JP2009252162A - Apparatus and method for generating map data - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a map data generating apparatus that correctly recognizes even an optical special object and generates an appropriate surrounding environment map. <P>SOLUTION: Included are: a temporary map data generation portion that generates temporary map data which contain a self-position at a current position and the entire area of which is divided into grids; a grid information accumulation portion that updates, registers and accumulates a past laser-scanning result for each grid of the temporary map data; a determination process portion that determines whether each grid is an obstacle or movable based on information of each grid accumulated in the grid information accumulation portion; and a true map data generation portion that generates true map data which reflect an actual environment based on a determination result by the determination process portion. The determination process portion determines whether each grid is an obstacle or not based on a percentage of reflected laser in a predetermined-direction area which reflects laser at a high percentage. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a map data generation device, a map data generation method, and a mobile robot.

  In order for an autonomous mobile robot to move autonomously, a map having accurate positional information of obstacles is required. As a method for creating a surrounding map by a mobile robot, a method using a laser distance meter (laser range finder) is known (for example, Patent Document 1).

JP 2005-326944 A

  The laser rangefinder recognizes an object in the surrounding environment by detecting a light receiving unit where the laser hits the object and returns. Therefore, in the case of an object with special optical properties (for example, window glass, mirror, lens) such as glass that has been mirror-polished, it may not be recognized correctly because it does not return well even if the laser hits it. It happens a lot. Then, a situation occurs in which the actual environment and the surrounding environment recognized by the mobile robot are different. As a result, there is a risk of generating a movement path that passes through the glass that cannot actually pass through, which is very dangerous.

  An object of the present invention relates to a map data generation device, a map data generation method, and a mobile robot that accurately recognize an optical special object and generate an appropriate surrounding environment map.

The map data generation apparatus of the present invention is a map data that can autonomously move in an environment including an optically special obstacle in addition to an obstacle that diffusely reflects laser light, and that is referred to by a mobile robot having a laser rangefinder. A temporary map data generating unit for generating temporary map data including the self-position of the mobile robot at the current position and divided into grids as a whole, and for each grid of the temporary map data A grid information storage unit that updates and stores past laser scanning results and determines whether each grid is an obstacle based on information for each grid stored in the grid information storage unit A processing unit, and a true map data generation unit that generates true map data reflecting an actual environment based on a determination result by the determination processing unit. Said grid information storage unit includes a laser incident direction alpha _i is a direction in which the laser is flying to a grid, the reflected laser incident direction beta _j is the direction in which the reflected laser flew reflected by the grid of these flying the laser And for each grid, the determination processing unit determines whether each grid is an obstacle based on a ratio of the reflected laser in a predetermined direction range in which the ratio of reflecting the laser is high. Features.

  In the present invention, the determination processing unit jumps in a predetermined direction range including the average reflection direction and an average reflection direction calculation unit that calculates an average reflection direction βave that is an average value of the reflected laser flying directions βj for each grid. A counting unit that counts the number of lasers M ′ and the number N ′ of reflected lasers, and a grid in which the ratio of reflected lasers in the predetermined direction range is higher than a predetermined threshold from the count value of the counting unit as an obstacle It is preferable to include a threshold determination unit for determining.

In the present invention, the average reflection direction calculating unit includes a standard deviation calculating unit that calculates a standard deviation σ _β of the reflected laser flying direction β _j using an average value βave of the reflected laser flying direction, and the average reflected direction is predetermined range of directions, including, it is preferable that the a laser traveling direction of the mean Betaave ± standard deviation sigma _beta.

In the present invention, the grid information storage unit further distance d _j measured by reflected laser has accumulated in the incident direction βj pair of reflected laser, the determination processing section is further flying times of the laser M An obstacle grid determination unit that determines a grid with a normal obstacle based on the ratio of the number of reflections and the number of reflections of the laser, and the grid other than that determined as a normal obstacle by the obstacle grid determination unit It is determined whether or not each grid is an optical special obstacle based on the ratio of the reflected laser in a predetermined direction range in which the ratio of reflecting the laser is high, and the grid information storage unit stores the optical registered in the true map data. A grid that corrects the information stored in the grid information storage unit by deleting laser scanning data that is determined to be inappropriate from the relationship with special obstacles It is preferred that broadcast correction unit is provided.

In the present invention, the grid information correction unit sequentially reads the reflection laser flying direction β _j and the measurement distance d _j by the reflection laser from the grid information storage unit, and reverses the flying direction β _j from the grid. generating a virtual laser having a length of measurement distance d _j, it is preferable to delete the information of the laser when the virtual laser is passing through in a direction out of the proper angle optical special obstacle.

In the present invention, the determination processing unit includes an average reflection direction calculation unit for calculating an average value in the form of the average reflection direction βave of the reflected laser incident direction beta _j for each grid, the grid information adjustment unit, a virtual laser It is preferable to delete the laser information when the direction of passing through the grid of the optical special obstacle deviates from the average reflection direction βave by a predetermined threshold or more.

The map data generation method of the present invention is capable of autonomously moving in an environment including an optically special obstacle in addition to an obstacle that irregularly reflects light, and is a map that is referenced by a mobile robot having a laser rangefinder. A map data generation method for generating data, including a temporary map data generation step for generating temporary map data including a self-position of the mobile robot at a current position and divided into grids as a whole, and a grid of the temporary map data Each grid can be moved as an obstacle based on the grid information accumulation process for updating and accumulating the past laser scanning results for each grid and the information for each grid accumulated in the grid information accumulation process. And a true map data for generating true map data reflecting the actual environment based on the judgment result of the judgment processing step. Comprising a data generating step, the grid information storage step, a laser incident direction alpha _i is a direction in which the laser is flying in a certain grid, directions reflected laser reflected by the grid of these flying the laser flew Is registered for each grid, and the reflected laser flying direction β _j is
The determination processing step is characterized by determining whether each grid is an obstacle or is movable based on a ratio of the reflected laser in a predetermined direction range in which a ratio of reflecting the laser is high.

  The mobile robot of the present invention is a mobile robot that can autonomously move in an environment including an optically special obstacle in addition to an obstacle that irregularly reflects light, and has a laser rangefinder, and generates the map data The device is built-in.

  According to the present invention, it is possible to accurately recognize an obstacle in the surrounding environment and obtain map data that appropriately reflects the surrounding environment. It was not possible to recognize optical special obstacles such as glass simply by registering the result of a single laser scan on a map as in the past. In this regard, in the present invention, after obtaining an angle region in which reflection is likely to occur, a past laser scanning result in this range is used as a basis for determination. Therefore, for example, even when there is an optically special obstacle that can be detected only within a limited angle range such as a glass plate, it is possible to accurately recognize this and generate accurate map data.

Embodiments of the present invention will be described with reference to the drawings.
(First embodiment)
A first embodiment of the present invention will be described.
FIG. 1 is a diagram showing an external configuration of the mobile robot 200. The mobile robot 200 includes a main body 210, a moving unit 220, and a measuring unit.
The structure of the moving means 220 is not particularly limited as long as the main body 210 can be moved, such as a wheel type, a crawler type, and a bipedal walking type, but FIG. 1 shows a case of a wheel type.

As a measuring means, a laser distance meter 230 and an odometry sensor 240 are provided.
The laser distance meter 230 includes a laser emitting unit and a light receiving unit, and is attached to the main body unit 210.
As shown in FIG. 2, the laser distance meter 230 can scan an area of 180 degrees in front, and can detect at least one degree as resolution of the angle.
The main body 210 has a built-in odometry sensor 240.
The specific configuration of the odometry sensor 240 is not limited, but may be an acceleration sensor or a speed sensor that detects the acceleration or speed of the mobile robot 200, or, if the moving means is a wheel type, the amount of rotation of the wheel. It may be an encoder that detects.

  The main unit 210 includes a control unit (not shown). The control unit generates a map data generation unit 300 that generates a map image, and moves the mobile robot 200 with reference to the generated map image. A movement control unit (not shown) for controlling.

The map data generation unit 300 will be described.
FIG. 3 is a diagram illustrating a configuration of the map data generation unit 300.
The map data generation unit 300 is based on the temporary map processing unit 310 that performs a process of creating a temporary map based on the measurement value obtained by the laser distance meter 230 and the measurement value obtained by the odometry sensor 240, and the accumulated data of the measurement value obtained by the laser distance meter 230. And a true map processing unit 320 that complements the temporary map created by the temporary map processing unit 310 and performs processing for creating a true map.

  The temporary map processing unit 310 determines the self-position of the mobile robot 200 on the basis of the laser scanning data storage unit 311 that temporarily stores the scanning data obtained by scanning the surroundings with the laser rangefinder 230 and the measurement value obtained by the odometry sensor 240. The absolute position of the mobile robot 200 whose self-position is corrected using the self-position prediction unit 312 to be predicted, the scan data stored in the laser scan data storage unit 311 and the self-position prediction by the self-position prediction unit 312 is included. A temporary map data generation unit 313 that generates a temporary map of the surrounding environment that is divided into grids as a whole.

The laser scanning data storage unit 311 temporarily stores the surrounding scanning data acquired by the laser distance meter 230.
The scan data includes, for example, the laser irradiation direction and the received light intensity in each irradiation direction as shown in FIG. 4, and further includes the time until the laser returns, etc. Shows the existence of the obstacle and the distance to the obstacle.

  The self position prediction unit 312 predicts the self position of the mobile robot 200 based on the sensor value obtained by the odometry sensor 240. That is, a point in the environment at the present time is predicted by taking into consideration the route and distance that the self travels to the self position before the movement.

The temporary map data generation unit 313 creates a temporary environment map based on the result of the laser scanning while integrating the self-position prediction by the self-position prediction unit 312 and the result of the laser scanning to correct the deviation of the self-position prediction. .
At this time, in correcting the self-position prediction, first, feature points are extracted based on the results of laser scanning before and after the movement, and matching is performed so that these feature points overlap.
Then, a deviation between the self-position prediction and the matching in the self-position prediction unit 312 is obtained, the self-position prediction obtained earlier is corrected, and self-position estimation is performed.
In this way, the self-position prediction by the odometry sensor 240 only increments the amount of movement, so there is a high risk that errors will accumulate and the actual position and the predicted position will deviate greatly. This is because the deviation cannot be repaired.
In this regard, robust self-position estimation is performed by making sequential corrections using the results of laser scanning.

  Then, a temporary map of the surrounding environment is created by adding the result of laser scanning to the estimated self-position.

  Here, an example of a temporary map in which a feature point (for example, a grid returned by reflecting a laser) is simply registered in order to match the results of laser scanning before and after movement is shown. Explain the differences from the actual environment that may be inherent.

In the working environment of the mobile robot 200, as shown in FIG. 1, there are a normal obstacle 910 and an optical special obstacle 920 that has a special action on laser light unlike a normal obstacle.
Here, a normal obstacle is an object which is opaque and has a rough surface.
Since such a normal obstacle reflects light irregularly, as shown in FIG. 5, a constant reflection intensity can be obtained regardless of the incident angle of light with respect to the surface of the obstacle.
Therefore, the presence of the obstacle can be recognized without error by receiving the reflected light from the obstacle at the light receiving unit of the laser distance meter 230.

On the other hand, the optical special obstacle is represented by a glass whose surface is mirror-polished and has optical transparency.
In such an optical special obstacle, when the incident angle is small, the laser is specularly reflected.
In addition, when the incident angle is large, not only the reflection from the surface but also the reflected light from the object on the other side transmitted through the optical special obstacle may be returned.
In the optical special obstacle, as shown in FIG. 6, the reflection intensity varies depending on the incident angle with respect to the surface of the optical special obstacle, and the mobile robot 200 faces the surface of the optical special obstacle (incident angle near 90 °). A predetermined reflection intensity can be obtained only when

When there is such an optical special obstacle, if the obstacle is registered in the map image data based on the result of the laser scanning by the laser rangefinder 230, map data different from the actual environment will be created.
For example, as shown in FIG. 7, when there is a glass plate 920 and an obstacle 910, the laser L ₁ having a shallow incident angle with respect to the glass 920 is specularly reflected, and thus the reflected light from the glass direction is actually reflected. May reflect the obstacle 910 at the end of the specular reflection.
Further, since most lasers pass through the glass 920 (for example, L ₂ ), the laser returning from the glass 920 cannot be recognized by the light receiving unit, and as a result, the region where the laser has passed is not in the glass itself. The existence will not be recognized.
Therefore, if the result of laser scanning is directly registered in the temporary map data, as shown in FIG. 8, the glass 920 that actually exists is not registered, and the virtual image 930 of the obstacle that does not actually exist is registered. Will be.

  The temporary map data generated by the temporary map data generation unit 313 is sent to the true map processing unit 320, and is corrected to true map data that accurately reflects the actual environment by a complementing process based on the accumulated data.

  The true map processing unit 320 accumulates past laser scanning results and generates a grid information accumulation map in which past laser scanning results are registered in each grid of the temporary map generated by the temporary map data generation unit 313. A grid information storage unit 330; a determination processing unit 340 that determines whether each grid is an obstacle based on information registered in the grid information storage map generated by the grid information storage unit 330; A true map data generation unit 390 that generates true map data that accurately reflects the actual environment based on the determination result by the unit 340.

  The grid information accumulation unit 330 accumulates past laser scanning results, and updates and registers past laser scanning results for each grid on the temporary map data newly generated by the temporary map data generation unit 313. I will do it. Here, the grid information storage unit 330 updates and registers the past laser scanning results for each grid as follows.

That is, in each grid, the number M of times the laser has come to the grid and the laser flying direction α _i (i = 1 to M) at the time of each flying are registered.
Further, when the grid reflects the laser, the number N of times the grid reflected the laser and the laser flying direction β _j (j = 1 to N) at each reflection are registered.
When the registration data for each grid is made into a table, for example, it is as follows.

In defining the direction here, every time the temporary map data is updated and created in accordance with the movement of the mobile robot 200, the 0 ° reference line is also updated according to the direction of the mobile robot 200, and the flying direction accordingly. Alternatively, the reference line may be fixed in the absolute coordinate system (world coordinate system) and the newly acquired laser scanning data may be corrected.
In any case, as long as it is consistent as a whole, the angle reference line is arbitrarily determined.
In some cases, the direction is defined in a range of 0 ° to 180 °, and 0 ° ≦ αi and βj <180 °.
For example, when a plate glass is considered, since reflection occurs on both the front surface and the back surface, the laser reflected from the front surface and the laser reflected from the back surface are distinguished, which is not preferable.
Therefore, in such a case, it is preferable to redefine the angle exceeding 180 ° by subtracting 180 ° and satisfying 0 ° ≦ αi and βj <180 °.

  In this way, past laser scanning results are updated and registered for every grid of the temporary map data, and a grid information accumulation map having information for each grid is generated.

  The determination processing unit 340 examines data for each grid of the grid information accumulation map generated by the grid information accumulation unit 330 and determines whether the grid is an obstacle or is movable.

FIG. 9 is a diagram illustrating a configuration of the determination processing unit 340.
The determination processing unit 340 includes an average reflection direction calculation unit 350 that calculates an average value βave of the reflected laser flying direction β _i for each grid, and a flying laser and a reflected laser that are near the average value of the reflected laser flying direction β _i . Count unit 360 that counts the number, and determines whether the grid is an obstacle grid by comparing the ratio of the flying laser and the reflected laser calculated based on the count result of count unit 360 with a predetermined threshold And a grid determination result storage unit 380 that stores the determination result in the threshold determination unit 370.

The average reflection direction calculation unit 350 extracts the laser flying direction β _j at the time of reflection from the past laser scanning results registered for each grid, and calculates the average value βave of the laser flying direction β _j .
The average reflection direction calculation unit 350 is provided with a standard deviation calculation unit 351. The standard deviation calculation unit 351 calculates the standard deviation σ _β of the laser flying direction β _j using the average value βave of the laser flying direction. calculate.

Counting unit 360 includes a laser count unit 361 for counting the number M 'of flying laser in the vicinity of the average value of the reflected laser incident direction beta _i, the number of reflected laser in the vicinity of the average value of the reflected laser incident direction beta _i A reflection laser counting unit 362 that counts N ′. That is, the count unit 360 counts the flying laser and reflecting the laser is in the range of _βave ± σ β in respectively the laser counting unit 361 and the reflected laser counting unit 362.

The threshold determination unit 370 has a predetermined threshold for determining whether or not the obstacle grid is an obstacle, and compares the ratio of the reflected laser in the vicinity of the average reflection direction with the threshold.
That is, determine the percentage of reflected laser in the range from the of βave ± σ β _'number N of counted reflected laser as in reflected laser counting unit 362' number M of flying laser counted by a laser count unit 361.
The ratio of the reflected laser is calculated as “N ′ / (M ′ + N ′)”.
Then, the ratio of the reflected laser is compared with a predetermined threshold value.

Here, when the obstacle is an optical special obstacle, the direction in which the laser is reflected and returned by the optical special obstacle is limited to a narrow range.
Specifically, if the optical special obstacle is a mirror surface or has a high light transmittance, if the laser is not applied from the direction facing the surface of the optical special obstacle, this optical reflection will be reflected or transmitted. Reflected light from special obstacles cannot be measured.
Since reflected light cannot be measured in most angular ranges, optical special obstacles are not registered in the map simply by plotting the results of each laser scan on map data.
In this regard, the determination target is narrowed down to the vicinity where the reflected light is measured, and a determination is made by taking into account the past laser scanning results, and if the reflection ratio is high in that range, it can be determined that the object is an obstacle. Even with a narrowed determination, it can be determined that a grid where the rate of laser transmission is still high is free space.

That is, if the ratio “N ′ / (M ′ + N ′)” of the reflected laser is equal to or greater than a predetermined threshold as a result of the determination by the threshold determination unit 370, the grid is determined to be a grid with an obstacle.
If the ratio “N ′ / (M ′ + N ′)” of the reflected laser is smaller than the predetermined threshold, it is determined that the grid is not an obstacle.
Such a determination result is sent to the grid determination result storage unit 380.

  The grid determination result storage unit 380 stores the determination results in the threshold determination unit 370 and sends the data to the true map data generation unit 390 when the determination results for all the grids are complete.

  The true map data generation unit 390 updates and registers the determination results for all the grids in the temporary map data grid, and creates true map data that accurately reflects the actual environment.

  After such map creation is performed by the map data generation unit 300, the movement route of the mobile robot 200 is autonomously selected with reference to the map data, and movement control is executed.

Next, the operation of the first embodiment having the above configuration will be described with reference to a flowchart.
FIG. 10 is a flowchart showing the operation procedure of the first embodiment.
First, sensing is performed when the mobile robot 200 moves.
That is, acquisition of odometry data by the odometry sensor 240 (odometry data acquisition step ST100) and acquisition of laser scanning data by the laser distance meter 230 (laser scanning step ST110) are executed.
The acquired laser scanning data is temporarily buffered in the laser scanning data storage unit 311 and sent to the grid information storage unit 330.

  The odometry data from the odometry sensor 240 is sent to the self-position prediction unit 312 and processed kinematically, and the self-position is predicted from the moving direction and the moving distance of the mobile robot 200 (self-position prediction step ST120).

Next, in ST130, a temporary map data generation step by the temporary map data generation unit 313 is executed.
In this step, the results of laser scanning before and after movement are matched with reference to self-position prediction.
Then, based on the odometry data and the laser scanning data, temporary map data including the absolute position of the mobile robot 200 whose self-position prediction is corrected and the whole being divided into grids is generated.
The temporary map data generated in this way is sent to the grid information storage unit 330, and the past laser scanning results are updated and registered for each grid (grid information storage step ST140).
That is, the number M of lasers that have come so far and the flying direction αi at the time of each laser flying are updated and registered for each grid, and for the reflected lasers, the number N of reflected lasers and the flying of the reflected lasers. The direction βj is updated and registered (see Table 1).

  Thus, when the update registration of data for all grids is completed, the update data for each grid is sequentially sent to the determination processing unit 340, and a determination processing step (ST150) for determining whether an obstacle or movement is possible for each grid. ) Is performed.

The determination processing step (ST150) will be described with reference to FIG.
FIG. 11 is a flowchart showing the procedure of determination processing step ST150.
In the determination processing step ST150, first, the grid information updated and registered in the grid information storage unit 330 is sequentially read (grid information reading step ST200).
Grid information read is transmitted to the average reflection direction calculating section 350, the average value βave direction beta _j of the reflected laser is flying is calculated (average reflection direction calculation step ST210).
Subsequently, in ST220, the standard deviation sigma _beta reflective laser direction βj is calculated using the average reflection direction βave calculated (standard deviation calculation step ST220)
Thus calculated average reflection direction βave and standard deviation sigma _beta is sent to the counting unit 360, count step ST230, ST240 is executed.

In the counting process, first, to count the number M 'of the laser the laser incident direction alpha _i an angular range of _βave ± σ β is around an average reflection direction Betaave calculated previously enters (laser counting step ST230) .
In addition, the number N ′ of reflection lasers in which the reflection laser flying direction β _j falls within the angle range of βave ± σ _β that is in the vicinity of the average reflection direction βave is counted (reflection laser counting step ST240).
These count results are sent to the threshold determination unit 370, and a threshold determination step (ST250) is executed.

In the threshold determination step ST250, determine the percentage of reflected laser in the range of βave ± σ _β, contrasting this ratio with a predetermined threshold value.
When the ratio “N ′ / (M ′ + N ′)” of the reflected laser is equal to or greater than the threshold, this grid is determined as a grid with an obstacle.
Further, when the ratio “N ′ / (M ′ + N ′)” of the reflected laser is equal to or less than the threshold value, it is determined that this grid is not an obstacle.
Such a determination result is sent to and stored in the grid determination result storage unit 380 (grid determination result storage step ST260).

ST200 to ST260 are repeated for each grid until threshold determination is performed for all grids, and determination data is sent to the true map data generation unit 390 when threshold determination is performed for all grids.
In the true map data generation unit 390, the determination result obtained for each grid is registered in each grid of the temporary map data, and true map data is generated (true map data generation step ST160).

As described above, according to the first embodiment having such a configuration, it is possible to accurately recognize an obstacle in the surrounding environment and obtain map data that appropriately reflects the surrounding environment.
It was not possible to recognize optical special obstacles such as glass simply by registering the result of a single laser scan on a map as in the past.
In this regard, in the present embodiment, the average reflection direction calculation unit 350 obtains an angle region where reflection is likely to occur, and the past laser scanning results in the vicinity of the average reflection direction are counted and extracted by the counting unit 360. The reflection ratio in the angle region is used as a basis for determination.
Therefore, for example, even when there is an obstacle that can be detected only within a limited angle range such as a glass plate, it is possible to accurately recognize this and generate accurate map data.

(Second Embodiment)
Next, a second embodiment of the present invention will be described.
The basic configuration of the second embodiment is the same as that of the first embodiment. However, in the second embodiment, the determination processing unit distinguishes between normal obstacles and optical special obstacles, and positions of the optical special obstacles. The laser scanning data determined to be inappropriate on the basis of the above is deleted from the accumulated data.

FIG. 12 is a diagram illustrating the grid information storage unit 330 and the determination processing unit 440 in the second embodiment.
Grid information storage section 330, where holding by updating registers the results of the past of the laser scanning for each grid, in the second embodiment, further, a distance d _j measured by reflecting a laser also accumulated as the accumulation data I will do it.
That is, it accumulates in pairs with the reflection laser flying direction βj and the distance dj measured by the reflection laser.

  In the second embodiment, the determination processing unit 440 includes an obstacle grid determination unit 441, an obstacle grid storage unit 442, an optical special obstacle determination unit 443, and a grid information correction unit 490.

The obstacle grid determination unit 441 determines whether each grid is an obstacle grid based on the data stored in the grid information storage unit 330.
That is, for each grid, the ratio of the reflected laser is compared with a predetermined threshold value.

The ratio of the reflected laser here is determined as “N / (M + N)” from the number M of lasers flying and the number N of laser reflections.
The obstacle grid determination unit 441 sends a determination result to the obstacle grid storage unit 442 for the grid determined to be an obstacle grid, and sends the other grid to the optical special obstacle determination unit 443.

  The obstacle grid storage unit 442 stores a grid in which the ratio “N / (M + N)” of the reflected laser is equal to or greater than a threshold value as an obstacle grid.

The optical special obstacle determination unit 443 has basically the same configuration as the determination processing unit in the first embodiment, and includes an average reflection direction calculation unit 450, a count unit 460, a threshold determination unit 470, and a grid determination result storage. Part 480.
Then, whether the count unit 460, the average reflection direction Betaave counts flying laser and reflecting the laser is in the range of _βave ± σ β is around obstacles by comparing the ratio of the reflected laser threshold of the reflected laser not Determine whether.

However, since what is sent to the optical special obstacle determination unit 443 is grid data excluding the obstacle grid, the threshold determination unit 470 calculates the ratio “N ′ / (M ′ + N ′)” of the reflected laser. A grid that is equal to or greater than the threshold is determined to be an optical special obstacle such as glass.
Therefore, in the grid determination result storage unit 480, the grid in which the ratio “N ′ / (M ′ + N ′)” of the reflected laser is greater than or equal to the threshold in the determination by the threshold determination unit 470 is stored as an optical special obstacle. The average reflection direction βave calculated by the average reflection direction calculation unit 450 is stored as the normal direction of the optical special obstacle.

  Based on the determination results stored in the obstacle grid storage unit 442 and the grid determination result storage unit 480, a grid of temporary map data indicating whether each grid is an obstacle, an optical special obstacle, or a free space And true map data is generated.

Next, the grid information correction unit 490 will be described.
The grid information correction unit 490 corrects the grid information by deleting the laser scanning data determined to be inappropriate from the relationship with the optical special obstacle registered in the true map data.
The grid information correction unit 490 sequentially reads the reflection laser flying direction β _j and the measurement distance d _j by the reflection laser from the grid information storage unit 330.
Then, a virtual laser having a length of the measurement distance d _j is generated from the grid in the direction opposite to the flying direction β _j and plotted on the true map data.

  For example, as shown in FIG. 13, when the grid Gxy reflects several lasers, the virtual laser Ls having a distance dj is plotted on the map in the direction opposite to the reflected laser flying direction βj from the grid Gxy.

When the optical special obstacle is registered in the true map data, it is investigated whether the virtual laser passes through the grid of the optical special obstacle.
Here, when there is a grid of optical special obstacles on the virtual laser, the passing direction of the virtual laser is compared with the normal direction of the optical special obstacle.
Virtual (if for example sigma _beta or higher) shift angle between the normal direction of the passage direction and optical special obstacle lasers case is above a predetermined threshold value, the laser range finder 230 in the direction that the laser is an off-proper angle That is, it returns to the light receiving part.

That is, since the laser is mirror-reflected by the optical special obstacle and returned, the information obtained by the laser is a virtual image 930 and does not reflect the actual environment.
Therefore, such inappropriate grid information is deleted from the grid information storage unit 330, and only the data reflecting the actual environment is left, and the grid information is corrected appropriately.

On the other hand, if the angle of deviation between the virtual laser passing direction and the normal direction of the optical special obstacle is smaller than a predetermined threshold (for example, smaller than σ _β ), this laser will cover the optical special obstacle (for example, a glass plate). It was transmitted from the front and returned. Since this is thought to simply reflect the normal obstacles that exist beyond the glass plate, there is no need to delete such a laser.

The operation of the second embodiment having such a configuration will be described.
The basic operation procedure is the same as in the first embodiment.
However, in the second embodiment, as shown in the flowchart of FIG. 14, the obstacle grid determination step ST310 is first executed after the grid information reading step ST300 in the determination processing step.
In the obstacle grid determination step ST310, it is determined whether or not each grid is an obstacle grid based on the data stored in the grid information storage unit 330. For each grid, the number M of lasers flying and the reflection of the laser are determined. From the number N, “N / (M + N)”, which is the ratio of the reflected laser, is calculated, and the ratio of the reflected laser is compared with a predetermined threshold value.

  If the ratio “N / (M + N)” of the reflected laser is greater than or equal to a predetermined threshold, the grid is determined to be a grid with a normal obstacle (ST311: YES) and stored in the obstacle grid storage unit. .

  On the other hand, if it is not an obstacle grid (ST311: NO), it is determined whether or not the grid is an optical special obstacle (ST320 to ST370).

ST320 to ST370 are the same procedure as ST210 to ST260 in the first embodiment. However, in the second embodiment, since the obstacle grid is separately determined in advance by the obstacle grid determination step ST310, the threshold determination in ST360 is performed. The grid exceeding the threshold in the process is an optical special obstacle.
In this way, in the determination processing step in the second embodiment, a normal obstacle, an optical special obstacle, and a free space are distinguished from each other and registered in true map data.

In 2nd Embodiment, after true map data are produced | generated, the grid information correction process by a grid information correction part is further performed.
FIG. 15 is a flowchart showing the procedure of the grid information correction process.

In the grid information correction step, first, grid information is read from the grid information storage unit 330 (grid information read step ST400).
At this time, the flying direction βj of the reflected laser accumulated for each grid and the distance dj measured by the reflected laser are sequentially read out.
Subsequently, in the virtual laser generation step ST410, a virtual laser having a length of the measurement distance d _j is generated from the grid in the direction opposite to the flying direction β _j .
The generated virtual laser is plotted on the true map data.

Obstacle grids, optical special obstacle grids, and free spaces are registered on the true map data, and it is investigated whether there are optical special obstacles on the path through which the virtual laser passes (passing grid). Investigation process ST420).
When there is no optical special obstacle on the passage path of the virtual laser (ST430: YES), it is determined that the flying direction β _j and the measurement distance d _j of the reflected laser are appropriate data.
And it returns to ST400 and repeats until investigation about all the grid information is complete | finished.

Here, when there is an optical special obstacle on the passage path of the virtual laser in the passing grid investigation step of ST420 (ST430: NO), the passage direction of the virtual laser is then compared with the normal direction of the optical special obstacle. Then, it is determined whether the passing direction is appropriate (passing direction determining step ST450).
When the deviation angle between the virtual laser passage direction and the normal direction of the optical special obstacle is smaller than a predetermined threshold (for example, σ _β ), the laser passage direction is appropriate (ST460: YES).

On the other hand, if the deviation angle between the passing direction of the virtual laser and the normal direction of the optical special obstacle is equal to or greater than a predetermined threshold, the laser has returned to the laser rangefinder 230 from a direction outside the proper angle. .
Therefore, such laser information is deleted from the data in the grid information storage unit 330 (grid information deletion step ST470).
In this way, appropriate correction of grid information is performed.

As described above, according to the second embodiment, after determining and distinguishing the normal obstacle and the optical special obstacle, the grid information indicates an inappropriate laser scanning result from the angular relationship with the optical special obstacle. Delete from the storage unit 330.
Accordingly, in the subsequent determination, since inappropriate data is not used as a basis, the accuracy of the determination is increased, and map data reflecting the surrounding environment more appropriately can be generated.

In addition, this invention is not limited to the said embodiment, The deformation | transformation in the range which can achieve the objective of this invention, improvement, etc. are included in this invention.
The standard deviation calculation unit is provided in the average reflection direction calculation unit, and the standard deviation σ _β in the reflected laser flying direction is used to determine the ratio of the flying laser and the reflected laser in the range of β ± σ _β. Regardless, the range may be delimited by other predetermined values.

The figure which shows the external appearance structure of a mobile robot. The figure showing a mode that the circumference is sensed with a laser distance meter. The figure which shows the structure of a map data generation part in 1st Embodiment. The figure which shows an example of scanning data. The figure which shows the light reflection intensity in a normal obstruction. The figure which shows the light reflection intensity in an optical special obstruction. The figure which shows an example of the path | route of a laser in case there exists a glass plate and a normal obstruction. The figure explaining the problem at the time of registering the result of laser scanning as it is. The figure which shows the structure of the determination process part in 1st Embodiment. The flowchart which shows the operation | movement procedure of 1st Embodiment. The flowchart which shows the procedure of the determination process process in 1st Embodiment. The figure which shows the structure of the determination process part in 2nd Embodiment. The figure which plotted virtual laser Ls on the map from grid Gxy in 2nd Embodiment. The flowchart which shows the operation | movement procedure of a determination process process in 2nd Embodiment. The flowchart which shows the procedure of a grid information correction process in 2nd Embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 200 ... Mobile robot, 210 ... Main part, 220 ... Moving means, 230 ... Laser distance meter, 240 ... Odometry sensor, 300 ... Map data generation part, 310 ... Temporary map processing part, 311 ... Laser scanning data storage part, 312 ... Self-position prediction unit, 313 ... provisional map data generation unit, 320 ... true map processing unit, 330 ... grid information storage unit, 340 ... determination processing unit, 350 ... average reflection direction calculation unit, 351 ... standard deviation calculation unit, 360 ... Count unit, 361 ... Laser count unit, 362 ... Reflected laser count unit, 370 ... Threshold value determination unit, 380 ... Grid determination result storage unit, 390 ... True map data generation unit, 440 ... Determination processing unit, 441 ... Obstacle grid determination 442 ... Obstacle grid storage unit 443 ... Optical special obstacle determination unit 450 ... Average reflection direction calculation unit 460 Counting unit, 470 ... threshold judgment unit, 480 ... Grid determination result storage unit, 490 ... Grid information correcting unit.

Claims (8)

A map data generation device that can autonomously move in an environment containing optically special obstacles in addition to obstacles that reflect light irregularly, and that generates map data to be referenced by a mobile robot having a laser rangefinder There,
A temporary map data generation unit that generates temporary map data that includes the self-position of the mobile robot at the current position and is entirely divided into grids, and updates and registers past laser scanning results for each grid of the temporary map data. Grid information accumulating unit that accumulates, a determination processing unit that determines whether each grid is an obstacle based on information for each grid accumulated in the grid information accumulating unit, and a determination result by the determination processing unit A true map data generation unit that generates true map data reflecting the actual environment based on
The grid information accumulating unit includes a laser flying direction α _i that is a direction in which a laser has come to a certain grid, and a reflected laser flying direction β _j that is a direction in which a reflected laser reflected by the grid among these lasers has come. Are registered for each grid,
The said determination process part determines whether each grid is an obstruction based on the ratio of the reflected laser in the predetermined direction range where the ratio which reflects a laser is high. The map data generation apparatus characterized by the above-mentioned. The map data generation device according to claim 1,
The determination processing unit calculates an average reflection direction βave that is an average value of the reflected laser flying directions βj for each grid, and
A counting unit that counts the number M ′ of lasers that have come in a predetermined direction range including the average reflection direction and the number N ′ of lasers that have reflected, respectively;
A map data generation device, comprising: a threshold determination unit that determines, as an obstacle, a grid in which a ratio of reflected lasers in the predetermined direction range is higher than a predetermined threshold based on a count value by the count unit. The map data generation device according to claim 2,
The average reflection direction calculation unit includes a standard deviation calculation unit that calculates the standard deviation σ _β of the reflected laser flying direction β _j using the average value βave of the reflected laser flying direction,
It said average predetermined direction range including the reflection direction, the map data generation apparatus, wherein the a laser traveling direction of the mean Betaave ± standard deviation sigma _beta. In the map data generation device according to any one of claims 1 to 3,
The grid information accumulating unit further accumulates the distance d _j measured by the reflection laser in a pair with the flying direction β _j of the reflection laser,
The determination processing unit further includes an obstacle grid determination unit that determines a grid with a normal obstacle based on a ratio between the number M of lasers flying and the number N of reflections of the laser, and the obstacle grid determination unit includes Determine whether each grid is an optical special obstacle based on the ratio of the reflected laser in a predetermined direction range where the ratio of reflecting the laser is high with respect to the grid other than that determined as a normal obstacle,
The grid information storage unit corrects the information stored in the grid information storage unit by deleting laser scanning data determined to be inappropriate from the relationship with the optical special obstacle registered in the true map data. A map data generation device characterized in that a grid information correction unit is provided. The map data generation device according to claim 4,
It said grid information adjustment unit, the grid information measured from the storage unit and the traveling direction beta _j of the reflected laser and the reflected laser by the measurement distance d _j from the grid in the opposite direction to the incident direction beta _j is read sequentially distance d _j Generates a virtual laser with a length of
When the virtual laser passes through an optical special obstacle in a direction away from an appropriate angle, information on the laser is deleted. In the map data generation device according to claim 5,
The determination processing section has a mean reflection direction calculation unit for calculating an average reflection direction βave an average value of the reflected laser incident direction beta _j for each grid,
The grid information correcting unit deletes the laser information when the direction in which the virtual laser passes through the grid of the optical special obstacle is shifted from the average reflection direction βave by a predetermined threshold or more. Data generator. A map data generation method capable of autonomously moving in an environment containing optically special obstacles in addition to obstacles that reflect light irregularly, and generating map data to be referenced by a mobile robot having a laser rangefinder There,
A temporary map data generation step for generating temporary map data including the self-position of the mobile robot at the current position and divided into grids as a whole, and updating and registering past laser scanning results for each grid of the temporary map data Grid information accumulation step for accumulating, a determination processing step for determining whether each grid is an obstacle or movable based on the information for each grid accumulated in the grid information accumulation step, and the determination processing A true map data generation step of generating true map data reflecting the actual environment based on the determination result of the step,
The grid information accumulating step includes a laser flying direction α _i which is a direction in which a laser has come to a certain grid, and a reflected laser flying direction β _j which is a direction in which a reflected laser reflected by the grid among these lasers has come. And for each grid,
The map data generation method characterized in that the determination processing step determines whether each grid is an obstacle or is movable based on a ratio of reflected lasers in a predetermined direction range in which a ratio of reflecting the laser is high. A mobile robot capable of autonomously moving in an environment including an optically special obstacle in addition to an obstacle that irregularly reflects light, and having a laser distance meter,
A mobile robot comprising the map data generation device according to any one of claims 1 to 6.

Images