JP4464893B2 - Mobile robot - Google Patents

Description

  The present invention relates to a mobile robot.

  Conventionally, as a method of moving a mobile robot to a destination in an environment where an obstacle exists, the position of the obstacle is detected by a plurality of obstacle detection sensors provided in the mobile robot, and the current position of the mobile robot is detected. Based on the information and the position information of the obstacle detected by the obstacle detection sensor, the mobile robot calculates a movement route that avoids the obstacle, and moves the mobile robot along the route (for example, patents). Reference 1).

  Here, the outline of the mobile robot of the prior art example 1 described in Patent Document 1 will be described with reference to FIGS. 7A, 7B, 7C, 8A, and 8B. 7A, 7B, and 7C show the configuration of the mobile robot described in Patent Document 1. FIG. 8A and 8B show a method for determining the movement path of the mobile robot shown in FIGS. 7A, 7B, and 7C.

  As shown in FIGS. 7A, 7B, and 7C, the mobile robot 171 includes a movable main body 171a including a wheel 178 necessary for movement of the robot, an auxiliary wheel 179, and a driving device 175 such as a motor, and a main body. Measured by an obstacle detection sensor 177 attached to the periphery of the portion 171a and detecting an obstacle present in any surrounding detection area using ultrasonic waves, infrared rays, or the like, and an encoder 161 attached to the wheel shaft, for example. A self-position measuring device 172 that measures the current position of the robot (hereinafter referred to as self-position) using the odometry computer 162, and information on the obstacle detected by the obstacle detection sensor 177 and the self-position measuring device 172. To avoid the obstacle detected from the self-position information of the mobile robot 171 measured by the Having a route calculation device 174 for calculating the.

  Such a movement path of the mobile robot 171 is determined as shown in FIG. 8A, for example. Specifically, for example, when the mobile robot 171 is moving toward the destination 183 as shown in FIG. 8A, there is no obstacle entering the detection area 182 of the obstacle detection sensor 177 in the traveling direction. The path 185 shown by the solid line arrow in FIG. 8A indicates that the mobile robot 171 is directed to the destination 183 from the self-position measured by the self-position measuring device 172 and the position information of the destination 183 in the route calculation device 174. And the mobile robot 171 moves along this path 185.

  However, when the mobile robot 171 moves, for example, when the obstacle 184 enters the detection area 182 of the obstacle detection sensor 177 in the traveling direction as shown in FIG. The direction and distance of the obstacle 184 are measured from the mobile robot 171, and the route calculation device 174 measures the obstacle by the obstacle detection sensor 177 in addition to the self-position information measured by the self-position measurement device 172 and the position information of the moving destination 183. The route is calculated based on the information of the obstacle 184 that has been made. The calculated route is, for example, an obstacle avoidance component 186 having a size inversely proportional to the distance between the mobile robot 171 and the obstacle 184 and a direction opposite to the obstacle 184, and a path 185 when the obstacle is not present. The synthesized route 187 is assumed. As described above, the mobile robot 171 moves along the route calculated in real time based on the obstacle information around the mobile robot 171, so that the mobile robot 171 avoids the obstacle 184 and reaches the destination 183.

  Further, unlike the mobile robot 171 shown in FIGS. 7A, 7B, 7C, 8A, and 8B, for example, the mobile robot has destination information and obstacle information within the mobile robot movement range as map information. In some cases, the map information is used as it is to calculate the movement route of the mobile robot (see, for example, Patent Document 2).

  Here, the outline of the mobile robot of the conventional example 2 described in the patent document 2 is demonstrated using FIG. 10A, FIG. 10B, FIG. 10C, and FIG. 10A, 10B, and 10C show the configuration of the mobile robot described in Patent Document 2. FIG. FIG. 11 shows a method for determining a movement path in the mobile robot shown in FIGS. 10A, 10B, and 10C.

  As shown in FIGS. 10A, 10B, and 10C, the mobile robot 201 includes a movable main body 201a including wheels 207, auxiliary wheels 208, and a driving device 205 such as a motor necessary for the movement of the robot 201; For example, an encoder value measured by an encoder 209 attached to a wheel shaft is used to measure a current position of the robot 201 (hereinafter referred to as a self-position) using an odometry computer 230, and where the destination is. located in information as to the position information indicating the obstacle is present anywhere in advance, a map database 203 such as to store the moving range of the map information of the mobile robot 201, the mobile robot 201 measured by the self-position measuring device 202 Avoid obstacles from self-location information and map information in map database 203 Having a route calculation device 204 for calculating the dynamic path.

  The movement path in the mobile robot 201 is determined as shown in FIG. The robot 201 and the destination 303 are set in the map information 301 based on the self-position information measured by the self-position measuring device. Here, for example, the map information is divided into meshes like the map information 301 according to the required moving accuracy, and the obstacle from the block where the mobile robot 201 exists to the block where the mobile destination 303 exists as shown in the enlarged view 307. A path that sequentially follows the block is determined so as not to pass through a certain block of the object 304. Here, since each block has a plurality of movable directions as indicated by 308, there are a plurality of routes exemplified by 309 following the movable direction. For this reason, a plurality of route candidates reaching the destination 303 are calculated as indicated by 305, and therefore the route 306 is uniquely selected according to conditions such as the shortest arrival route.

Japanese Patent Laid-Open No. 7-110711 JP-A-6-138940

  However, in the case of the mobile robot 171 of the conventional example 1 shown in FIGS. 7A, 7B, 7C, 8A, and 8B, the route is calculated based on the obstacle information detected by the obstacle detection sensor 177. However, since there is a limit to the range of the detection area 182 of the obstacle detection sensor 177, the destination of the calculated route cannot be predicted. This results in an inefficient movement in which the mobile robot enters a dead-end passage or obstacle pit. Furthermore, in the worst case, there is a risk of falling into a so-called deadlock state where the user cannot move at the dead end.

  Details thereof will be described with reference to FIG. FIG. 9 shows a case where the mobile robot 171 of the conventional example 1 enters an inefficient movement or a deadlock state.

  As shown in FIG. 9, for example, when an obstacle 194 having an opening 197 through which the mobile robot 171 can pass and a depression 198 that is a dead end 198 exists in the direction of the destination 183, Even if the robot 171 moves along the route 195 toward the destination 183, the destination is dead end. Therefore, the mobile robot 171 takes an ideal route 196 that avoids the entire obstacle 194 in advance. It is desirable to go to 183.

  However, since the detection area 182 of the obstacle detection sensor 177 of the mobile robot 171 is a limited range around the mobile robot 171, as described above, the mobile robot 171 is near the opening 197 in the depression of the obstacle 194. If it is detected that there are obstacles on both sides of the mobile robot 171 and the width of the opening 197 is a width that allows the mobile robot 171 to pass through, the path 195 enters the back side of the opening 197. It will end up. At this time, the dead end 198 of the depression has not been detected yet. Then, the mobile robot 171 moves along the path 195, reaches the dead end 198, detects that it is finally impossible to pass, and escapes from the depression to select a different path. Moreover, in the worst case, a moving component that tries to move in the direction of the path 195 at the dead end and a component that tries to avoid the dead end are combined, so that it becomes impossible to move and a deadlock state may occur. is there.

  On the other hand, in the mobile robot 201 of Conventional Example 2 shown in FIGS. 10A, 10B, 10C, and 11, the route is calculated based on the map information stored in the map database 203. This map information Since the route calculation using is performed for the entire moving range, for example, when there are many obstacles 304 or the map information is vast, the amount of calculation in the route calculation is enormous. It becomes difficult to perform processing in real time by a processor mounted on a small mobile robot.

For example, as described in the above example, the calculation of the shortest route from a certain point to the moving destination 303 requires a calculation amount proportional to the square of the reference point (reference area / movement accuracy) by graph theory. It is known that there is. For example, it is necessary to have 100 × 100, that is, 10,000 reference points even if only 1 m square is moved with an accuracy of 1 cm, and the calculation amount necessary for route calculation is K × 10 ⁸ (K is a proportional constant) and a huge calculation amount. It becomes. On the other hand, as a comparison, the calculation amount in the case of the robot of Conventional Example 1 is a calculation amount proportional to the number of sensors.

  Accordingly, an object of the present invention is to solve such a problem, and in a mobile robot that moves in an environment where an obstacle exists, it is possible to realize a movement that can realize a real-time and efficient movement to a destination. To provide a robot.

  In order to achieve the above object, the present invention is configured as follows.

According to the first aspect of the present invention, a movable robot body,
A self-position measuring device for measuring the self-position of the main body,
A map database for storing map information of the movement range of the main body,
An obstacle set on the map information based on the self-position information measured by the self-position measuring device and the map information stored in the map database and serving as an obstacle to the movement of the main body An obstacle information extraction unit that extracts the obstacle information in a detection area of a virtual sensor capable of detecting the object information;
A route calculation device that calculates a movement route for the main body to move based on the obstacle information extracted by the obstacle information extraction unit;
A virtual sensor setting change device for setting and changing an extraction condition for the obstacle information extraction unit to extract the obstacle information;
A mobile robot is provided.

  According to the second aspect of the present invention, it further includes an obstacle detection sensor that detects an obstacle in a detection region around the main body, and the route calculation device is extracted by the obstacle information extraction unit. In addition to obstacle information, the mobile robot according to the first aspect is provided that calculates a movement path for the main body to move based on detection information from the obstacle detection sensor.

According to a third aspect of the present invention, the virtual sensor setting change device is configured to extract the map information stored in the map database, the self-position information measured by the self-position measurement device, and the obstacle information extraction. The obstacle information extraction unit performs the obstacle information based on at least one of the obstacle information extracted by a part, detection information of the obstacle detection sensor, and a movement route calculated by the route calculation device. The mobile robot according to the first or second aspect, in which the extraction condition for extracting the image is changed is provided.

  According to such a configuration, sensor information is calculated virtually from the map information, so that “there is a limit to the distance and range to the obstacle that can be detected” or “cannot be detected due to the surface physical properties of the obstacle” or “sensor Because there is no physical limitation peculiar to actual obstacle detection sensors such as `` cannot detect each other due to interference '', it is possible to set a free detection area according to the obstacle environment and detect obstacles with high accuracy Can do. Therefore, it is possible to realize an efficient movement to the destination as compared with a case where only an actual obstacle detection sensor is used. Further, the total calculation amount required for the route calculation of the present invention is proportional to the calculation amount in the obstacle information extraction unit because whether or not there is an obstacle in the detection region (sensor detection region / movement accuracy). This is the sum of the calculated amount and the calculated amount of the route calculation of the conventional example 1, so that the amount of calculation can be greatly reduced as compared with the case of calculating the route directly from the map, and it is mounted on a small mobile robot. It is also possible to perform processing in real time with such a processor.

  As described above, according to the present invention, by extracting obstacle information in the detection area of the virtual sensor from the map information, it is possible to detect obstacles that cannot be detected due to physical limitations of the actual obstacle sensor. Furthermore, since the amount of calculation at the time of route calculation is greatly reduced compared to the case of calculating the route directly from the map information, even a processor mounted on a small mobile robot can perform processing in real time. Can do.

Embodiments according to the present invention will be described below in detail with reference to the drawings.
Before continuing the description of the present invention, the same parts are denoted by the same reference numerals in the accompanying drawings.

  A mobile robot according to an embodiment of the present invention will be described with reference to FIGS. 1A to 1E.

  FIG. 1A is a diagram showing a configuration of the mobile robot 20 of the present embodiment, FIG. 1B is a diagram showing an actual mobile robot 20 and a first detection area 3 of its obstacle detection sensor 4, and FIG. 1C is a map 13 It is a figure which shows the 2nd detection area | region 21 of the upper mobile robot 20 and its virtual sensor (it mentions later for details). 1D is a diagram showing an avoidance path B calculated by the mobile robot 20, and FIG. 1F is a diagram showing a configuration of a mobile robot 20B different from the mobile robot 20 shown in FIG. 1A.

  As shown in FIG. 1A, the mobile robot 20 includes a movable rectangular parallelepiped main body 2, and a plurality of obstacle detection sensors 4 (four are arranged on both sides of the main body 2 in FIG. 1A). The self-position measuring device 5, the map database 11, the obstacle recognizing device 22, the route calculating device 6, and the driving device 7 are provided.

  The main body 2 is configured to be movable so as to include four wheels 2w necessary for movement of the robot 20 and a driving device 7 such as a motor.

  The plurality of obstacle detection sensors 4 (four in FIG. 1A arranged at the upper part on both sides of the main body 2) are attached to the periphery of the main body 2 and use ultrasonic waves, infrared rays, lasers, or the like. An obstacle present in the first detection area 3 formed around the area 2 is detected.

  The self-position measuring device 5 measures the current position of the robot 20 (hereinafter referred to as self-position) using an encoder value measured by an encoder attached to the wheel shaft of the wheel 2w, for example.

  The map database 11 has information indicating where to find the destination 8, the position information indicating the obstacle 9, 16 are present anywhere in advance, a map information of the movement range of the body portion 2.

  The obstacle recognition device 22 detects known obstacles 9 and 16 recorded in the map database 11 and existing in the second detection area 21 of the virtual sensor in the memory area 12.

  The route calculation device 6 includes information on the obstacles 9 and 16 recognized by the obstacle recognition device 22 and an unknown obstacle 23 existing around the main body 2 detected by the obstacle detection sensor 4 (see FIG. 1B). And an avoidance route (a route for avoiding an obstacle and moving to a moving destination) B in the mobile robot 20 is calculated based on the above information and the self-position information of the mobile robot 20 measured by the self-position measuring device 5.

  The driving device 7 moves the main body 2 based on the calculated avoidance path B.

  The obstacle recognizing device 22 forms a map (map information) 13 as map graphic data in the memory area 12, forms the known obstacles 9 and 16, the main body 2 on the map 13, and And setting a virtual sensor having a second detection area 21 capable of detecting known obstacles 9 and 16 existing in a second detection area 21 different from the first detection area 3 in the main body 2. Thus, the obstacles 9 and 16 existing in the second detection area 21 of the virtual sensor are detected on the map 13.

  The virtual sensor is not an actual sensor, but can set a second detection region 21 having a detection function virtually equivalent to the sensor on the map 13. The obstacle recognition device 22 The known obstacles 9 and 16 existing in the second detection area 21 can be recognized and extracted. More specifically, the second detection area 21 of the virtual sensor has a circle drawn in front of the mobile robot 20 when the mobile robot 20 rotates when turning (when turning to avoid an obstacle or the like). A distance that is larger than the turning radius of the obstacle, and that is longer than the depth of the obstacle that has the largest depression as viewed from the mobile robot 20 among obstacles that are known in advance on the map 13. A triangular or rectangular region with a This region may be the maximum region, and as one aspect, the second detection region 21 of the virtual sensor may be always set to this maximum region. Alternatively, as another aspect, as will be described in detail later, when no obstacle is detected, the mobile robot 20 is instructed to stop when the mobile robot 20 is moving along the movement path of the mobile robot 20 that is smaller than the maximum area. Is the length of the distance traveled by the mobile robot 20 before stopping (however, the distance varies depending on the speed of the mobile robot 20), and when the mobile robot 20 turns (for avoiding obstacles, etc.) An area having a width through which a circle drawn at the time of rotation (at the time of turning) passes (a width of at least twice the radius of rotation) is set as a minimum area, and this minimum area is set as a second detection area 21 to detect an obstacle. In some cases, the maximum area may be changed. Further, on the map 13, an area where the second detection area 21 of the virtual sensor is set as the minimum area (for example, an area assumed to have relatively few obstacles), and the second detection area 21 of the virtual sensor are displayed. An area to be set as the maximum area (for example, an area that is assumed to have a relatively large number of obstacles) may be set in advance.

  According to such a virtual sensor, unlike an actual sensor, it is possible to freely set a region and detection accuracy regardless of physical constraint conditions. For example, a virtual sensor can detect a long-distance area or a large area that cannot be detected by an actual sensor, an area on the back side of an object, a complicated and complicated area, and the like. Furthermore, it is possible for a virtual sensor to acquire information with such high accuracy as impossible with an actual sensor. Also, with a virtual sensor, there is no need to worry about interference between sensors, and problems such as mounting positions, sensor drive sources, wiring, and piping that occur when using actual sensors are also virtual sensors. If so, there is no need to think about it. Furthermore, if it is a virtual sensor, the setting of the sensor, for example, the detection area and detection accuracy, the program for setting the virtual sensor is switched to another program, or the parameter value used in the program for setting the virtual sensor is changed. It is possible to change the setting freely by doing so. Therefore, when using a virtual sensor, it is possible to set detection settings with coarse accuracy during normal times, and to change detection settings with higher accuracy when obstacles are found. Or, when using a virtual sensor, the detection range is usually expanded in a place where it is known in advance that there are many obstacles by setting to detect only a narrow front region of the robot 20, and the detection is performed on the entire periphery of the robot 20. It can be changed to such a setting. In this way, it is possible to set the detection accuracy and area of the virtual sensor according to need, in other words, according to time and place.

  The obstacle detection by the virtual sensor having such a configuration is performed by detecting information on each location in the second detection area 21 in the map 13 (information on presence of an obstacle at each location, for example, the second detection area 21). Information on the presence or absence of obstacles, the shape of obstacles in the second detection area 21, the distance and direction from the robot 20 to the obstacles in the second detection area 21) from the map database 11. This is made possible by partially acquiring the signal. Therefore, in the case of a normal obstacle detection sensor 4, as shown by X in FIG. However, in the virtual sensor as described above, the information in the second detection area 21 is partially acquired from the map database 11, so that the obstacle Identification device 22 is able to detect an area X as a behind the obstacle 9 to such mobile robot 20. That is, the obstacle recognizing device 22 is in the positional relationship between the obstacles 9 and 16 if the obstacles are known obstacles 9 and 16 registered in advance in the map database 11 and are present in the second detection area 21. Regardless (in other words, even if the obstacles overlap each other when viewed from the mobile robot 20), all of them can be detected. Note that the second detection area 21 in the virtual sensor is, for example, closer to the main body part 2 than the first detection area 3 so that the traveling direction side of the main body part 2 can be detected from the first detection area 3. It is set in a triangular shape that extends farther toward the direction of travel.

  Further, as described above, the obstacle recognition device 22 recognizes only obstacles existing in the second detection region 21 and recognizes obstacles existing outside the second detection region 21. Do not do. Thereby, only a part of the map 13 within the range of the second detection area 21 can be set as a calculation target when the avoidance route of the mobile robot 20 is calculated.

  A method of calculating the avoidance route B that prevents the mobile robot 20 from entering the obstacle 16 having no exit in such a configuration will be described. It is assumed that the position information of the obstacle 16 is registered in the map database 11.

  Here, as shown in FIG. 1B, in the mobile robot 20 that is actually traveling along the route A toward the destination 8, the current position of the mobile robot 20 is measured by the self-position measuring device 5 and obtained. The self-location information and the map information in the map database 11 are sent to the obstacle recognition device 22. In the obstacle recognizing device 22, as shown in FIG. 1C, a map 13 is formed based on the sent map information in the memory area 12, and the destination 8 and obstacles registered in advance in the map database 11 are formed. 9 and 16 are formed on the map 13. In addition, a mobile robot 20 is formed on the map 13 based on the self-position information sent from the self-position measuring device 5, and a virtual sensor is set on the mobile robot 20, and a second detection area 21 of the virtual sensor is displayed. Is set. In addition, there are known obstacles 9 and 16 registered in advance in the map database 11 and unknown obstacles 23 not registered in the map database 11 around the mobile robot 20 that is actually moving. The obstacle 16 has an opening 16a through which the mobile robot 20 can enter and a dead end 16b on the back side of the opening 16a. The destination 8 is located behind the obstacle 16. Is present.

  Such a moving path of the mobile robot 20 enters the first detection area 3 of the obstacle detection sensor 4 in the traveling direction while the mobile robot 20 is moving toward the destination 8, for example. When there is no obstacle, the mobile robot 20 calculates a route in the route calculation device 6 so as to go to the destination 8 from the self-position measured by the self-position measurement device 5 and the position information of the destination 8. The mobile robot 20 moves according to the route.

  At this time, the obstacle detection sensor 4 of the moving mobile robot 20 detects a part of the obstacle 16 on the opening 16a side and a part of the obstacle 23, and the route calculation device 6 in the mobile robot 20 detects the obstacle. If the avoidance route is calculated based only on the information of the obstacle 23 detected by the obstacle detection sensor 4, as already described in the prior art, the obstacle detection sensor 4 falls within the range of the first detection region 3. Since there is a limit, it is impossible to predict the destination of the calculated route, and as shown in FIG. 1B, an avoidance route that goes to the inside of the obstacle 16, that is, an avoidance route in the same direction as the route A As a result, the mobile robot 20 may fall into a deadlock state or a permanent motion inside the obstacle 16.

  However, in the mobile robot 20 of this embodiment, as shown in FIG. 1C, the obstacle recognition device 22 exists in the second detection area 21 of the virtual sensor set for the mobile robot 20 on the map 13. Since the known obstacle 16 is recognized and the second detection region 21 can be arbitrarily set, as shown in FIG. 1C, it is farther toward the traveling direction side of the main body 2 than the first detection region 3. If the second detection area 21 is set in a triangular shape that extends to the far side, it is possible to detect the inner side of the obstacle 16 and detect that the dead end portion 16b exists on the inner side of the obstacle 16. can do.

  As described above, by using the actual obstacle detection sensor 4, it is possible to detect the unknown obstacle 23 that cannot be detected by the virtual sensor, that is, not registered in the map database 11, and to use the virtual sensor. Thus, a portion of the known obstacle 16 that is not covered by the first detection region 3 of the obstacle detection sensor 4 can be detected. Therefore, by using the virtual sensor, it is possible to perform obstacle detection with higher accuracy than when only the obstacle detection sensor 4 is used.

  Then, the self-position information measured by the self-position measuring device 5, the information of the obstacle 23 not registered in the map database 11 obtained in advance by the obstacle detection sensor 4, and the obstacle recognition device 22 are detected. Information about the known obstacle 16 is sent to the route calculation device 6.

  Then, in the route calculation device 6, as shown in FIG. 1D, based on the self-position information of the mobile robot 20 and the position information of the obstacle 23 and the obstacle 16, An avoidance route B that can avoid the obstacle 16 having the dead end portion 16b is calculated.

  At this time, as described above, the obstacle recognizing device 22 recognizes only the obstacle existing in the second detection area 21 and recognizes the obstacle existing outside the second detection area 21. Do not do. Thereby, the route calculation device 6 can set only a part of the map 13 within the range of the second detection region 21 as a calculation target when calculating the avoidance route of the mobile robot 20. Compared with the case where the avoidance route is calculated using the entire range as the calculation target, the amount of calculation can be greatly reduced.

  Thereby, even if it is a processor mounted in a small mobile robot, a avoidance path | route can be calculated rapidly. Even if it is necessary to detect a new obstacle while moving and calculate the avoidance route again, as described above, the calculation amount is greatly reduced. Calculations can be made. For the route calculation, the detection information of the virtual sensor and the detection information of the actual obstacle detection sensor 4 can be handled as the same sensor information, and a function using the sensor information as an input (for example, given in Conventional Example 1) Thus, the sensor information, that is, a function that calculates the path of the mobile robot by adding the correction amount of the movement component according to the direction and distance of the obstacle to the movement component to the destination is solved. To calculate the route. An example of such a function is as follows.

(Equation 1)
Do (robot path) = F ([sensor information])
Example: F ([Sensor information]) =
Dt (moving component to the destination)
+ G (avoidance gain) * L1 (distance to obstacle 1) * D1 (direction of obstacle 1)
+ G (avoidance gain) * L2 (distance to obstacle 2) * D1 (direction of obstacle 2)
+. . .
(Continued by the number of obstacles)

Do, Dt, D1, D2 ... are vectors.

  In the mobile robot 20 shown in FIG. 1A, the obstacle recognition device 22 performs data processing for recognizing a known obstacle 9 on the map 13, and the route calculation device 6 performs data processing for calculating an avoidance route. However, these data processes may be performed by a single computing device (not shown). In this case, delivery of obstacle detection information by the virtual sensor is performed using a memory in the apparatus or an internal communication function.

  Further, as shown in FIG. 1E, the obstacle detection sensor 4 actually detects the obstacle 9 in the obstacle recognition device 22 of the mobile robot 20 </ b> B based on information on the known obstacle 9 recognized using the virtual sensor. You may provide the converter 24 which converts into the same signal as the output signal at the time, for example, the same kind of signal. In this case, since the output signal of the virtual sensor can be made the same as the output signal of the actual obstacle detection sensor 4 by the conversion device 24, for example, the setting of the second detection area 21 in the virtual sensor is changed. Thus, the effect of adding a sensor or changing its installation position can be verified without actually adding a sensor or changing its installation position. Conversely, a virtual sensor can be easily used instead of the actual obstacle detection sensor 4. Thereby, even in a mobile robot that is in an experimental stage or being adjusted, it is possible to verify the effect of adding a sensor or the like without actually adding a sensor or changing its installation position.

  According to the embodiment, the map graphic data (map information) is formed based on the map database 11, and the known obstacle and the main body 2 are formed on the graphic data. By setting a virtual sensor capable of detecting the known obstacles 9 and 16 existing in the second detection area 21 different from the first detection area 3 by the actual sensor, for example, provided in the main body 2 Even if the first detection area 3 of the obstacle detection sensor 4 is not covered, if it is a known obstacle 9, 16 whose position information is stored in the map database 11, it is detected by this virtual sensor. By using a virtual sensor, it is possible to perform obstacle detection with higher accuracy than when only the obstacle detection sensor 4 provided in the main body 2 is used. Further, the second detection area 21 in the virtual sensor detects the known obstacles 9 and 16 that have entered the second detection area 21, and the known obstacle that exists outside the second detection area 21. Since the objects 9 and 16 are not detected, when the route calculation device 6 calculates the avoidance route, information on the known obstacles 9 and 16 that have entered the second detection area 21 in the virtual sensor and the main body Since the avoidance route can be calculated based on the unknown obstacle information among the obstacles detected by the obstacle detection sensor 4 in the unit 2, for example, all the graphic data of the map is calculated as the route calculation. Compared with the case where the calculation is performed at the time, the amount of calculation can be greatly reduced.

  Therefore, by providing a virtual sensor when calculating the avoidance route, the amount of calculation at the time of route calculation is greatly reduced, so even a processor mounted on a small mobile robot can avoid it in real time. A route can be calculated. Further, even if it is necessary to detect a new obstacle during the movement and calculate the avoidance route again, the new avoidance route can be similarly calculated in real time.

  Next, a more specific example of the mobile robot according to the embodiment of the present invention will be described with reference to FIGS. 1F to 6B.

  As shown in FIG. 1F, FIG. 1G, and FIG. 1H, a mobile robot 51 of a more specific example of the above embodiment of the present invention includes a movable rectangular parallelepiped main body 51a and positions of the main body 51a. A self-position measuring device 53 for measuring, a map database 52 for storing map information of a moving range to the moving destination of the main body 51a, and a calculation for the virtual sensor information calculating device 54 to calculate virtual sensor calculation information Arbitrary detection on the map information based on the virtual sensor setting changing device 57 for setting and changing the condition, the self-position information 73 measured by the self-position measuring device 53 and the map information stored in the map database 52 The virtual information which extracts the obstacle information which becomes an obstacle with respect to the movement of the main body 51a in the region and calculates the calculation information of the virtual sensor under the set calculation condition Sensor information calculation device 54 (in other words, based on the self-position information 73 measured by the self-position measurement device 53 and the map information stored in the map database 52, the main body portion in an arbitrary detection region on the map information) An obstacle information extraction unit that extracts obstacle information that becomes an obstacle to the movement of 51a, and virtual sensor calculation information calculated by the virtual sensor information calculation device 54 (in other words, the obstacle information extraction unit) And a route calculation device 55 that calculates a movement route for the main body 51a to move based on the obstacle information extracted by the computer 54. Further, the mobile robot 51 further includes an input device 39 for inputting obstacle information related to obstacles, information related to virtual sensor settings, and information related to the destination to the map database 52 and the virtual sensor setting changing device 57, and various types of information. And an output device 38 such as a display for outputting information (for example, map information, virtual sensor setting information, movement route information, etc.).

  Here, the main body 2 of the mobile robot 20 in FIG. 1A corresponds to the main body 51a of the mobile robot 51. Similarly, the obstacle detection sensor 4 corresponds to the obstacle detection sensor 56, and The position measurement device 5 corresponds to the self-position measurement device 53, the map database 11 corresponds to the map database 52, the obstacle recognition device 22 corresponds to the virtual sensor setting change device 57 and the virtual sensor information calculation device 54, and the route calculation. The device 6 corresponds to the route calculation device 55, and the driving device 7 corresponds to the driving device 61.

  The obstacle detection sensor 56 detects an obstacle in an arbitrary detection region around the main body 51a, and the route calculation device 55 moves based on the detection information of the obstacle detection sensor 56 in addition to the calculation information of the virtual sensor. It is assumed that the route can be calculated.

  Further, the robot 51 has the virtual sensor setting changing device 57 for setting and changing the calculation conditions for the virtual sensor information calculating device 54 to calculate the calculation information of the virtual sensor. Map information stored in the map database 52, self-position information 73 measured by the self-position measurement device 53, calculation information of a virtual sensor calculated by the virtual sensor information calculation device 54, and the obstacle sensor It is also possible to change the calculation conditions for calculating the calculation information of the virtual sensor based on the 56 detection information and the movement route calculated by the route calculation device 55.

  Here, as a specific specification example of the robot 51, the movable main body 51a shown in FIGS. 1F, 1G, and 1H includes two left and right drive wheels 59 that can be driven independently and a caster type auxiliary. It is assumed that the moving device 58 is composed of two small auxiliary wheels 60. The left and right drive wheels 59 can be controlled at a predetermined rotational speed by the drive device 61 using the left and right motors 61a, respectively, and can be turned and turned depending on the difference in rotational speed between the two drive wheels 59. . The main body 51a has a shape approximating a vertically long rectangular parallelepiped, and the left and right drive wheels 59 and the left and right auxiliary wheels 60 are arranged at four corners. The front two wheels are the drive wheels 59 and the rear two wheels are The auxiliary wheel 60 is used. These two drive wheels 59 and two auxiliary wheels 60 correspond to the four wheels 2w in FIG. 1A.

  The self-position measuring device 53 includes an encoder 62 attached to the rotational drive shafts of the two drive wheels 59, and an odometry calculation device 63 that calculates the self-position from the value of the encoder 62. The route calculation device 55 calculates real-time self-position information 73 of the robot 51 by performing odometry calculation based on the rotational speeds of the two drive wheels 59 obtained from the two encoders 62. The calculated position measurement information is specifically the position and posture (traveling direction) of the main body 51a of the robot 51. Further, the speed information of the robot 51 can be calculated based on the difference of the time-series self-position information 73.

  A plurality of photoelectric sensors 64 and an ultrasonic sensor 65 are used as the obstacle detection sensor 56 for obstacle detection.

  As shown in FIG. 2, a plurality of photoelectric sensors 64 each capable of detecting a substantially rectangular detection region such as 64s are arranged around the main body 51a of the robot 51 (specifically, the main body 51a The proximity region is detected so as to surround the body 51a of the robot 51 (one at the center of the front and rear surfaces and two at the center of the left and right sides). In addition, a plurality of ultrasonic sensors 65 having a long and narrow detection area such as 65s are arranged in front (specifically, two are arranged on the front surface of the main body 51a), and the obstacle 40 in front is detected. As the detection values by the obstacle detection sensors 56, the photoelectric sensor 64 uses the inaccessible area 40a-6 of the robot 51 as the detection value, and the ultrasonic sensor 65 uses the distance L to the obstacle 40 as the detection value. Therefore, the first detection region 56s of the obstacle detection sensor 56 (the first detection region 4 of the obstacle detection sensor 4 of the mobile robot 20 in FIG. 1A) is detected from the detection region 64s of the photoelectric sensor 64 and the detection region 65s of the ultrasonic sensor 65. 1 corresponding to the detection area 3).

  As map information 70 stored in the map database 52, as shown in FIG. 4A, obstacle information 72 relating to the position, size, and shape of the obstacle 40 and destination information 71 are registered. When calculating the virtual sensor, the information of the mobile robot 51 is superimposed on the map information 70 based on the self-position information 73.

  In addition, the calculation conditions of the virtual sensor are set (setting of the second detection area 41 of the virtual sensor (corresponding to the second detection area 21 of the virtual sensor of the mobile robot 20 in FIG. 1A)). As shown in FIG. 3A, in the mobile robot 51 without the virtual sensor setting changing device 57 as shown in FIG. 1I, the second detection area 41 is located in front of the mobile robot 51 when the mobile robot 51 turns (obstacles). The obstacle 40 has a width that allows a circle to be drawn at the time of rotation necessary for turning (for turning such as avoiding the rotation) (a width of at least twice the turning radius 42) and is known in advance in the map information 70. The obstacle 40G having the largest depression as viewed from the mobile robot 51 is a rectangular area having a distance longer than the depression depth 40G-1. Here, it is assumed that the depth 40G-1 of the depression of the obstacle 40G is so deep that it cannot reach the elongated detection region 65s of the ultrasonic sensor 65. 1I can input the virtual sensor setting information from the input device 39 to the virtual sensor information calculation device 54 to arbitrarily set the second detection area 41 of the virtual sensor. However, in this example, the setting of the virtual sensor cannot be changed while the robot is moving.

  In addition, since the mobile robot 51 includes the virtual sensor setting change device 57 that changes the calculation conditions for the virtual sensor information calculation device 54 to calculate the calculation information of the virtual sensor, the mobile robot 51 starts moving. Sometimes, once the second detection area 41 of the virtual sensor is set, the area can be fixed, and various information such as obstacle information input from the obstacle detection sensor 56 to the virtual sensor setting changing device 57 can be used. The setting of the second detection area 41 of the virtual sensor can be changed by the virtual sensor setting changing device 57 during the movement of the mobile robot 51 based on the information.

  An example in which the setting of the second detection area 41 of the virtual sensor is changed by the virtual sensor setting changing device 57 during the movement of the mobile robot 51 based on various types of information, for example, obstacle detection information will be described.

  As the second detection area 41 of the virtual sensor, for example, as shown in FIG. 3B, when the mobile robot 51 is moving normally (before obstacle detection), the robot 51 is moving along the movement path. When the robot 51 turns (avoids obstacles, etc.) with a length 43 that is a distance traveled until the robot 51 stops after the stop command is issued to the robot 51 (however, the distance varies depending on the speed of the mobile robot 51). For example, the region 41g can have a width through which a circle drawn at the time of rotation (a width of at least twice the rotation radius 42) passes.

  Then, as shown in FIG. 3B, for example, when an obstacle 40G enters the second detection area 41g when the virtual sensor is normal (before obstacle detection), map information is obtained from the detection position of the obstacle 40G. The shape and position of the obstacle 40G detected from 70 are extracted by the virtual sensor setting changing device 57. This time, in addition to the normal detection area 41g, the robot 51 around the obstacle 40 turns (obstruction). An additional area 41h having a width (at least twice the rotation radius 42) through which a circle drawn during rotation during turning to avoid an object or the like can also be changed to the second detection area 41.

  In addition, after the robot 51 avoids the obstacle, as shown in FIG. 3B, when the mobile robot 51 returns to the normal movement (before the obstacle discovery), the additional area 41h is eliminated, and the normal detection area You may make it only 41g.

  Then, the virtual sensor calculation information calculated by the virtual sensor information calculation device 54 while performing detection by the second detection region 41 of the set virtual sensor is the robot 51 measured by the self-position measurement device 53. The robot 51 is moved in the second detection area 41 of the virtual sensor extracted based on the self-location information 73 and the map information 70 stored in the map database 52 and set on the map information 70. It means the information that becomes an obstacle to it. As a specific example of the calculation information of the virtual sensor, information on the obstacle can be taken into consideration and the robot 51 can move, and the distance from the mobile robot 51 to the obstacle 40 is taken into consideration. And an angle range in which the mobile robot 51 can move. Such information on the distance from the mobile robot 51 to the obstacle 40 and the angle range in which the mobile robot 51 can move is shown in the figure depending on the presence or absence of an obstacle in the second detection area 41 as follows. 3C to FIG. 3F and input to the route calculation device 55. 3C to 3F exemplify the case where the second detection area 41 of the virtual sensor is not changed during the movement by the virtual sensor setting change device 57, but even when the second detection area 41 is changed during the movement as described above. The same calculation is performed.

    (1) Based on the map information 70 stored in the map database 52 and the self-position information of the mobile robot 51 measured by the self-position measuring device 53, as shown in FIG. When the virtual sensor information calculation device 54 can determine that there is no 40, the virtual sensor information calculation device 54 is omnidirectional in front of the mobile robot 51 such that the obstacle distance is infinite (∞) and the movable angle is 41c-3. The calculation information is calculated. Here, the second detection area 41 of the virtual sensor is a rectangular detection area 41c-2 extending in front of the mobile robot 51, and the following (2) to (4) are similar areas. ing.

    (2) As shown in FIG. 3D, based on the map information 70 stored in the map database 52 and the self-position information of the mobile robot 51 measured by the self-position measuring device 53, the mobile robot 51 is moved forward by 2 When the virtual sensor information calculation device 54 can determine that there are two opposing obstacles 40d-6 and 40d-7, the vehicle can pass between the two obstacles 40d-6 and 40d-7. When the virtual sensor information calculation device 54 determines that a simple path 40d-5 is formed, the distance from the mobile robot 51 to the obstacle 40d-6 near the front of the mobile robot 51 is determined from the mobile robot 51 to the obstacle 40d. The distance between the two obstacles 40d-6 and 40d-7 is an angle direction that enters the path 40d-5 and an angle direction that avoids the path 40d-5. The angular range 40d-3, the calculation information to be movable angle of the mobile robot 51, the virtual sensor information calculation unit 54 calculates. Here, whether or not the passage 40d-5 through which the robot 51 can pass is formed between the two obstacles 40d-6 and 40d-7 can be determined by the virtual sensor as follows. For example, algorithmically, if there are two obstacles arranged opposite to each other, the distance between the two obstacles is set to (the overall width of the robot 51) (the dimension for the safety margin). ) Is determined by the virtual sensor information calculation device 54, and if it is determined by the virtual sensor information calculation device 54 that it is more than such a width size, the robot 51 is allowed to pass. While the process of FIG. 3D is performed, if the virtual sensor information calculation apparatus 54 determines that the width is less than such a width dimension, the robot 51 performs the process of FIG. Further, information such as the width and length of the robot 51 and the turning radius when turning is included in the information when setting the virtual sensor.

    (3) As shown in FIG. 3E, based on the map information 70 stored in the map database 52 and the self-position information of the mobile robot 51 measured by the self-position measuring device 53, the mobile robot 51 is directly in front of the front. When the virtual sensor information calculation device 54 can determine that there is an obstacle 40e-6, the distance from the mobile robot 51 to the obstacle 40e-6 in the front direction of the mobile robot 51 is determined from the mobile robot 51 to the obstacle 40e-. The virtual sensor information calculation device 54 calculates the calculation information that sets the distance 40e-4 to 6 and the movable robot 51 in the angle direction 40e-3 that avoids the obstacle 40e-6.

    (4) As shown in FIG. 3F, based on the map information 70 stored in the map database 52 and the self-position information of the mobile robot 51 measured by the self-position measuring device 53, the mobile robot 51 is directly in front of the front. When the virtual sensor information calculation device 54 can determine that there is an obstacle 40f-6, and the depression of the obstacle 40f-6 stops due to the movement of the mobile robot 51, the passage 40f-7 or a path 40f that cannot pass through is displayed. When the virtual sensor information calculation device 54 can determine that the value is -5, the obstacle 40f-6 is considered as an obstacle 40f-6 in which the opening of the depression of the obstacle 40f-6 is blocked, and the mobile robot 51 moves in front of the mobile robot 51. The distance from 51 to the obstacle 40f-6 is set as the distance 40f-4 from the mobile robot 51 to the obstacle 40f-6. The angular orientation 40f-3 of avoiding 40f-6, the calculation information to be movable angle of the mobile robot 51, the virtual sensor information calculation unit 54 calculates.

  Next, the route calculation device 55 calculates the movement route of the mobile robot 51 as shown in FIGS. 4B and 4C.

  First, based on the map information 70 stored in the map database 52 and the self-position information of the mobile robot 51 measured by the self-position measurement device 53, if there is no obstacle in the traveling direction of the robot 51, the virtual sensor information calculation device If determined at 54 (as shown in FIG. 3C), as shown in FIG. 4B, the direction 71b-3 to the destination 71b-2 connecting the mobile robot 51 and the destination 71b-2, and the self-position An angle difference in the traveling direction 51b-4 of the current mobile robot 51 measured by the measuring device 53 is calculated by the route calculation device 55, and a turning speed component 51b-5 proportional to the angle difference is added to the straight speed component. The travel route 51b-6 is calculated by the route calculation device 55, and the straight speed component of the robot 51 is set by the distance to the obstacle and the destination and the turning speed component.

  As described above, the route calculating device 55 can move along the moving route 51b-6 by calculating such a moving speed. The moving speed calculated by the route calculating device 55 is input to the driving device 61, and the mobile robot 51 moves. If there are no obstacles, the robot 51 moves at the maximum speed.

  Here, the direction 71b-3 toward the destination 71b-2 connecting the mobile robot 51 and the destination 71b-2 can be individually obtained by the virtual sensor information calculation device 54 or the route calculation device 55 as necessary. . In the virtual sensor information calculation device 54, when the detection setting of the virtual sensor detects the area of the direction 71b-3 toward the destination 71b-2 as the second detection area 41, the direction 71b toward the destination 71b-2 -3 is calculated. In this case, the direction 71b-3 toward the destination 71b-2 can be calculated by the virtual sensor information calculation device 54 from the self-location information sent to the virtual sensor and the destination information of the map information. On the other hand, the route calculation device 55 also calculates a direction 71b-3 to the destination 71b-2 to be used for route calculation (here, the current traveling direction of the robot 51 and the target (the destination 71b). -2) is also used for calculating the angle difference with respect to 2). As in the case of the virtual sensor information calculation device 54, the route calculation device 55 can also calculate from the self-location information and the information on the destination of the map information. Further, when the current traveling direction 51b-4 of the mobile robot 51 is obtained by the self-position measuring device 53, it can be calculated by a method called odometry, for example. In the case of this example, the position and direction of the robot 51 can be calculated by integrating the rotational speeds of both wheels of the robot 51.

Further, both the turning speed component 51b-5 and the straight traveling speed component can be obtained by the route calculation device 55. By the way, although it is possible to set parameters such as various gains, here, it is assumed that necessary numerical values are included in the algorithm in advance, and the description of the setting device etc. is omitted to simplify the explanation. To do. As for the turning speed component, as described in this specification, the difference between the “current traveling direction” and the “destination direction” is obtained (or the “traveling direction” and the “movable and inaccessible area”). Except for the angle closest to the destination direction))),
The magnitude obtained by multiplying the difference by a proportional gain is defined as a turning speed component. By doing so, direction control is performed such that the robot 51 faces the direction of the destination. Further, the linear velocity component can be calculated as follows. First, the traveling speed is set according to the distance to the destination and the distance to the obstacle. Here, the traveling speed is the maximum rotational speed that the robot motor can continuously generate, and the maximum speed is defined as the distance at which deceleration starts when near the destination or near an obstacle. As Xd, for example, if the distance between the destination and the obstacle is x, if there is no destination or obstacle during the distance Xd, the speed is 100%. If there is a destination or obstacle during the distance Xd, it is given by the following formula.

(Equation 2)
[Advance speed] = [Maximum speed] * (1-[Deceleration gain] * (Xd-x))

  Also, although it is a linear velocity component, if the robot tries to move at a large linear velocity when the above-mentioned turning component is large, the robot may fall to the outside in the turning direction due to centrifugal force.

(Equation 3)
[Linear speed component] = [Progression speed] * (1-[Swivel deceleration gain] * | Turning speed component |)

  Further, as described above, the maximum speed is defined as the “maximum speed”, which is the maximum speed at which the motor of the robot 51 can continuously output. Specifically, in the case of this example, it can be calculated by the following formula.

(Equation 4)
[Maximum speed] = [Wheel radius] * [Maximum motor speed] * [Gear ratio]

  Further, it is assumed that the setting relating to the maximum speed is included in the algorithm of the route calculation device 55 as described above.

  Further, as shown in FIG. 4C, an obstacle is present in the traveling direction of the robot 51 based on the map information 70 stored in the map database 52 and the self-position information of the mobile robot 51 measured by the self-position measuring device 53. When it is determined by the virtual sensor information calculation device 54 that it exists, and there is information on the obstacle 40 in the calculation information of the obstacle detection sensor 56 or the virtual sensor (as shown in FIGS. 3D to 3F) In the case), the route calculation device 55 calculates the following route.

  As shown in FIG. 4C, when the obstacle 40c-9 exists in the direction of the destination or in the vicinity of the robot 51, the obstacle detection is performed within the movable angle 51c-7 of the robot 51 calculated as the calculation information of the virtual sensor. The turning speed component 51c− is in an angle range excluding the inaccessible area 40c-8 detected by the sensor 56 and in the direction closest to the destination direction 71c-3 connecting the mobile robot 51 and the destination 71c-2. The movement route 51c-6 is calculated by the route calculation device 55 by adding 5 to the straight traveling speed component. Further, the route calculation device 55 calculates the speed reduced according to the distance from the mobile robot 51 to the obstacle 40. The movement speed calculated by the route calculation device 55 is input to the drive device 61, and the mobile robot 51 is driven.

  The turning speed component 51c-5 and the straight traveling speed component are obtained in the same manner as described above. However, the turning speed component is described in this specification, and the difference between the “current traveling direction” and the “angle closest to the destination direction excluding the movable and inaccessible area” is calculated. The magnitude obtained by multiplying by a proportional gain is the turning speed component.

  The linear velocity component is calculated as follows. First, the traveling speed is set according to the distance to the destination and the distance to the obstacle. Here, the traveling speed is the maximum rotational speed that the robot motor can continuously generate, and the maximum speed is defined as the distance at which deceleration starts when near the destination or near an obstacle. As Xd, for example, if the distance between the destination and the obstacle is x, if there is no destination or obstacle during the distance Xd, the speed of 100% is set as the traveling speed. If there is a destination or obstacle during the distance Xd, it is given by the following formula.

(Equation 5)
[Advance speed] = [Maximum speed] * (1-[Deceleration gain] * (Xd- x))

  In addition, although it is a linear velocity component, if the robot 51 tries to move at a large linear velocity when the turning component is large, the robot 51 may fall outside in the turning direction due to centrifugal force. .

(Equation 6)
[Linear speed component] = [Progression speed] * (1-[Turning deceleration gain] * | Turning speed component |)

  As described above, when the obstacle 40 is in the traveling direction of the mobile robot 51, the mobile robot 51 takes a movement path for avoiding the obstacle 40, and the mobile robot 51 passes through the obstacle 40 (in other words, As soon as there are no obstacles in the first and second detection areas, the route toward the destination 71c-2 is taken again (calculation as shown in FIG. 4B) toward the destination 71c-2.

  In order to move the mobile robot 51 along the movement route calculated by the route calculation device 55, the rotational speeds of the left and right drive wheels 59 of the left and right motors 61a of the drive device 61 are controlled as follows. Realize with. That is, the straight speed component may be given by the average speed of the two left and right drive wheels 59, and the turning speed component may be given by the speed difference between the two left and right drive wheels 59.

  Further, the basic processing flow of the mobile robot 51 having the above-described configuration is as shown in FIGS. 6A and 6B as follows.

  In the mobile robot 51 in which the setting of the second detection area 41 of the virtual sensor is not changed during the movement, processing is performed according to the basic flow as shown in FIG. 6A.

    Step S1: First, the moving destination of the robot 51 is input to the map database 52 by the input device 39. The destination on the map information 70 is updated by the input, and the subsequent steps until the destination is reached are executed. In the input of the moving destination, it is assumed that the destination coordinates and the reaching condition distance used for arrival determination are input.

    Step S2: Self-position information 73 of the robot 51 is acquired by the self-position measuring device 53.

    Step S3: The virtual sensor information calculation device 54 calculates virtual sensor calculation information based on the self-location information 73 and the map information 70 acquired in step S2.

    Step S4: The self-location information 73 of the robot 51 acquired in step S2 and the destination information of the map information 70 are compared by the route calculation device 55 to determine whether or not the destination has been reached. At this time, the distance from the self position (current position) of the mobile robot 51 to the destination is calculated by the route calculation device 55 from the coordinates of the self position (current position) and the coordinates of the destination, and the distance is calculated in step S1. If the route calculation device 55 determines that the distance is within the input arrival condition distance, it is assumed that the robot 51 has reached the destination, the destination information is cleared from the map information 70, and the movement of the robot 51 is driven. The process ends at 61 (step S7), and a new destination input (step S1) is waited for.

    Step S5: If the route calculation device 55 determines that the destination has not been reached, that is, the distance from the robot 51's own position to the destination is greater than the arrival condition distance, in steps S2 and 3 Based on the calculated information, the movement route of the robot 51 is calculated by the route calculation device 55.

    Step S6: The movement of the robot 51 is controlled by the driving device 61 so as to realize the movement of the robot 51 along the movement path calculated in Step S4. After performing Step S5, the process returns to Step S2.

  Further, in the mobile robot 51 in which the setting of the second detection area 41 of the virtual sensor is changed during the movement using the obstacle detection sensor 56 and the virtual sensor setting changing device 57, the flow shown in FIG. Process.

    Step S11: First, the moving destination of the robot 51 is input to the map database 52 by the input device 39. The destination on the map information 70 is updated by the input, and the subsequent steps until the destination is reached are executed. In the input of the moving destination, it is assumed that the destination coordinates and the reaching condition distance used for arrival determination are input.

Step S12: Various information is acquired by the obstacle detection sensor 56 and the self-position measuring device 53. Specifically, the following steps S12-1 and S12-2 are performed.
Step S12-1: Self-position information 73 of the robot 51 is acquired by the self-position measuring device 53.
Step S12-2: Obstacle detection information is acquired by the obstacle detection sensor 56.

Step S13: Based on the information acquired in step S12, the calculation conditions for the virtual sensor calculation information are set, and the virtual sensor calculation information is calculated by the virtual sensor information calculation device 54. Specifically, the following step S13-1, step S13-2, and step S13-3 are performed.
Step S13-1: Based on the self-location information 73, the obstacle detection information acquired in Step S12, and the map information 70 stored in the map database 52, if necessary, the virtual sensor setting changing device 57 performs a virtual sensor. Change the calculation condition setting for the calculation information.
Step S13-2: Based on the self-location information 73 acquired in Step S12, the virtual sensor calculation information is virtualized from the map information 70 stored in the map database 52 according to the calculation condition according to the virtual sensor setting change device 57. It is calculated by the sensor information calculation device 54.
Step S13-3: Based on the virtual sensor calculation information calculated in step S13-2, if necessary, the virtual sensor setting change device 57 has changed and changed the setting of the virtual sensor calculation information calculation condition. Based on the map information 70 stored in the map database 52 under the calculation conditions according to the virtual sensor setting change device 57, the virtual sensor information calculation device 54 calculates the virtual sensor calculation information again.

    Step S14: The self-location information 73 acquired in step S12 and the destination information of the map information 70 are compared by the route calculation device 55 to determine whether or not the destination has been reached. At this time, the distance from the self position (current position) of the mobile robot 51 to the destination is calculated by the route calculation device 55 from the coordinates of the self position (current position) and the coordinates of the destination, and the distance is calculated in step S11. If the route calculation device 55 determines that the distance is within the input arrival condition distance, it is assumed that the robot 51 has reached the destination, the destination information is cleared from the map information 70, and the movement of the robot 51 is driven. The process ends at 61 (step S17) and waits for a new destination input (step S1).

    Step S15: If the route calculation device 55 determines that the destination has not been reached, that is, the distance from the robot 51's own position to the destination is greater than the arrival condition distance, in steps S12 and S13. Based on the acquired information, the movement route of the robot 51 is calculated by the route calculation device 55.

    Step S16: The movement of the robot 51 is controlled by the driving device 61 so as to realize the movement of the robot 51 along the movement path calculated in Step S15. After performing Step S16, the process returns to Step S12 again.

  With the mobile robot 51 as described above, even in the basic flow in which the setting of the second detection area 41 of the virtual sensor is not changed during the movement, the actual obstacle detection sensor 56 has the physical characteristics of the sensor. It is possible to detect a wide range of obstacles that cannot be detected by the second detection area 41 of the virtual sensor. For example, a dead end path is detected in advance by the second detection area 41 of the virtual sensor. Therefore, it is possible to avoid inadvertent movement or deadlock by accidentally entering the dead end passage. Also, in the calculation amount in the route calculation, the calculation amount at the time of calculating the virtual sensor is proportional to (search calculation of the detection area of the second detection region 41 of the virtual sensor = detection area / accuracy). However, even when the entire moving range is set as the detection region, the calculation can be performed with a very small calculation amount as compared with the conventional example 2 that is proportional to the square of (detection area / accuracy). Therefore, it is possible to realize real-time and efficient movement to the destination in a mobile robot in an environment where an obstacle exists.

  In the present invention, since a virtual sensor can be set, characteristics (for example, the size and direction of a detection region) can be freely set without being particularly restricted by the physical detection characteristics of an actual sensor. For example, a search is performed on the back side of the obstacle or at a distant place, the shape of the obstacle is recognized based on the map information 70 of the map database 52 as described in the above example, and the surroundings are detected. Information that cannot be detected by this sensor can also be acquired. Furthermore, there is no need to consider the problems of detection accuracy and detection area, the number of sensors and mounting problems, the problem of interference between sensors, the problem of the influence of the surrounding environment, etc. that can occur when an actual sensor is mounted.

  Further, the obstacle detection sensor 56 detects an unknown obstacle that is not registered in the map database 52 and a moving obstacle by combining the obstacle detection sensor 56 that is an actual sensor with the virtual sensor. It is possible to deal with these unknown obstacles and moving obstacles. Then, the obstacle detected by the actual obstacle detection sensor 56 may be registered in the map information of the map database 52 by the map registration device 69 (see FIG. 1H). It is possible to calculate more accurate virtual sensor calculation information by updating.

  In addition, by installing the virtual sensor setting change device 57, it is possible to change the calculation settings according to the robot and surrounding conditions, so that the detection area can be set small and coarse in places where there are few obstacles. By increasing the detection area or increasing the accuracy only when necessary, high-accuracy detection can be realized while suppressing the overall calculation amount. It is possible to change not only the accuracy and the detection area, but also the characteristics as required. For example, it is possible to combine the functions of a plurality of sensors only by switching calculation settings.

  Here, the optimal virtual sensor setting method of the present invention is set to the minimum necessary detection area and accuracy in accordance with the movement characteristics of the robot and the characteristics of the obstacles as described in the above example. . The smaller the detection area and the coarser the detection accuracy, the smaller the amount of calculation and the smaller the burden on a processing device such as a computing device. Further, not only the detection area but also the detection characteristic may be set as appropriate as necessary.

  Here, the detection characteristic means “what can be extracted (detected)” as information by a virtual sensor, and includes, for example, the following.

(1) Information on the presence or absence of an obstacle in the second detection area of the virtual sensor. Information on the location and direction of the nearest obstacle.
This makes it possible to use the virtual sensor as an actual sensor.

(2) Information on whether passage is permitted or not.
Information for determining whether or not a robot can pass through a dead alley or a maze by a virtual sensor.
(The second detection area is sequentially extended in the direction in which the robot movement path continues to detect whether or not the exit of the passage is reached.)

(3) Information on the type of obstacle (for example, weight, material, etc.).
What is necessary is just to register into the map database 52 as a property of an obstacle like the position and shape of an obstacle.
If it is a light obstacle, you can choose to push it away by the robot.
In addition, if the obstacle is a brittle material, it can also be used as a judgment material for the robot to be careful.

  Further, when the obstacle detection sensor 56 and the virtual sensor are combined, it is possible to perform the role sharing by utilizing the advantages of the actual obstacle detection sensor 56 and the virtual sensor. For example, as shown in the above example, the actual obstacle detection sensor 56 includes a detection area around the robot 51 for final safety ensuring and a front of the robot 51 for avoiding an unknown obstacle in advance. Used for long range detection, virtual sensors are difficult to detect with actual sensors such as detecting obstacles on the movement path of the robot 51 according to the movement status of the robot 51 and collecting detailed information around the obstacles when detecting obstacles It is good to detect a region.

  Further, although it is an optimal moving route calculation method, it is preferable to use a route calculation method based on the information of the actual obstacle detection sensor 56 as it is. The calculation information itself of the virtual sensor itself has the same information content as the actual obstacle detection sensor 56 and there is no need to distinguish between them, and conversely, a virtual sensor is used instead of the actual obstacle detection sensor 56. When a sensor having a target specification is being developed and is available on the market, a virtual sensor (for example, a replaceable virtual device comprising the virtual sensor calculation device 54 and the conversion device 50 in FIG. This is because it is possible to easily replace the sensor 56z) with an actual sensor (see the arrow in FIG. 5).

  Using the characteristics, the calculated information of the virtual sensor is further converted into the same signal as the output signal when the obstacle detection sensor 56 actually detects the obstacle, for example, the same type of signal as shown in FIG. A conversion device 50 may be provided. In this case, since the output signal of the virtual sensor can be made the same as the output signal of the actual obstacle detection sensor 56 by the conversion device 50, for example, by changing the setting of the detection area in the virtual sensor, The effect of adding a sensor or changing its installation position can be verified without adding a sensor to the sensor or changing the actual sensor installation position. Conversely, a virtual sensor can be easily used instead of the actual obstacle detection sensor 56. Thereby, even in a mobile robot that is in an experimental stage or being adjusted, it is possible to verify the effect of adding a sensor or the like without actually adding a sensor or changing its installation position.

  Here, in the comparison with a virtual sensor, the following are mentioned as a typical example of an actual obstacle detection sensor.

(1) A sensor that determines the presence or absence of an obstacle in a region (eg, area sensor).
(2) A sensor that detects the distance to an obstacle (eg, ultrasonic sensor, laser sensor, etc.).
(3) A sensor that detects the presence or absence of obstacles and distance information in a certain angle range (eg, photoelectric sensor, laser scanner).

  The virtual sensor mentioned in this example in FIG. 5 functions as a variation of the above (3). For example, if a virtual sensor is used as a sensor that detects an angular region without an obstacle, it is synonymous with detection of an angular range in which the robot can move. As for an actual sensor, a sensor that not only physically detects a value but also calculates the value in a meaningful form within the sensor and outputs it can be considered as a sensor. For example, the sensor that performs the scanning of (3), a distance sensor that uses a stereo camera, and the like correspond to this.

  That is, the detection information obtained based on the information detected by the physical device of the sensor can be referred to as “detection information” by the sensor.

  On the other hand, “calculated information” of a virtual sensor is “information extracted from information stored in a map database”. The actual sensor is physically limited by the actual sensor, but if it is a virtual sensor, anything can be extracted as long as there is information in the map database. Conversely, the necessary data can be extracted from the database. The information that can be detected is not limited, and as a result, the “calculation information” of the virtual sensor includes the contents of the “detection information” of the actual sensor.

  As described above, the map information stored in the map database 52, the self-position information 73 measured by the self-position measurement device 53, and the virtual sensor information calculation device 54 are calculated by the virtual sensor setting change device 57. Example of changing calculation condition for calculating calculation information of virtual sensor based on calculated information of virtual sensor, detection information of obstacle sensor 56, and movement route calculated by route calculation device 55 More specific examples will be given below.

  First, a case where the calculation condition is changed by the virtual sensor setting changing device 57 based on the map information will be described. Assuming that FIG. 3A is the normal second detection area 41, as shown in FIG. 12A, when the robot 51 approaches the area III where there are many obstacles in the map information stored in the map database 52. (I state of the robot 51), the calculation condition is changed by the virtual sensor setting changing device 57 from the normal setting state of the second detection area 41s-1 (FIG. 12B), the second detection for the area with many obstacles. The state is changed to the setting state (FIG. 12C) of the area 41s-2. In the normal setting state of the second detection area 41 s-1, as shown in FIG. 12B, the obstacle is detected only in the area in front of the robot 51. On the other hand, the setting state of the second detection area 41s-2 for the area with many obstacles is not only the area in front of the robot 51 but also the entire area around the robot 51, as shown in FIG. 12C. In the direction, the current safe stop possible distance (the distance traveled until the speed stops) is set, and the obstacle is detected in the set area. On the other hand, when the robot 51 goes out from the region III where there are many obstacles in the map information (the state of the robot 51), the virtual sensor setting changing device 57 uses the second detection for the region where there are many obstacles. It changes so that it may return to the setting state (FIG. 12B) of the normal 2nd detection area | region 41s-1 from the setting state (FIG. 12C) of area | region 41s-2.

  Next, a case where the calculation condition is changed by the virtual sensor setting change device 57 based on the self-position information (for example, speed information) will be described. Assuming that FIG. 3A is the normal second detection area 41, when the moving speed of the robot 51 becomes equal to or higher than the threshold, the virtual sensor setting changing device 57 sets the calculation condition to the second detection area for slow speed. The state is changed from the setting state of 41s-3 (FIG. 13A) to the setting state (FIG. 13B) of the second detection region 41s-4 for high speed. The setting state of the second detection area 41s-3 for slow speed is the same as the setting state of the normal second detection area 41s-1 in FIG. 12B. On the other hand, the setting state of the second detection area 41s-4 for high speed is a distance (speed) at which the robot 51 can safely stop at the current speed in front of the robot 51, as shown in FIG. 13B. The distance traveled to stop is set, and obstacles are detected in the set area. On the other hand, when the moving speed of the robot 51 becomes less than the threshold, the calculation condition is changed by the virtual sensor setting changing device 57 from the setting state of the second detection area 41s-4 for the high speed (FIG. 13B). It changes so that it may return to the setting state (FIG. 13A) of the 2nd detection area | region 41s-3.

  Next, when the calculation condition is changed by the virtual sensor setting change device 57 based on the calculation information of the virtual sensor, the virtual sensor has detected an obstacle as described above (in the second detection area 41). When there is an obstacle), the virtual sensor setting changing device 57 changes the calculation condition from the normal setting state of the second detection area 41g (FIG. 3A) to the second after the obstacle detection including the additional area 41h. To the set state of the detection area 41 (FIG. 3B). On the other hand, when the virtual sensor no longer detects an obstacle (no obstacle in the second detection area 41), the setting state of the second detection area 41 for detecting the obstacle (including the additional area 41h) ( From FIG. 3B), it changes so that it may return to the setting state (FIG. 3A) of the normal 2nd detection area 41g.

  Next, in the case where the calculation conditions are changed by the virtual sensor setting changing device 57 based on the movement path, a movement stop command is issued to the robot 51 along the movement path of the robot 51 as shown in FIG. 3B. Due to the length of the distance traveled until the robot 51 stops (distance depending on the speed), when the robot 51 turns, the circle drawn by the robot 51 from the normal setting state of the second detection area 41g (FIG. 3A). The detection area is changed to include the width that the trajectory passes (more than twice the turning radius 42). When the turning operation of the robot 51 is completed, the detection state including the width (twice or more of the turning radius 42) through which the circular orbit drawn by the robot 51 passes is returned to the normal setting state of the second detection region 41g (FIG. 3A). Change as follows.

  Furthermore, a case where the calculation condition is changed by the virtual sensor setting change device 57 based on the detection information of the obstacle detection sensor 56 will be described. FIG. 14A shows a normal setting area, in which a first detection area 56s of the obstacle detection sensor 56 and a second detection area 41 of the virtual sensor are set. As shown in FIG. 14B, when the obstacle 40 is detected in the first detection area 56s of the obstacle detection sensor 56, the calculation conditions are calculated by the virtual sensor setting changing device 57, and FIG. As shown in FIG. 14C, not only the area in front of the robot 51 but also the surroundings of the robot 51, that is, all directions, the current safe stop distance (the distance traveled to the speed stop) is additionally set as a virtual sensor. Then, the obstacle is detected by the detection area 56t in which the additionally set area and the first detection area 56s are combined, and the second detection area 41. On the other hand, as shown in FIG. 14D, when the obstacle 40 is no longer detected in the detection region 56t in which the additionally set region and the first detection region 56s are combined, the calculation condition is set by the virtual sensor setting change device 57, As shown in FIG. 14E, a change is made to return to the normal setting area.

  In the present embodiment, an independently driven two-wheeled mobile robot with auxiliary wheels is assumed, but other moving mechanisms may be employed. For example, a moving mechanism that bends by a rudder such as an automobile may be used, a foot-shaped moving mechanism that does not use wheels, or a ship-type moving robot that moves on the sea. In that case, a virtual sensor setting that matches the characteristics of each moving mechanism or a method for calculating a moving route may be employed. Further, in the present embodiment, it is assumed that the movement is in a two-dimensional plane, but it may be adapted to movement in a three-dimensional space. Also in this case, the detection area of the virtual sensor can be easily adapted only by setting it three-dimensionally. Therefore, the technology of the present invention can also be applied to the movement of a flying object such as an airship or an airplane, or the tip of a manipulator.

  It should be noted that, by appropriately combining any of the various embodiments and more specific examples and more specific examples, the respective effects can be achieved.

  The mobile robot according to the present invention can realize real-time and efficient movement to a destination in a mobile robot in an environment where obstacles exist, so that it can be used in public places such as factories, stations, and airports. It can also be applied to robots that work independently and home robots.

  Although the present invention has been fully described in connection with preferred embodiments with reference to the accompanying drawings, various variations and modifications will be apparent to those skilled in the art. Such changes and modifications are to be understood as being included therein, so long as they do not depart from the scope of the present invention according to the appended claims.

FIG. 1A is a diagram showing a configuration of a mobile robot according to one embodiment of the present invention. FIG. 1B is a diagram showing an actual mobile robot and detection areas of its sensors. FIG. 1C is a diagram showing detection areas of the mobile robot and its virtual sensor on the map. FIG. 1D is a diagram illustrating an avoidance path calculated by the mobile robot illustrated in FIG. 1A. FIG. 1E is a diagram showing a configuration of a mobile robot different from the mobile robot shown in FIG. 1A. FIG. 1F is a perspective view, a perspective plan view, and a block diagram showing an outline of the mobile robot according to the embodiment of the present invention. FIG. 1G is a perspective view, a perspective plan view, and a block diagram showing an outline of the mobile robot according to the embodiment of the present invention. FIG. 1H is a perspective view, a perspective plan view, and a block diagram showing an outline of the mobile robot according to the embodiment of the present invention. FIG. 1I is a block diagram showing an outline of a mobile robot which is a modification of the embodiment of the present invention and does not have a virtual sensor setting change device. FIG. 2 is a diagram illustrating a detection area of the obstacle detection sensor and its detection value. FIG. 3A is a diagram illustrating a detection area of a virtual sensor (details will be described later) in a mobile robot according to one embodiment. FIG. 3B is a diagram illustrating a detection area of a virtual sensor (details will be described later) in a mobile robot according to another embodiment. FIG. 3C is a diagram illustrating calculation information of the virtual sensor when there is no obstacle. FIG. 3D is a diagram illustrating calculation information of the virtual sensor when there is an obstacle on the passage. FIG. 3E is a diagram illustrating calculation information of the virtual sensor when a general obstacle exists. FIG. 3F is a diagram illustrating calculation information of the virtual sensor when there is an obstacle that cannot be passed through. FIG. 4A is a diagram showing map information stored in a map database. FIG. 4B is a diagram illustrating a route calculation method when there is no obstacle. FIG. 4C is a diagram illustrating a route calculation method when there is an obstacle. FIG. 5 is a diagram illustrating the effect of the signal conversion apparatus. FIG. 6A is a diagram illustrating a processing flow of a basic mobile robot. FIG. 6B is a diagram illustrating a processing flow of the mobile robot when an obstacle sensor or a virtual sensor setting change device is used. FIG. 7A is a perspective view, a perspective plan view, and a block diagram showing an outline of a mobile robot of Conventional Example 1. 7B is a perspective view, a perspective plan view, and a block diagram showing an outline of the mobile robot of Conventional Example 1. FIG. FIG. 7C is a perspective view, a perspective plan view, and a block diagram showing an outline of the mobile robot of Conventional Example 1. FIG. 8A is a diagram showing a method for determining the movement route of the mobile robot shown in FIGS. 7A, 7B, and 7C. FIG. 8B is a diagram showing a method for determining the movement path of the mobile robot shown in FIGS. 7A, 7B, and 7C. FIG. 9 is a diagram illustrating a case where the mobile robot of Conventional Example 1 moves in an inefficient manner or falls into a deadlock state. FIG. 10A is a perspective view, a perspective plan view, and a block diagram showing an outline of a mobile robot of Conventional Example 2. FIG. 10B is a perspective view, a perspective plan view, and a block diagram showing an outline of a mobile robot of Conventional Example 2. FIG. 10C is a perspective view, a perspective plan view, and a block diagram showing an outline of a mobile robot of Conventional Example 2. FIG. 11 is a diagram illustrating a method of determining a movement route in the mobile robot illustrated in FIGS. 10A, 10B, and 10C. FIG. 12A is an explanatory diagram for describing an example in which the calculation condition is changed by the virtual sensor setting change device based on the map information. FIG. 12B is an explanatory diagram for explaining a setting state of a normal second detection region. FIG. 12C is an explanatory diagram for describing a setting state of a second detection area for an area with many obstacles. FIG. 13A is an explanatory diagram for explaining a setting state of a second detection area for a low speed. FIG. 13B is an explanatory diagram for describing a setting state of the second detection area 41s-4 for high speed. FIG. 14A is an explanatory diagram for explaining a state in which the first detection area of the obstacle detection sensor and the second detection area of the virtual sensor are set as a normal setting area. FIG. 14B is an explanatory diagram for explaining a state when an obstacle is detected in the first detection region of the obstacle detection sensor. In FIG. 14C, not only the area in front of the robot but also the surroundings of the robot, that is, all directions, the current safe stop distance (the distance to reach the speed stop) is additionally set as a virtual sensor, and the additional setting is performed. It is explanatory drawing for demonstrating the state which detects an obstruction with the detection area | region where the area | region which combined the 1st detection area | region, and the 2nd detection area | region. FIG. 14D is an explanatory diagram for explaining a state when an obstacle is no longer detected in the detection region where the additionally set region and the first detection region are combined. FIG. 14E is an explanatory diagram for explaining a state in which the normal setting area is restored.

Explanation of symbols

DESCRIPTION OF SYMBOLS 2,51a ... Main-body part, 3,56s ... 1st detection area | region, 4,56 ... Obstacle detection sensor, 5,53 ... Self-position measuring device, 6,55 ... Path | route calculation device, 7, 61 ... Drive device, 8, 71b-2 ... Destination, 9 ... Obstacle, 11, 52 ... Map database, 12 ... Memory area, 13 ... Map, 16 ... Obstacle, 16a ... Opening, 16b ... Dead end, 20, 20B, 51 ... Mobile robot, 21, 41g ... Second detection area, 22 ... Obstacle recognition device, 23 ... Obstacle, 24 ... Conversion device, 38 ... Output device, 39 ... Input device, 40a-6 ... Unaccessible area,
40d-4, 40e-4 ... Distance, 40d-5 ... Passage, 40d-6, 40d-7, 40e-6, 40f-6 ... Obstacle, 40e-3 ... Angular direction, 40f-7 ... Dead end , 40f-5 ... passage shape, 41s-1, 41s-2, 41s-3, 41s-4 ... second detection region, 40G ... obstacle, 40G-1 ... depth of depression, 41h ... additional region, 42 ... turning radius, 51b-4 ... traveling direction, 51b-5 ... turning speed component, 51b-6 ... movement path, 54 ... virtual sensor information calculation device, 56t, 64s, 65s ... detection region, 57 ... virtual sensor setting changing device 62 ... Encoder, 63 ... Odometry calculation device, 64 ... Photoelectric sensor, 65 ... Ultrasonic sensor, 70 ... Map information, 71b-3 ... Direction, 73 ... Self-position information.

Claims (3)

A movable robot body,
A self-position measuring device for measuring the self-position of the main body,
A map database for storing map information of the movement range of the main body,
An obstacle set on the map information based on the self-position information measured by the self-position measuring device and the map information stored in the map database and serving as an obstacle to the movement of the main body An obstacle information extraction unit that extracts the obstacle information in a detection area of a virtual sensor capable of detecting the object information;
A route calculation device that calculates a movement route for the main body to move based on the obstacle information extracted by the obstacle information extraction unit;
A virtual sensor setting change device for setting and changing an extraction condition for the obstacle information extraction unit to extract the obstacle information;
A mobile robot comprising:   In addition to the obstacle information extracted by the obstacle information extraction unit, the path calculation device further includes an obstacle detection sensor that detects an obstacle in a detection region around the main body, and the obstacle The mobile robot according to claim 1, wherein a movement path for the main body to move is calculated based on detection information from a detection sensor. The virtual sensor setting changing device includes the map information stored in the map database, the self-position information measured by the self-position measurement device, and the obstacle information extracted by the obstacle information extraction unit. The obstacle information extraction unit changes the extraction condition to extract the obstacle information based on at least one of detection information of the obstacle detection sensor and a movement route calculated by the route calculation device. Item 3. The mobile robot according to Item 1 or 2.

Images