstacle detection in a real time. Disclosure of the invention

The present invention has been made in order to solve the problems of the related art, and it has been made possible to detect an obstacle without increasing a calculation load on a controller, and to perform an obstacle detection without reducing accuracy. An object of the present invention is to provide a mapping method for obstacle detection of an unmanned vehicle capable of detecting an object and an obstacle detection device.

The mapping method for detecting an obstacle in an unmanned vehicle according to the present invention includes: In the mapping method for detecting an obstacle of an unmanned vehicle, which detects an obstacle on a traveling course in which the unmanned vehicle travels automatically,

The area including the traveling course is divided by a predetermined distance with a straight line parallel to the coordinate axes X and y to form a number of areas, and the area through which the traveling course passes is stored,

In automatic driving, if at least one of the memorized passing areas and the area where the detected obstacle is located matches, the obstacle is determined to be on the traveling course. are doing.

According to this configuration, the traveling area is divided at predetermined distances by straight lines parallel to the two coordinate axes x and y, and a large number of areas are formed surrounded by adjacent straight lines parallel to the X and y directions. . When the driving course is mapped to this area map, the area where the driving course passes is stored. Then, the position of the obstacle detected during the automatic travel of the unmanned vehicle is detected, and when the area where the obstacle is located coincides with any one of the areas where the traveling course passes, the obstacle is detected. It is determined that the vehicle is on the traveling course.Therefore, it is only necessary to determine whether or not the traveling course is mapped in the area where the obstacle position is mapped, so that the calculation time required for the determination is short. . Therefore, since the calculation load can be reduced, the calculation speed of the controller does not have to be high and the controller does not have to be expensive. Furthermore, since there is ample time for real-time control processing, other automatic control processing can be executed without fail.

Further, it is desirable that the section of the curved section of the traveling course is a section that is finely divided from the section of the straight section of the traveling course.

With this configuration, in the curve section, the width of the traveling course in the direction perpendicular to the traveling direction becomes narrow, and an obstacle located near the curve can be accurately detected. This eliminates the possibility that an obstacle near the traveling course is erroneously determined to be on the traveling course, and eliminates the problem that the unmanned vehicle stops by mistake. Therefore, the operating rate of unmanned vehicles is improved.

An obstacle detection device for an unmanned vehicle according to the present invention,

Obstacle detection of unmanned vehicles that detects obstacles on the traveling course where unmanned vehicles run automatically In the device,

Obstacles that store a map composed of a large number of areas by dividing the area containing the driving course by a predetermined distance with a straight line parallel to the coordinate axes X and y, and storing the memory area corresponding to these areas Storage area map for detection

Creates an area map that calculates the coordinate data of points at predetermined distances on the traveling course and writes the specified data including the calculated coordinate data to the memory corresponding to the area that includes the position of the coordinate data Department and

An obstacle position calculation unit that calculates the position of the obstacle;

The calculated obstacle position is mapped to the memory area, and at least one of the areas including the mapped obstacle position and the coordinate data position written to the memory area is mapped. An obstacle determining unit that determines that the calculated obstacle is on the traveling course when the two areas meet.

According to this configuration, each area surrounded by adjacent straight lines in the X and y directions is converted into an area map for obstacle detection configured by memory areas corresponding to each area. To memorize. The area map creation unit writes the predetermined data (map number or 0 or 1 representing mapping correspondence) into the memory area corresponding to the calculated coordinate data, thereby obtaining a travel command. Map to the obstacle detection area map storage unit. Then, the obstacle determination unit stores the memory in the obstacle detection area map storage unit corresponding to the position of the obstacle calculated by the obstacle position calculation unit, and corresponds to each point on the traveling course. If any one of the memories in the obstacle detection area map storage unit matches, the obstacle is determined to be on the traveling course. Accordingly, the controller requires only a short calculation time for the determination, reduces the calculation load, and allows time margin, as in the configuration of the above method.

The maps to be stored are: i) a wide-area map that divides an area including a traveling course by a predetermined distance with a straight line parallel to the coordinate axes X and y to form a large number of areas; A detailed map that defines the area as an area that is more subdivided than the area of the wide area map, and iii) the area of the curve section of the driving course And a force map formed in an area that is finely divided. According to this configuration, the obstacle detection area map has three types of maps: a wide area map, a detailed map, and a curve map. The wide area map represents, for example, the outline of the position where the traveling course passes in the entire traveling coordinate system X-y. The detail map shows each area of the wide area map through which the traveling course passes in more detail. In the curve map, only the curve section is represented by a finer area. Since only the necessary area is represented as a more detailed area, the memory in the obstacle detection area map storage unit can be configured with a small capacity and can be used efficiently. BRIEF DESCRIPTION OF THE FIGURES

FIG. 1 is a functional configuration block diagram of the obstacle detection device according to the present invention.

Fig. 2 is an explanatory diagram of a wide area map of the area map storage unit for obstacle detection according to the present invention.

FIG. 3 is an explanatory diagram of a detailed map of the obstacle detection area map storage unit according to the present invention.

FIG. 4 is a flowchart for creating a detailed map and a curve map according to the present invention.

FIG. 5 is an explanatory diagram of a detailed map creation procedure according to the present invention.

FIG. 6 is an explanatory diagram of the next step in the detailed map creation procedure according to the present invention. FIG. 7 is an explanatory diagram of a force map generation procedure according to the present invention.

FIG. 8 is an explanatory diagram of an obstacle detection method according to the related art. BEST MODE FOR CARRYING OUT THE INVENTION

Preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

In FIG. 1, a controller 10 is a general computer system mainly composed of, for example, a micro computer. [Controller] 0 Storage section], an area map creation section 13, an obstacle detection area map storage section 12, an obstacle position calculation section 14, and an obstacle determination section 15. . Code storage unit 1 stores position data of each point at a predetermined distance on the traveling course, and each position data is a traveling coordinate system _x- representing an area where the unmanned dump truck 1 automatically travels. It is stored as each coordinate data at y. The obstacle detection area map storage unit 12 (hereinafter referred to as “obstacle map storage unit 12”) forms an obstacle detection area map corresponding to the traveling coordinate system X-y space and stores it. Is stored. That is, the traveling coordinate system X-y space is divided at predetermined intervals by a straight line parallel to the two coordinates x and y, and the memo corresponding to each area surrounded by the adjacent parallel lines in the X and y directions. It stores the area map for obstacle detection configured by the criterion. The obstacle map storage unit 12 is composed of rewritable memory, for example, a RAM, a floppy disk, a hard disk, and the like.

The area map creating section 13 maps the traveling course 6 to the obstacle detection area map. That is, based on the coordinate data of each point of the traveling course 6 stored in the course data storage unit 11, the area map creating unit 13 generates the coordinate data of the points on the traveling course 6 at every predetermined distance. Calculate, and write a predetermined data to each memory area (corresponding to each area) in the obstacle map storage unit 12 corresponding to the calculated coordinate data. Further, the obstacle position calculation unit 14 calculates the position of the obstacle 2 based on the current position data of the unmanned dump 1, the distance data to the obstacle 2 and the traveling direction data of the unmanned dump 1. . Then, the obstacle determination unit 15 maps the position of the obstacle 2 calculated by the obstacle position calculation unit 14 to each bit in the obstacle map storage unit 12, and It is determined whether the data and one of the bits mapped to the run τ course 6 match. If they match, it is determined that the obstacle 2 is on the traveling course 6.

The position detection means 21 detects the current position of the unmanned dump truck 1 in the traveling course coordinate system, and outputs this position data to the obstacle position calculation unit 14. As the position detecting means 21, for example, a means for detecting an absolute coordinate position by a GPS system or the like may be used. Also The position detecting means 21 is based on traveling direction data (angle data) detected by a gyro or the like and traveling distance data detected by the number of times of a wheel or the like, and determines a relative position from a known reference position. May be obtained by calculation. The obstacle distance detecting means 22 detects the distance to the obstacle 2 and outputs the distance data to the obstacle position calculating section 14 and is constituted by, for example, a laser radar or the like. The traveling direction detecting means 23 detects the traveling direction of the unmanned dump truck 1 and outputs the traveling direction data to the obstacle position calculation unit 14 and is constituted by, for example, a jar. If the position detecting means 21 detects relative coordinate positions as described above, it is better to share the traveling direction data.

As described above, the area map creating unit 13 maps the traveling course 6 to the area map for obstacle detection. The area map for obstacle detection in this embodiment includes a wide area map 31 shown in FIG. 2 and a detailed map 32 shown in FIG. The wide area map 31 is provided with a traveling coordinate system X—y in an area covering the entire area of the traveling course 6, and the traveling coordinate system Xy space is defined by a straight line parallel to both coordinate axes X and y at predetermined distances. It is divided and configured. In this case, the area surrounded by the adjacent straight lines in the X and Y directions is called "area". The distance between the straight lines delimiting the wide area map 31 is large, so that the traveling course 6 can be roughly represented. Then, the traveling course 6 stored in the course data storage unit 1 かに is roughly mapped on the wide area map 31. That is, the area map creating unit 3 assigns a unique number to each area of the wide area map 31 that includes the coordinates of the points at predetermined distances on the traveling course 6 and Is stored in the memory in the obstacle map storage unit 12 corresponding to the wide area map 31. This unique number represents the detail map 32 number corresponding to this area.

The detailed map 32 shows each area of the wide area map 31 numbered in detail. In Figure 3, these numbers represent the areas of Nos. 5, 6, and 7 of the wide area map 31 in Figure 2 in detail. Each detailed map 3 2 is composed of the same area as the wide area map 31, but each area of this detailed map 32 is more detailed than the wide area map 31. Has been cut. In the example of FIG. 3, one side of the area of the wide area map 31 is divided into 16 equal parts. In addition, although the details will be described later, the detailed map 32 corresponding to the curve section of the traveling course 6 includes a curve map 3 3 in order to more accurately map the traveling course 6. Is created. The distance of one side of each section of the curve map 33 is further divided than the above section of the detailed map 32.

Each area of the detailed map 32 and the curve map 33 is configured to correspond to a predetermined bit of a predetermined address of the memory in the obstacle map storage unit 12. Then, the area map creating unit 13 also determines points for each predetermined distance on the traveling course 6 for each area of the detailed map 32 and the curve map 33 as in the case of the wide area map 31. Map. Then, the area map creating unit 13 writes “Γ” in a memory bit in the obstacle map storage unit 12 corresponding to each of the mapped areas. As will be described later, the coordinate data of the points mapped to the detailed map 32 and the curve map 33 are calculated and calculated based on the coordinate data stored in the course data storage unit 11. Can be

Next, in accordance with the flowchart shown in FIG. 4, a processing method in which the area map creating unit 13 maps the detailed map 32 and the curve map 33 will be described in detail. The step number of the flow is denoted by S, and S 1 represents Step 1.

S 1: n is initialized to “n = 0” for repetitive processing.

S2: From the coordinate data of each point on the traveling course 6 stored in the course data storage unit 11, the coordinate data (xO, yO) of the automatic traveling start position (start position) P0 is obtained. Ask. In this step, the coordinate data (X n, y 11) of the point P n on the traveling course 6 for each predetermined distance is calculated with this start position as a starting point, that is, n = 0 as an initial value. Find in. Here, the predetermined distance is set, for example, as a travel distance per second when the unmanned dump truck 1 travels on the travel course 6 at a specified vehicle speed. The points of P n (X n, yn) calculated in this way are mapped on the detailed map 32 as shown in FIG.

S3: Based on the coordinates (Xn, yn) of the obtained point Pa, a detailed mapping of the point Pn is performed. Map to step 32. Also, “1” is written to a predetermined bit of the obstacle map storage unit 12 corresponding to the area where the point P n is mapped.

S 4: The bits of the obstacle map storage unit 12 corresponding to the three areas close to the point P n (xn, yn) and adjacent to the area where the point P n is mapped are also “ Here, the above three adjacent areas are obtained as follows: The area where the point P n is mapped is represented by X and y as shown in FIG.铀 is divided into two equal parts to form four sub-areas.Of these four sub-areas, a sub-area including the point Pn is obtained, and a detailed map adjacent to the sub-area including the point Pn is obtained. Three areas are required for 3 2. The processing in this step is performed so that the traveling course 6 has a permissible range of obstacle 2 detection.

S 5: Next, the coordinates of the point Pnc based on the formulas “Xnc = Xn−1 + (VcosΘ) Ζ2” and “ync = yη−1 (Vsin ^) Z2” Calculate and calculate data (xny nc). The point Pnc is a point interpolating between the points Pn-1 and Pn as shown in Fig. 6, and the traveling course 6 is continuously mapped to the detailed map 32 so that the traveling course 6 is mapped. It is calculated. Here, Θ represents the angle formed by the straight line connecting points P n-1 and P n as x 轴, as shown in Fig. 5, and corresponds to the traveling direction from point P n-1 to point P n I do. When n = 0, the processing of this step is not performed.

S6: Map the point Pnc to the detailed map 32 based on the coordinate data (xnc.ync) of the point Pnc, and store the obstacle map 1 corresponding to the area where the point Pnc is mapped. Write “Γ” to the predetermined bit of 2.

S7: In the bit of the obstacle map storage unit 12 corresponding to the three areas close to the point Pnc (xnc, ync) and adjacent to the above-mentioned area where the point Pnc is mapped. Also, write "1" to make traveling course 6 have an allowable value range. At this time, "Γ" may be already written in the bit to be written. Here, the above three adjacent areas are obtained in the same manner as in the previous step.

S 8: From three consecutive points Pn-2 (xn-2, yn-2), Pn-1 (xn-1, yn-1) and Pn (xn, yn), near point Pn Judge whether there is a curve or not. Go to S 9 when making a curve Proceed, otherwise proceed to S13. Here, when the angle between the straight line connecting the points Pn-2 and Pn-1 and the straight line connecting the points Pn-1 and Pn is larger than a predetermined value, it is determined to be a curve.

S 9: Create a curve map 33 by further subdividing the detailed map 32 corresponding to the curve section (between P11-1 and Pn). A unique number representing the curve map 33 is stored in association with each area. In FIG. 7, the distance of one side of the area of the curve map 33 is shown as an example in which one side of each area of the detailed map 32 is divided into two equal parts. However, the present invention is not limited to this. For example, by further dividing the section more finely, the mapping accuracy of the curve section is improved.

S 10: For the points P ti and P nc, the points P ti and P nc are calculated using the expressions “X a0 = xa + (V cos θ). Calculate and calculate the aO coordinate data (xaO, yao). Where a = n, nc. The point PnO is a point that interpolates between the points Pnc and Pn as shown in FIG. The point P ncO is a point for interpolating between the point P n-1 and the point P nc. In this way, the traveling course 6 is continuously mapped to the curve map 33. Here, Θ represents the angle formed by the straight line connecting the point P n-1 and the point P nc or the straight line connecting the point P nc and the point Ρ ιι to the x-axis, as described above.

S 1 1: Each point P n, P nc based on the coordinates P n (X n, yn), P nc (X nc, y nc) and P aO (x a0, y aOJ Then, PaO is mapped to the curve map 33, and "j" is written in a predetermined bit of the obstacle map storage unit 12 corresponding to the area where each point is mapped.

5 1 2: Each of the three areas near the points P 11, P nc, and PaO and adjacent to the above-mentioned area of the force map 33 where each point is mapped is determined. "Γ" is also written in the bit of the obstacle map storage unit 12 corresponding to each of the obtained adjacent areas, and the allowable width is given to the running course 6. At this time, the bit to be written is already " 1 "may be written. Here, three adjacent areas are obtained in the same manner as described above.

5 1 3: Check whether n satisfies “n ≥ N” vigorously, and if satisfied, end. Then, the process ends, and if not, the process proceeds to S14.

3 1 4: Advance 1 by 1, return to 8 2 and repeat the process.

As described above, each point on the traveling course 6 is mapped to the wide area map 31, the detailed map 32, and the curve map 33. In addition, regarding each area of the detailed map 32 and the curve map 33 including each point on the traveling course 6, the bit of the obstacle map storage unit 12 corresponding to each area includes “ Γ is written.

Next, obstacle detection will be described. At the time of automatic traveling control, it is determined whether or not the obstacle 1 is on the traveling course 6 in the following procedure. The obstacle distance detection means 22 detects the distance to the obstacle 2 and outputs the obstacle distance data to the obstacle position calculation unit 14. At the same time, the traveling direction detecting means 23 outputs the traveling direction data to the obstacle position calculating section 14. The obstacle position calculation unit 14 inputs the current travel position from the position detection means 21 as coordinate data on the travel coordinate system X—y, and uses this coordinate data, the obstacle distance data described above, and travel. The coordinate data of the position of the obstacle 2 is calculated based on the direction data.

Then, the obstacle determination unit 15 maps the position of the obstacle 2 calculated by the obstacle position calculation unit 14 to the detailed map 32 and the curve map 33. The obstacle determining unit 15 determines whether or not "1" is written in the bit of the obstacle map storage unit i2 corresponding to the mapped area. The obstacle determination unit 15 determines that “If 2 is written, it determines that the obstacle 2 is on the traveling course 6 and stops traveling or detouring to prevent collision with the obstacle 2. In this way, the unmanned dump 1 is avoided. Thus, when judging the presence or absence of the obstacle 2, "Γ" is added to the bit of the obstacle map storage unit 12 corresponding to the area of the obstacle 2. Since it is only necessary to judge whether or not it is written, the calculation load of obstacle judgment is greatly reduced.

The straight section of the traveling course 6 is mapped to the detailed map 32, and the curved section is mapped to the force map 33 composed of a smaller area than the straight section. In Fig. 7, the two-dot chain line A continuously connects the corners of each section so that the running course 6 has the maximum width for each section mapped to the curve map 33. It is a curve. Dotted line B is a curve that continuously connects the corners of each section of the traveling course 6 mapped on the detailed map 32 so that the section has the maximum width. The width of the traveling course 6 sandwiched between the lines A and A is smaller than the width sandwiched between the lines B and B. As described above, since the curve section is mapped to the force map 33 in a small section, the mapping accuracy in the curve section can be improved. As a result, the conventional misjudgment, that is, two obstacles not located on the traveling course 6 <, because the obstacle is near the traveling course 6 in the force section, There is no such thing as being erroneously determined to be present.

In addition, if the area of the area map is made smaller as described above, the accuracy of obstacle detection can be improved. However, if the entire road surface on which the unmanned dump truck 1 travels is divided into small sections from the beginning to improve accuracy, the memory capacity of this section map will be extremely large. In addition, the area through which traveling course 6 passes is a part of the entire area map, so if it is fine from the beginning, a lot of wasteful memory will be added. Therefore, in the present invention, in order to use the storage memory of the area map efficiently, the area map for obstacle detection has three types of maps, that is, the wide area map 31 and the detailed map 32. And a curve map 33. The wide area map 31 shows the outline of the position where the traveling course 6 passes on the entire road surface of the traveling coordinate system Xy, and the detailed map 32 shows each area of the wide area map 31 where the traveling course 6 passes. Is shown in more detail. In the curve map 33, only the curve section of the traveling course 6 is represented by a smaller area. Since only the necessary area is represented in a more detailed area in this way, the memory in the obstacle map storage unit 12 can be configured with a small capacity and can be used efficiently. Industrial applicability

The present invention makes it possible to detect an obstacle without increasing the calculation load on a controller, and furthermore, a method and a method for detecting an obstacle in an unmanned vehicle that can detect an obstacle without reducing accuracy. Useful as an object detection device.

Claims

The scope of the claims

1. In the mapping method of obstacle detection of an unmanned vehicle which detects an obstacle (2) on a traveling course (6) in which an unmanned vehicle (1) travels automatically,

The area including the traveling course (6) is sectioned at predetermined intervals by a straight line parallel to the coordinate axes x and y to form a number of areas, and an area through which the traveling course (6) passes is stored.

If, during automatic driving, the area where the detected obstacle (2) is located and at least one of the memorized passage areas meet, the obstacle (2) is moved along the traveling course ( 6) A mapping method for obstacle detection of unmanned vehicles, which is determined to be above.

2. The unmanned person according to the range 1 of the blue sky seeking, wherein the area of the curved section of the traveling course (6) is a section that is finely divided from the area of the straight section of the traveling course (6). A method for detecting obstacles in vehicles.

3. In an unmanned vehicle obstacle detecting device that detects an obstacle (2) on a traveling course (6) in which an unmanned vehicle (1) automatically travels,

The area including the running course (6) is divided by a predetermined distance with a straight line parallel to the coordinate axes x and y to form a large number of areas, and is constituted by memory areas corresponding to the areas. An obstacle detection area map storage unit (12) for storing a map,

Coordinate data of points at predetermined distances on the traveling course (6) is calculated, and predetermined data including the calculated coordinate data is stored in a memory area corresponding to an area including the position of the coordinate data. Area map creation unit (13) to write in the

An obstacle position calculator (14) for calculating the position of the obstacle (2);

The calculated obstacle (2) position is mapped to the memory area, and the mapped obstacle (2) position and the coordinate data position written in the memory area are mapped. An obstacle determining unit (15) that determines that the calculated obstacle (2) is on the traveling course (6) when at least one of the areas including An obstacle detection device for an unmanned vehicle, comprising:

4. The stored map is

A wide area map (31) for forming a large number of areas by dividing an area including the traveling course (6) by a predetermined distance at a predetermined distance by a straight line parallel to the coordinate axes x and y;

A detailed map (32) which forms an area through which the traveling course (6) passes by an area which is divided more finely than an area of the wide area map (3);

A curve map (33) formed by defining a section of a curve section of the traveling course (6) in a section divided more finely than a section of the detailed map (32). An obstacle detection device for an unmanned vehicle according to range 3.