JP6052045B2 - Autonomous mobile - Google Patents

Description

  The present invention relates to an autonomous mobile body having an environment map creation function, a route planning function, and an autonomous traveling function.

  Conventionally, an autonomous mobile body that autonomously moves in the surrounding environment is known. In order for an autonomous mobile body to move autonomously in the surrounding environment, an environment map that represents an area where an object (hereinafter referred to as an obstacle) in the moving space exists and an area where the object does not exist is required. Various methods have been devised for obtaining such an environmental map. Recently, SLAM (Simultaneous Localization and Mapping) is a technique for performing self-location estimation and environmental map creation in real time while moving. Is attracting attention. A mobile robot that uses SLAM to generate an environmental map using terrain data obtained as a result of distance measurement by a laser range finder (LRF) or a camera has been proposed (for example, Patent Document 1). .

JP 2010-92147 A

In the autonomous mobile body, for example, based on the control amount of the actuator (motor) that drives the wheel, the movement amount of the autonomous mobile body is cumulatively added to estimate the self position. In this case, the estimated self-position varies greatly depending on the wear of the wheels. Similarly, when wheel slips or sensor measurement errors are accumulated, there is a problem that a self-position estimated by the autonomous mobile body is shifted, and autonomous travel according to the travel schedule cannot be performed.
As described above, in the case of an autonomous mobile body that autonomously travels while estimating its own position on the environment map based on position information of obstacles located around the body of the autonomous mobile body, based on the control of the actuator Even if there is a deviation in the estimated self-position, the self-position on the environmental map can be corrected by map matching.
However, even in such an autonomous mobile body, there is a possibility that the self position cannot be accurately estimated in a space where there are almost no obstacles around the main body or a space where similar shapes are continuous.
In addition, there is an autonomous moving body that creates a travel schedule while creating an environment map based on an operator's manual operation and reproduces the travel schedule by autonomous travel. In such an autonomous mobile body, when the travel start position is set in a space where there are almost no obstacles around the main body as described above, or in a space where similar shapes are continuous, the travel start position is specified during reproduction travel. Cannot autonomously travel according to the travel schedule.
An object of the present invention is to enable an autonomous mobile body to reliably estimate its own position and appropriately travel according to a travel schedule.

Hereinafter, a plurality of modes will be described as means for solving the problems. These aspects can be arbitrarily combined as necessary.
An autonomous mobile body according to an aspect of the present invention includes a travel unit, a travel control unit, an obstacle information acquisition unit, a map creation unit, and a travel command generation unit.
The traveling unit includes a plurality of actuators to which control amounts are input. The travel control unit generates a control amount to be input to the actuator based on the travel command. The obstacle information acquisition unit acquires position information of obstacles existing around each predetermined control cycle. The map creation unit creates a local map indicating the presence or absence of an obstacle in the peripheral area of the main body based on the obstacle position information. The travel command generating unit outputs a forced stop travel command to the travel control unit when the local map satisfies a predetermined condition. Further, when the situation is approaching a predetermined condition, the operator can be notified of this by voice output or image output. For example, if the condition is to forcibly stop at a moving amount of 15 m from a predetermined position, a warning sound or voice notification and / or a warning lamp, a warning display, etc. may be sent when the moving amount from the predetermined position exceeds 10 m. it can.
When a predetermined condition that makes it difficult to estimate the position of the vehicle with the local map created by the map creation unit is satisfied, the travel command generation unit generates a forced stop travel command and travels by the travel unit. Stop.
In this way, in a place where it is difficult to estimate the self position, it is possible to prevent a situation in which the vehicle travels at a position different from the predetermined travel schedule due to wheel wear, slip, sensor measurement error, or the like.

Here, when a local map in which the ratio of obstacles in the surrounding area is equal to or less than a predetermined value continues for a predetermined travel distance or a predetermined travel time, the travel command generation unit determines that the local map has a predetermined condition. It may be determined that
In general, when there are almost no obstacles around the main body, the estimation of the self-position on the environment map depends only on the amount of movement of the traveling unit. Therefore, in such a case, there is a possibility that the autonomous mobile body is traveling at a position different from the predetermined travel schedule due to wheel wear and slip, sensor measurement error, and the like. In such a case, the autonomous mobile body according to the present invention can prevent traveling in a position different from the travel schedule by outputting a forced stop travel command based on the above determination.

In addition, when a local map in which an obstacle existing in the surrounding area has a simple shape with a small shape change continues for a predetermined travel distance or a predetermined travel time, the travel command generator generates a predetermined condition for the local map. You may judge that it was satisfied.
For example, when parallel wall surfaces continue along the travel route or when circular wall surfaces continue, local maps having the same shape are continued based on the obstacle information. Even in such a case, the autonomous mobile body according to the present invention can prevent traveling in a position different from the predetermined travel schedule by outputting a forced stop travel command based on the above determination. it can.

The autonomous mobile body may further include a storage unit that stores an environment map of the surrounding environment and a self-position estimation unit that estimates a self-position on the environment map. In that case, the travel command generation unit compares the environment map around the self-location estimated by the self-location estimation unit with the local map, and the state that differs by more than a predetermined ratio is a predetermined travel distance or a predetermined When the running time continues, it is determined that a predetermined condition is satisfied.
In this case, the local map created by the map creation unit is compared with the environmental map located in the vicinity of the self-location. be able to.

In the local map created by the map creation unit before the start of traveling, if the percentage of obstacles in the surrounding area is below a predetermined value, the obstacles in the surrounding area have a simple shape with little shape change. In this case, or when it differs from the environment map around the self-position by a predetermined ratio, the travel command generation unit may determine that the local map satisfies a predetermined condition.
In the case of the above conditions, the autonomous mobile body may travel at a position different from the predetermined travel schedule. The reason is as follows. First, when the position where there are almost no obstacles in the surrounding area or where there are many positions having obstacles of similar shape is set as the travel start position, the autonomous mobile body selects a travel start position different from the travel schedule. there's a possibility that. Secondly, when the shape of the obstacle on the local map is greatly different from the shape of the obstacle on the environmental map, there is a high possibility that it is different from the travel start position of the travel schedule.
In such a case, it is possible to prevent the traveling command generation unit from traveling in a position different from the predetermined traveling schedule by outputting a forced stopping traveling command.

The autonomous mobile body manually trains from a travel start position to a travel end position by an operator's manual operation while creating a travel schedule including a set of passing point data and passing point data at the passing time, and a traveling schedule. While reproducing, a reproduction travel mode in which the vehicle travels autonomously from the travel start position to the travel end position may be executed. Further, the travel command generation unit may determine whether or not the local map satisfies a predetermined condition in the teaching travel mode.
In this case, it is possible to create an appropriate travel schedule by eliminating passing point data that makes it difficult to estimate the self position in the teaching travel mode.

The travel command generation unit may determine whether or not the local map satisfies a predetermined condition in the reproduction travel mode.
In this case, in the reproduction travel mode, if the travel start position is greatly deviated from the travel schedule, a forced stop can be executed to further prompt the operator to reset the travel start position.

  According to the configuration of the present invention, it is possible to provide an autonomous mobile body capable of reliably performing self-position estimation and appropriately traveling according to a travel schedule.

The block diagram which shows schematic structure of the autonomous mobile body by one Embodiment. The control block diagram of the robot for cleaning. The control flowchart which shows the outline of control operation | movement of an autonomous mobile body. The control flowchart of the autonomous mobile body in teaching traveling mode. Explanatory drawing which shows an example of the scanning range of a laser range sensor. Explanatory drawing which shows the other example of the scanning range of a laser range sensor. The table which shows an example of the positional information on the obstruction detected by the laser range sensor. Explanatory drawing of the local map in the coordinate system centering on a sensor. Explanatory drawing of a driving | running route. The control flowchart of a self-position estimation process. Explanatory drawing which shows an example of posterior probability distribution. Explanatory drawing which shows an example of prior probability distribution. The control flowchart of likelihood calculation processing. Explanatory drawing which shows an example of likelihood distribution obtained by likelihood calculation processing. Explanatory drawing which shows an example of posterior probability distribution obtained by the posterior probability calculation process. Explanatory drawing which shows the conceptual structure of a global map update process. The control flowchart of the autonomous mobile body in reproduction travel mode. Explanatory drawing of the update process of a dynamic environment map. Explanatory drawing of likelihood calculation at the time of self-position estimation.

(1) Schematic Configuration Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In addition, in each figure, the same code | symbol is attached | subjected to the same element and the overlapping description is abbreviate | omitted.

FIG. 1 is a block diagram showing a schematic configuration of an autonomous mobile body according to an embodiment of the present invention.
The autonomous mobile body 1 includes a control unit 11, a distance measuring sensor 31, an operation unit 32, a display unit 33, and a traveling unit 25.

  The distance measuring sensor 31 is a sensor that detects an obstacle on the front side in the traveling direction of the autonomous mobile body 1. The distance measuring sensor 31, for example, irradiates a target with laser light pulsed by a laser oscillator, and receives reflected light reflected from the target with a laser receiver, thereby calculating a distance to the target. A range finder (LRF) can be used. The distance measuring sensor 31 can operate the laser beam in a fan shape at a predetermined angle using a rotating mirror. Such a distance measuring sensor 31 can also be arranged behind the autonomous mobile body 1.

The operation unit 32 is a user interface that is operated by an operator when the autonomous mobile body 1 travels by manual operation, and includes a throttle that receives a travel speed instruction, a handle that receives a travel direction instruction, and the like.
The display unit 33 displays information related to the current driving situation and various other information, and includes a liquid crystal display, an LED lamp, and the like.
The traveling unit 25 includes a plurality of traveling wheels (not shown) for traveling on the traveling route, and includes a plurality of traveling motors (not illustrated) as actuators for driving the traveling wheels.

The control unit 11 is a computer including a CPU, a RAM, and a ROM, and performs running control by executing a predetermined program.
The control unit 11 includes an obstacle information acquisition unit 21, a map creation unit 22, a travel command generation unit 23, and a travel control unit 24.

The obstacle information acquisition unit 21 acquires position information of obstacles existing around the main body 1a (see FIGS. 5 and 6) for each predetermined control cycle.
The map creation unit 22 creates a local map indicating the presence / absence of an obstacle in a peripheral region at a predetermined distance from the main body 1a based on the obstacle position information.
The travel control unit 24 generates a control amount to be input to the actuator based on the travel command.
The travel command generation unit 23 outputs a travel command for forced stop to the travel control unit 24 when the local map satisfies a predetermined condition.

(2) Functional Block A first embodiment in which the autonomous mobile body of the present invention is applied to a cleaning robot will be described below.
FIG. 2 is a control block diagram of the cleaning robot in which the first embodiment of the present invention is employed.
This cleaning robot 40 executes the teaching travel mode and the reproduction travel mode. In the teaching travel mode, a travel schedule consisting of a collection of passage point data and passage point data at the passage time and environmental map restoration data are created while manually traveling from the travel start position to the travel end position by an operator's manual operation. In the reproduction travel mode, the vehicle travels autonomously from the travel start position to the travel end position while reproducing the travel schedule.
The cleaning robot 40 includes the autonomous mobile body 1, a throttle 42, and a cleaning unit user interface 43. The cleaning robot 40 further includes a cleaning unit (not shown). The cleaning unit is provided at the lower part of the autonomous mobile body 1 and is a device for cleaning the floor surface along with the movement of the autonomous mobile body 1.

  The throttle 42 corresponds to the operation unit 32 of FIG. 1 and accepts an instruction input by an operator's manual operation. The throttle 42 can be independently provided with a left and right throttle grip for receiving a control instruction input according to a rotation angle around the axis. Further, the throttle 42 may be a combination of a throttle grip that receives the forward speed and a handle that receives an instruction of the steering direction. Further, the throttle 42 can be a combination of a throttle lever and other input means.

  The cleaning unit user interface 43 receives an instruction for a cleaning operation by a cleaning unit (not shown) from an operator, and can be configured by a combination of operation buttons, a touch panel, and other operation switches, for example.

  The autonomous mobile body 1 includes a control board 51, a breaker 52, a front laser range sensor 53, a rear laser range sensor 54, a bumper switch 55, an emergency stop switch 56, a speaker 57, and a traveling unit 25.

The breaker 52 is a main power switch, and turns on / off the supply of power supply voltage from a battery (not shown) to each part of the autonomous mobile body 1 based on the operation of the operator.
The front laser range sensor 53 detects position information of an obstacle positioned in front of the autonomous mobile body 1, and reflects the laser beam irradiated in the horizontal direction within a predetermined angle range and reflected from the obstacle. The light is received and the distance from the main body 1a to the obstacle is measured.
The rear laser range sensor 54 detects positional information of an obstacle located behind the autonomous mobile body 1, and reflects the laser beam irradiated from the obstacle in the horizontal direction within a predetermined angle range. The light is received and the distance from the main body 1a to the obstacle is measured.

The bumper switch 55 can be a pressure-sensitive switch attached to the outer periphery of the vehicle body of the autonomous mobile body 1, detects contact with an obstacle, and outputs a detection signal.
The emergency stop switch 56 receives an instruction to urgently stop the operation of the autonomous mobile body 1, and is a switch that can be operated by an operator.
The speaker 57 notifies the operator of various information during operation of the autonomous mobile body 1 by sound.

  The control board 51 is a circuit board on which a CPU, a ROM, and a RAM are mounted, and includes a control unit 11, a teaching data creation unit 26, a SLAM processing unit 63, a nonvolatile memory 64, an obstacle information acquisition unit 21, a storage unit 27, and a travel control. Part 24 is included.

  The storage unit 27 stores various data, and stores the travel schedule and environment map restoration data created in the teaching travel mode in association with the time or an identifier associated with the time.

The control unit 11 is a microprocessor mounted on the control board 51, and controls the operation of each unit of the autonomous mobile body 1 by executing a predetermined program.
The teaching data creation unit 26 creates a travel schedule that is a set of passage time data corresponding to the passage time and the passage time in the teaching travel mode.

  The SLAM processing unit 63 is a functional block that executes SLAM (Simultaneous Localization and Mapping) processing that simultaneously performs self-position estimation and environment map creation, and includes a map creation unit 22, a self-position estimation unit 72, and a map matching unit 73. .

  The map creation unit 22 stores the obstacle position information acquired by the obstacle information acquisition unit 21 in the storage unit 27 as environmental map restoration data in association with the acquired time. Further, the map creation unit 22 creates a local map (local map) in the peripheral area of the main body 1a based on the obstacle position information acquired by the obstacle information acquisition unit 21. Furthermore, the map creation unit 22 creates a global map (environment map) for self-position estimation based on the created local map. Further, in the reproduction travel mode, the map creation unit 22 reads the environmental map restoration data corresponding to the time before the current passing point from the storage unit, and updates the global map.

The self-position estimation unit 72 estimates the current self-position by adding the movement amount from the previous self-position to the previous self-position.
The map matching unit 73 compares the local map created based on the position information of the obstacle located around the main body 1a with the updated global map, and the self-position estimated by the self-position estimation unit 72 Correct.

The nonvolatile memory 64 stores a boot program for the control unit 11, a program related to traveling control, various parameters, and the like.
The obstacle information acquisition unit 21 acquires position information of obstacles located around the main body 1a based on detection signals of the front laser range sensor 53 and the rear laser range sensor 54.
The storage unit 27 stores the travel schedule created by the teaching data creation unit 26 and the environment map restoration data created by the map creation unit 22 or an identifier associated with time.

The travel command generation unit 23 outputs a travel command to the travel control unit 24. The travel command generation unit 23 outputs a travel command to the travel control unit 24 based on an instruction input from an operator input via the throttle 42 in the teaching travel mode. The travel command generation unit 23 generates a travel command based on the self-position on the global map estimated by the SLAM processing unit 63, the updated global map, and the travel schedule in the reproduction travel mode, and the travel control unit 24 Output to.
Furthermore, the travel command generation unit 23 determines whether or not the position information of the obstacle on the local map created by the map creation unit 22 satisfies a predetermined condition, and determines that the predetermined condition is satisfied. Then, a forcible stop travel command is output to the travel control unit 24.

  The travel control unit (motion controller) 24 generates a control amount of the actuator of the travel unit 25 based on the input travel command, and inputs it to the travel unit 25.

  The travel control unit 24 includes a calculation unit 81, a determination unit 82, and a movement control unit 83. The determination unit 82 interprets the input travel command. The computing unit 81 calculates the control amount of the actuator of the traveling unit 25 based on the traveling command interpreted by the determining unit 82. The movement control unit 83 outputs the control amount calculated by the calculation unit 81 to the traveling unit 25.

The traveling unit 25 includes a pair of motors 93 and 94 corresponding to two traveling wheels (not shown), and includes encoders 95 and 96 that detect the respective rotational positions, and motor drive units 91 and 92. ing.
The motor drive unit (motor driver) 91 performs feedback control of the motor 93 based on the control amount input from the travel control unit 24 and the rotational position of the motor 93 detected by the encoder 95. Similarly, the motor drive unit 92 performs feedback control of the motor 94 based on the control amount input from the travel control unit 24 and the rotational position of the motor 94 detected by the encoder 96.

  The travel command generation unit 23, the teaching data creation unit 26, the obstacle information acquisition unit 21, the travel control unit 24, and the SLAM processing unit 63 may be functional blocks that are realized by the control unit 11 executing programs. It is also possible to configure each with an independent integrated circuit.

(3) Control Operation FIG. 3 is a control flowchart showing an outline of the control operation of the autonomous mobile body. Hereinafter, when the control unit 11 in FIG. 2 executes a predetermined program, each function of the travel command generation unit 23, the teaching data creation unit 26, the SLAM processing unit 63, the obstacle information acquisition unit 21, and the travel control unit 24 is performed. A description will be given assuming that the block is realized.

  In step S31, the control unit 11 determines whether or not a mode selection has been performed by the operator. Specifically, when an instruction input is received by an operation of the operation unit 32 by the operator, or when an instruction input signal is received from the remote controller, the control unit 11 proceeds to step S32.

In step S32, the control unit 11 determines whether or not the selected mode is the teaching travel mode. When it is determined that the instruction input for selecting the teaching travel mode has been received, the control unit 11 proceeds to step S33, and otherwise proceeds to step S34.
In step S33, the control unit 11 executes the teaching travel mode, and then proceeds to step S36.
In step S34, the control unit 11 determines whether or not the selected mode is the reproduction travel mode. If the control unit 11 determines that an instruction input for selecting the reproduction travel mode has been received, the control unit 11 proceeds to step S35.

In step S35, the control unit 11 executes the reproduction travel mode, and then proceeds to step S36.
In step S <b> 36, the control unit 11 determines whether an end instruction has been accepted or whether the taught travel schedule has been completed. Specifically, the control unit 11 receives an instruction input indicating that the process is ended by an operation of the operation unit 32 by the operator, receives an instruction input signal indicating that the process is ended by the remote controller, or teach travel If it is determined that the travel schedule created by the mode has been completed, the process is terminated; otherwise, the process proceeds to step S31.

(3-1) Teaching Travel Mode FIG. 4 is a control flowchart of the autonomous mobile body in the teaching travel mode.
In step S40, the control unit 11 performs initial setting. Specifically, the control unit 11 acquires position information of an obstacle located around the main body 1a based on the detection signals of the front laser range sensor 53 and the rear laser range sensor 54, and based on this, the control unit 11 Create a local map around.
When the global map is not stored in the storage unit 27 in advance, the control unit 11 determines that the self position (x1, y1) at time (t1) created based on the position information of the obstacle is the origin of the absolute coordinate system ( Alternatively, the local map is replaced with a global map so that the predetermined coordinates are obtained.

FIG. 5 is an explanatory diagram illustrating an example of a scanning range of the laser range sensor.
The autonomous mobile body 1 has a front laser range sensor 53 attached to the front end thereof, scans the pulsed laser beam in a fan shape with a rotating mirror, and reflects it from an obstacle positioned in front of the autonomous mobile body 1. Receive reflected light.
In the example shown in FIG. 5, the front laser range sensor 53 transmits laser light to the scanning region 101 of about 180 degrees forward. The front laser range sensor 53 receives the reflected light reflected from the obstacle at a predetermined period, compares the transmitted pulse of the transmitted laser light with the received pulse of the received reflected light, and detects the distance to the obstacle. .

A rear laser range sensor 54 is attached to the autonomous mobile body 1 at the rear end. Similar to the front laser range sensor 53, the rear laser range sensor 54 scans the pulsed laser beam in a fan shape with a rotating mirror, and as a result, reflects from an obstacle located behind the autonomous mobile body 1. Receive light.
In the example shown in FIG. 5, the rear laser range sensor 54 transmits laser light to the scanning region 103 that is about 180 degrees rearward. However, the autonomous mobile body 1 is provided with a mask area 105 that does not detect obstacles in the teaching travel mode described later, and detects position information of obstacles located in the first scanning area 115 and the second scanning area 117. . Note that the mask region 150 is provided corresponding to an operator located behind the autonomous mobile body 1.

FIG. 6 is an explanatory diagram showing another example of the scanning range of the laser range sensor.
In the example shown in FIG. 5, the front laser range sensor 53 and the rear laser range sensor 54 scan the laser beam at an angle of about 180 degrees, whereas in the example shown in FIG. The difference is that the laser beam is scanned at an angle of.
Specifically, the scanning region 101 by the front laser range sensor 53 is an angle range of 240 degrees forward. Further, the scanning area 103 by the rear laser range sensor 54 has an angle range of 240 degrees behind, and detects position information of obstacles located in the first scanning area 115 and the second scanning area 117 excluding the mask area 105.

FIG. 7 is a table showing an example of position information of an obstacle detected by the laser range sensor.
The front laser range sensor 53 and the rear laser range sensor 54 irradiate a predetermined scanning region with laser light, further receive reflected light from an obstacle, and then irradiate the transmitted pulse of laser light and the received reflected light. The distance R to the obstacle is calculated as an output value based on the received pulse. The front laser range sensor 53 and the rear laser range sensor 54 also output the sensor angle N when the distance R to the obstacle is obtained.
The control unit 11 acquires the outputs of the front laser range sensor 53 and the rear laser range sensor 54 for each predetermined angle, and based on this, a position information table in which the sensor angle N and the distance R to the obstacle are associated with each other. create. The position information table shown in FIG. 7 is an obstacle position information table obtained by obtaining sensor outputs every 0.72 degrees.

  The control unit 11 executes a local map creation process using the sensor coordinate system. Specifically, the control unit 11 detects the sensor based on the obstacle position information table (for example, the obstacle position information table shown in FIG. 7) acquired by the front laser range sensor 53 and the rear laser range sensor 54. Create a local map, which is an environmental map in the vicinity of its own position in the central coordinate system.

  For example, as shown in FIGS. 5 and 6, the control unit 11 divides the scanning areas 101 and 103 scanned by the front laser range sensor 53 and the rear laser range sensor 54 into a grid having a predetermined area, Based on the object position information table, the obstacle existence probability in each grid is calculated.

FIG. 8 is an explanatory diagram of a local map in a coordinate system centered on a sensor.
The control unit 11 uses the sensor center of the front laser range sensor 53 or the rear laser range sensor 54 as the origin 132, and the obstacle existence probability with respect to the grid on the scanning line 131 corresponding to the sensor angle N of the obstacle position information table. To decide. Here, the obstacle existence probability of the grid where the obstacle exists is set to “1.0”, the obstacle existence probability of the grid where it is unknown whether the obstacle exists is set to “0”, the obstacle, the origin, The obstacle existence probability of the grid located between “−1+ (r / R) ² = −1.0 to 0” is assumed.

The control unit 11 sets a grid on the scanning line 131 where the distance r from the origin 132 coincides with the distance R to the obstacle as the detection grid 133, and sets the obstacle existence probability of the grid to “1.0”. To do.
The control unit 11 determines that there is no obstacle on the grid on the scanning line 131 and whose distance r from the origin 132 is less than the distance R to the obstacle. In FIG. 8, the obstacle existence probability of the intermediate grids 134 to 138 located between the origin 132 and the detection grid 133 is set to “−1+ (r / R) ² = −1.0 to 0”.
The control unit 11 determines that it is unknown whether or not there is an obstacle in the grid on the scanning line 131 where the distance r from the origin 132 is further than the distance R to the obstacle. Thus, the obstacle existence probability of the unknown grid 139 is set to “0”.

Accordingly, as shown in FIG. 5 or FIG. 6, the detection grid 107 having an obstacle existence probability of “1.0” around the front laser range sensor 53 and the rear laser range sensor 54, and the obstacle existence probability A local map based on the sensor coordinate system is created, including the intermediate grid 109 with “−1.0 to 0” and the unknown grid 111 with the obstacle existence probability “0”.
However, the grid located in the mask area 105 of the rear laser range sensor 54 is set to “0” as the obstruction existence probability as the unknown grid 113.
In the example shown in FIGS. 5 and 6, it is assumed that the detection grid 107 exists linearly in front of the autonomous mobile body 1, and therefore there is a wall surface or a large obstacle that constitutes the travel route. Can do.
Thus, in the local map creation process of the sensor coordinate system, the obstacle existence probability of each grid located around the front laser range sensor 53 and the rear laser range sensor 54 is calculated, and this is calculated as the sensor coordinates. Output as a local map of the system.

  In step S41, the control unit 11 determines whether or not the travel start position is appropriate. Specifically, the local map created based on the obstacle position information acquired by the obstacle information acquisition unit 21 has a position where there are almost no obstacles around the main body 1a, or a simple shape is continuous. If it is a position, the control unit 11 determines that the local map satisfies a predetermined condition.

FIG. 9 is an explanatory diagram of a travel route.
As shown in FIG. 9, the autonomous mobile body 1 travels in an annular travel route 901, and arbitrary passing points in the travel schedule are defined as a first passing point 911 to a fifth passing point 915.
At each of the first passing points 911 to the fifth passing points 915, the map creating unit 22 creates local maps 921 to 925 of the surrounding area based on the position information of the obstacles.

  When the third passing point 913 in FIG. 9 is selected as the travel start position in the teaching travel mode, the local map 923 in the peripheral area of the main body 1a does not include obstacle information. Therefore, the obstacle existence probability of each grid of the local map 923 is “−1.0 to 0”.

In FIG. 9, there are a plurality of passing points where a local map having an obstacle existence probability “−1.0 to 0” of each grid is created in the vicinity of the third passing point 913 of the travel route 901. Accordingly, when such a third passing point 913 is selected as the travel start position, it is impossible to determine whether or not the travel start position matches the travel start position of the travel schedule in the reproduction travel mode.
Further, in the reproduction travel mode, even if the travel start position can be coincident with the third passing point 913 by accident, the travel direction (posture θ) cannot be specified. Difficult to do.

When the second passing point 912 in FIG. 9 is selected as the travel start position in the teaching travel mode, the local map 922 in the peripheral area of the main body 1a has an obstacle presence probability of the grid at positions corresponding to the parallel wall surfaces on both the left and right sides. Is “1.0”, and the obstacle existence probability of the grid located between them is “−1.0 to 0”.
In the example shown in FIG. 9, a simple shape in which the parallel wall surfaces are present on the left and right sides is continuous before and after the second passing point 912 in the traveling direction, and similar in the vicinity of the second passing point 912. There are multiple passing points that are local maps. Further, there is a point in the travel route 901 having the same shape as the local map even at a position away from the second passing point 912. For example, the local map 924 at the fourth passing point 914, The local map 925 at the fifth passing point is a similar local map.
Therefore, when such a second passing point 912 is selected as the travel start position, it becomes impossible to determine whether or not the travel start position matches the travel start position of the travel schedule in the reproduction travel mode.

  In this way, when the ratio of obstacles present in the local map created by the map creator 22 is very low before the start of traveling, or when the percentage of obstacles is below a predetermined value, The control unit 11 determines that the local map satisfies the predetermined condition (No in Step S41), and proceeds to Step S42. If not (Yes in Step S41), the control unit 11 proceeds to Step S43.

  In step S <b> 42, the control unit 11 forcibly stops traveling in the traveling unit 25. Specifically, the travel command generation unit 23 generates a forced stop travel command and outputs it to the travel control unit 24. The travel control unit 24 transmits a control signal for stopping the motors 93 and 94 of the travel unit 25 based on the forced stop travel command. Thereafter, the control unit 11 ends the process.

  In step S <b> 43, the control unit 11 creates a control amount for the actuator of the traveling unit 25 based on the traveling command input by the operator and outputs the control amount to the traveling unit 25. In the teaching travel mode, a travel speed input by an operator operating the throttle 42 and an instruction input related to steering are received, and travel control in the travel route is performed. The instruction input by the operator can also receive an instruction input signal by radio from the remote controller.

In step S44, the control unit 11 executes a data measurement process. Specifically, the control unit 11 irradiates the laser beam with the front laser range sensor 53 and the rear laser range sensor 54, receives the reflected light reflected from the obstacle, and has an obstacle related to the distance and direction of the obstacle. Get material information.
Moreover, the control part 11 acquires the information regarding the movement amount in the fixed time of the autonomous mobile body 1 further. Specifically, the control unit 11 acquires information on the rotational positions of the corresponding motors 93 and 94 from the encoders 95 and 96 of the traveling unit 25, and calculates the movement amount within a predetermined time.
The amount of movement within a certain period of time is the amount of movement from the two-dimensional coordinates (x (t−1), y (t−1)) at the previous measurement time (t−1) of the autonomous mobile body 1 and the traveling direction. It can be expressed as a movement amount (dx, dy, dθ) having a change amount.

  In step S45, the control unit 11 executes a local map creation process using the sensor coordinate system. Specifically, based on the obstacle position information table acquired by the front laser range sensor 53 and the rear laser range sensor 54 (for example, the obstacle position information table shown in FIG. 7), a coordinate system centered on the sensor is used. A predetermined range located around the main body 1a is divided into grids of a predetermined size, and a local map in which the calculated obstacle existence probability is associated with each grid is created.

  In step S46, the control unit 11 determines whether or not the created local map satisfies a predetermined condition. Specifically, when the ratio of obstacles present in the local map created by the map creation unit 22 is less than or equal to a predetermined value, or the obstacles included in the local map are simple shapes Therefore, the control unit 11 determines that the local map satisfies the predetermined condition, and proceeds to step S47. In addition, when there is a global map stored in advance in the storage unit 27, the global map located in the vicinity of the self position is compared with the created local map, and the obstacle existence probability in each corresponding grid is large. Even if they are different, the control unit 11 determines that the local map satisfies the predetermined condition, and proceeds to step S47.

In step S47, the control unit 11 determines whether or not the state where the created local map satisfies a predetermined condition continues for a predetermined travel distance or continues for a predetermined time. Specifically, for example, when the vehicle is passing in the vicinity of the third passing point 913 in FIG. 9, the state in which the ratio of obstacles in the local map is equal to or less than a predetermined value continuously for a predetermined distance or a predetermined time. Arise. Therefore, in such a case, the control unit 11 determines that an accurate self-position on the global map cannot be specified (Yes in Step S47), and proceeds to Step S48.
Further, for example, when the vehicle is passing near the second passing point in FIG. 9, a state where the obstacle existing in the local map has a simple shape continues for a predetermined distance or a predetermined time. Even in such a case, the control unit 11 determines that an accurate self-position on the global map cannot be specified (Yes in Step S47), and proceeds to Step S48.
Further, the control unit 11 compares the global map stored in the storage unit 27 with the local map created by the map creation unit 22 and determines that a greatly different state continues for a predetermined distance or a predetermined time. In such a case (Yes in step S47), the process proceeds to step S48. For example, if the layout in the travel route has changed significantly from the time when the global map is created, the control unit 11 determines that an accurate self-position on the global map cannot be specified, and proceeds to step S48. .
In step S47, when the control unit 11 determines that the state in which the local map satisfies the predetermined condition is not continuous for a predetermined distance or for a predetermined time (No in step S47), the control unit 11 proceeds to step S49.

In step S <b> 48, the control unit 11 forcibly stops traveling in the traveling unit 25. Specifically, the control unit 11 causes the travel command generation unit 23 to generate a forced stop travel command and causes the travel control unit 24 to output the forced stop travel command.
Thereafter, the control unit 11 ends the process.

  In step S49, the control unit 11 estimates the self position on the global map. Specifically, the self-position estimating unit 72 adds the amount of movement (dx, dy, dθ) between time (t−1) and time (t) to the self-position at time (t−1) to obtain the time ( Estimated as the self-position of t). Further, the control unit 11 assumes a three-dimensional grid space (x, y, θ) having two-dimensional coordinates (x, y) and a traveling direction (θ) as parameters of the self-position, and each grid before and after the movement. Map matching is performed by calculating the probability distribution of the self-position.

FIG. 10 is a control flowchart of the self-position estimation process.
In step S101, the control unit 11 executes a prior probability calculation process.
Using the time (t) that is the current time, the control unit 11 uses the posterior probability distribution at time (t−1) and the movement amount (dx, dy, dθ between time (t−1) and time (t). ) To calculate the prior probability distribution of the self-position at time (t).

FIG. 11 is an explanatory diagram illustrating an example of the posterior probability distribution, and FIG. 12 is an explanatory diagram illustrating an example of the prior probability distribution.
The control unit 11 adds the amount of movement (dx, dy, dθ) between time (t−1) and time (t) to the posterior probability distribution at time (t−1) shown in FIG. Shift.
The control unit 11 further obtains a prior probability distribution at time (t) as shown in FIG. 12 by performing a convolution operation on the normal distribution.
The control unit 11 executes such a prior probability calculation process for each grid in the three-dimensional grid space (x, y, θ), thereby taking into account wheel slips and the like, while taking into account wheel slips and the like. The probability distribution of the self position is updated by adding the movement amounts (dx, dy, dθ) acquired from (1).

In step S102, the control unit 11 executes a likelihood calculation process.
FIG. 13 is a control flowchart of likelihood calculation processing.
In step S131, the control unit 11 extracts temporary self-position coordinates. Specifically, the control unit 11 extracts one coordinate that exceeds the threshold from the prior probabilities at the time (t) calculated by the prior probability distribution process, and uses this as a temporary self-position coordinate. The control unit 11 calculates the likelihood in the extracted temporary self-position coordinates based on the prior probability distribution at time (t), the local map, and the global map being created.

In step S132, the control unit 11 performs a coordinate conversion process. Specifically, the control unit 11 converts a local map based on the sensor coordinate system into a local map based on the temporary self-position coordinates.
The control unit 11 uses the temporary self-position coordinates extracted based on the prior probability distribution as (x, y, θ), the coordinates of each grid of the local map by the sensor coordinate system as (lx, ly), and absolute coordinates. The coordinates (gx, gy) of each grid of the local map of the system are calculated by Equation 1.

In step S133, the control unit 11 executes map matching processing.
The control unit 11 performs map matching processing between the global map being created and the local map of the absolute coordinate system, and calculates the obstacle existence probability of each grid on the local map of the absolute coordinate system based on the map matching process.
The control unit 11 sets the coordinates of the i-th grid to (gx_occupy (i), gy_occupy (i)) when N grids having an obstacle existence probability “1.0” exist in the local map of the absolute coordinate system. The score s indicating the degree of coincidence is calculated by the following formula 2 with the obstacle existence probability at the coordinates (gx_occupy (i), gy_occupy (i)) on the global map as GMAP (gx_occupy (i), gy_occupy (i)) To do.



The control unit 11 calculates a score s for each of a plurality of temporary self-positions, and sets it as the likelihood distribution of the self-position.
FIG. 14 is an explanatory diagram illustrating an example of a likelihood distribution obtained by the likelihood calculation process.

  In step S103, the control unit 11 executes a posteriori probability calculation process. Specifically, the control unit 11 multiplies the prior probability distribution at the time (t) by the likelihood distribution at the time (t) to each grid of the local map in the absolute coordinate system, thereby obtaining the time map at the time (t). Calculate the posterior probability.

FIG. 15 is an explanatory diagram showing an example of the posterior probability distribution obtained by the posterior probability calculation process. FIG. 15 shows a posterior probability distribution obtained by multiplying the prior probability distribution at time (t) shown in FIG. 11 by the likelihood distribution at time (t) shown in FIG.
The control unit 11 can execute the posterior probability calculation process for a plurality of temporary self-positions, and can set the one having the maximum posterior probability as the self-position.

  With the above-described configuration, the difference in prior probability can be reflected in the posterior probability with respect to the temporary self-position having the same likelihood value, and further, the front laser range sensor 53, the rear laser Based on the output signals of the range sensor 54 and the encoders 95 and 96 of the traveling unit 25, the self-position can be estimated by dead-reckoning. Further, since the a posteriori probability at the time (t-1) is used when the prior probability calculation at the time (t) is performed, slip and wear of wheels (not shown) in the traveling unit 25, the front laser range sensor 53, disturbance factors such as noise of the rear laser range sensor 54 can be prevented, that is, the self-position can be accurately estimated.

  In step S <b> 50, the control unit 11 uses the obstacle information acquisition unit 21 to acquire position information of obstacles located around the current position on the global map.

In step S <b> 51, the control unit 11 creates environmental map restoration data by associating the position information of the obstacle acquired by the obstacle information acquisition unit 21 with the time, and stores the data in the storage unit 27.
Here, the environmental map restoration data described in the storage unit 27 divides the periphery of the self-location on the global map into unit grids of a predetermined area, and further associates obstacle existence probabilities with the respective grids. It is possible to associate the position information with the time (t) at which the position information is acquired.
When a global map is stored in the storage unit 27 in advance, the step S51 can be omitted.

  In step S52, the control unit 11 performs a global map update process. Specifically, the control unit 11 adds the local map of the absolute coordinate system obtained by performing coordinate transformation with the self-position estimated by the self-position estimation process to the global map at time (t-1). A global map at time (t) is created.

FIG. 16 is an explanatory diagram showing a conceptual configuration of the global map update process. In the example shown in FIG. 16, a 4 × 4 grid in the vicinity of the self-position is targeted, (a) is a global map at time (t−1), and (b) is an absolute coordinate system local at time (t). The map, (c), shows the global map at the updated time (t).
As shown in FIG. 16, the global map at time (t−1) and the local map at time (t) have obstacle existence probabilities stored in the respective grids. As described above, the obstacle existence probability of a grid that is estimated to have an obstacle is “1.0”, and the obstacle existence probability of a grid that is estimated to be no obstacle is “−1”. It is assumed that the obstacle existence probability of the grid where the existence of the obstacle is unknown is set to “0”.

In the global map update process, the control unit 11 converts each grid (FIG. 16B) of the local map converted into the absolute coordinate system based on the self position estimated by the self position estimation process to each grid of the global map. Corresponding to (FIG. 16A), the obstacle existence probability of each grid is added.
When the obstacle existence probability of each grid becomes “1.0” or more and “−1.0” or less by the calculation, the control unit 11 performs a truncation calculation to calculate “1.0”. And “−1.0”. As a result, the obstacle existence probability of the grid determined to be repeatedly present is close to “1.0”, and the obstacle existence probability of the grid determined to be the non-repetitive obstacle is “−1”. .0 ".
In this way, the control unit 11 updates the global map by calculating the obstacle existence probability of each grid of the global map at time (t) shown in FIG.

The global map at time (t) includes a range necessary for the control unit 11 to execute the self-position estimation process at least at the next time (t + 1), and is temporarily stored in the storage unit 27. .
In this global map update process, the obstacle existence probability of each grid of the global map is updated by adding the obstacle existence probability of each corresponding grid of the local map of the absolute coordinate system, so that the wheel slip or sensor It is possible to minimize the influence of noise based on the measurement error.

  In the teaching travel mode, the control unit 11 can connect the local maps in the travel route, create an overall global map, and store the global map in the storage unit 27. In that case, in order to solve the circular route problem, it is preferable to divide the map into partial maps and to select and set connection points of the partial maps. In step S51, when the environmental map restoration data corresponding to the passage time is created and stored in the storage unit 27, the global map in the vicinity of the passing point is left in the storage unit 27, and the information of the global map after passing is stored. It is also possible to delete. In this case, it is not necessary to consider the problem that the circular route does not close in the environment map.

  In step S <b> 53, the control unit 11 creates a travel schedule by the teaching data creation unit 26 and stores it in the storage unit 27. Specifically, the passing point data (x, y, y) including the coordinates (x, y) on the global map of the autonomous mobile body 1, the posture (θ) indicating the traveling direction of the autonomous mobile body 1, and the time (t). θ, t) is stored in the storage unit 27.

  In step S54, the control unit 11 determines whether an end instruction has been accepted. Specifically, the control unit 11 performs an end process when an instruction input to end the process is received by an operation of the operation unit 32 by the operator or when an instruction input signal to end the process is received from the remote controller. Is terminated, otherwise the process proceeds to step S43.

(3-2) Reproduction Travel Mode FIG. 17 is a control flowchart of the autonomous mobile body in the reproduction travel mode.
In step S171, the control unit 11 executes data measurement processing. Specifically, the control unit 11 receives the reflected light reflected from the obstacle by irradiating the laser beam with the front laser range sensor 53 and the rear laser range sensor 54, and thereby the obstacle related to the distance and direction of the obstacle. Get material information. The data measurement process here is the same as the data measurement process in the teaching travel mode, and detailed description thereof is omitted.

  In step S172, the control unit 11 executes a local map creation process using the sensor coordinate system. Specifically, based on the obstacle position information table acquired by the front laser range sensor 53 and the rear laser range sensor 54 (for example, the obstacle position information table shown in FIG. 7), a coordinate system centered on the sensor is used. A local map in which the obstacle existence probability is made to correspond to each grid in the vicinity of the self-location is created.

  In step S173, the control unit 11 determines whether or not the local map created by the map creation unit 22 satisfies a predetermined condition. Specifically, when the ratio of obstacles present in the local map created by the map creation unit 22 is less than or equal to a predetermined value, or the obstacles included in the local map are simple shapes Therefore, the control unit 11 determines that the local map satisfies the predetermined condition, and proceeds to step S174. In addition, when the global map stored in the vicinity of the self position among the global maps stored in the storage unit 27 is compared with the created local map, the obstacle existence probability in each corresponding grid is greatly different. In addition, the control unit 11 determines that the local map satisfies the predetermined condition, and proceeds to step S174.

In step S174, the control unit 11 determines whether or not the state where the created local map satisfies a predetermined condition is continuous for a predetermined travel distance or continuous for a predetermined time.
Specifically, for example, when the vehicle is passing in the vicinity of the third passing point 913 in FIG. 9, the state in which the ratio of obstacles in the local map is equal to or less than a predetermined value continuously for a predetermined distance or a predetermined time. Arise. Therefore, in such a case, the control unit 11 determines that an accurate self-position on the global map cannot be specified (Yes in Step S174), and proceeds to Step S175.
Further, for example, when the vehicle is passing near the second passing point in FIG. 9, a state where the obstacle existing in the local map has a simple shape continues for a predetermined distance or a predetermined time. Even in such a case, the control unit 11 determines that an accurate self-position on the global map cannot be specified (Yes in Step S174), and proceeds to Step S175.
Further, the control unit 11 compares the global map stored in the storage unit 27 with the local map created by the map creation unit 22 and determines that a greatly different state continues for a predetermined distance or a predetermined time. In the case, the process proceeds to step S175. For example, when the layout in the travel route has changed significantly from the time when the global map is created, the control unit 11 determines that an accurate self-position on the global map cannot be specified, and proceeds to step S175. .
In step S174, when the control unit 11 determines that the state in which the local map satisfies the predetermined condition is not continuous for a predetermined distance or for a predetermined time (No in step S174), the control unit 11 proceeds to step S176.

In step S175, the control unit 11 forcibly stops traveling in the traveling unit 25. Specifically, the control unit 11 causes the travel command generation unit 23 to generate a forced stop travel command and causes the travel control unit 24 to output the forced stop travel command.
Thereafter, the control unit 11 ends the process.

  In step S176, the control unit 11 estimates the self position on the global map. Specifically, the self-position estimating unit 72 adds the amount of movement (dx, dy, dθ) between time (t−1) and time (t) to the self-position at time (t−1) to obtain the time ( Estimated as the self-position of t). Further, the control unit 11 assumes a three-dimensional grid space (x, y, θ) having two-dimensional coordinates (x, y) and a traveling direction (θ) as parameters of the self-position, and each grid before and after the movement. Map matching is performed by calculating the probability distribution of the self-position.

  In step S177, the control unit 11 compares the global map corresponding to the estimated self-location with the local map created from the location information of the obstacle acquired by the obstacle information acquisition unit 21, and sets the current self-location. Estimate the time in the corresponding travel schedule. At this time, the control unit 11 can use a global map stored in the storage unit 27 in advance or a global map created in the teaching travel mode and stored in the storage unit 27. In addition, when the environmental map restoration data is stored in the storage unit 27 in the teaching travel mode, the control unit 11 restores the global map in the vicinity of the own position using the environmental map restoration data, You can also compare with a local map.

In step S <b> 178, the control unit 11 reads from the storage unit 27 the passing point data corresponding to the estimated time in the travel schedule stored in the storage unit 27.
In step S179, the control unit 11 executes a global map update process. Specifically, the control unit 11 adds the local map of the absolute coordinate system obtained by performing coordinate transformation using the self-position estimated by the self-position estimation process to the global map at time (t−1), Update the global map at (t). In this global map update process, the obstacle existence probability of each grid of the global map is updated by adding the obstacle existence probability of each grid corresponding to the local map in the absolute coordinate system, so that wheel slip and wear The influence of noise based on the measurement error of the sensor can be reduced as much as possible.
In the teaching travel mode, the control unit 11 leaves the global map in the vicinity of the passing point in the storage unit 27 when the environmental map restoration data corresponding to the passage time is created and stored in the storage unit 27. It is also possible to delete global map information.

  In step 180, the control unit 11 creates a control amount for the actuator of the traveling unit 25 and outputs it to the traveling unit 25. Specifically, the control unit 11 determines the control amount of the actuator based on the self-position on the global map and the passing point data read from the storage unit 27 so as to follow the travel schedule. Based on the travel command, the travel unit 25 is input.

  In step S181, the control unit 11 determines whether or not to end the reproduction travel mode. Specifically, when the final stop point data in the travel schedule stored in the storage unit 27 is reached, when the emergency stop switch 56 is operated by the operator, the control unit 11 uses the bumper switch 55 to obstruct the obstacle. When the obstacle position information detected by the obstacle information acquisition unit 21 is close to a predetermined distance or less, the control unit 11 determines to end the reproduction travel mode. When the control unit 11 determines not to end the reproduction travel mode, the control unit 11 proceeds to step S181.

As described above, in the teaching travel mode described with reference to FIG. 4, when the travel start position is inappropriate (No in step S41), the travel command generation unit 23 generates a forced stop travel command. Thus, traveling by the traveling unit 25 is stopped (step S42). Therefore, in the subsequent reproduction travel mode, it is possible to avoid a situation in which the travel start position cannot be specified and travel is impossible.
Further, in both the teaching traveling mode (FIG. 4) and the reproduction traveling mode (FIG. 17), even when shapes that are difficult to estimate the self-position continue (Yes in step S47 in the teaching traveling mode, the reproducing traveling mode). In this case, Yes in step S174), the travel command generating unit 23 generates a forced stop travel command to stop the travel by the travel unit 25 (step S48 in the teaching travel mode, in the reproduction travel mode) Is step S175). In this case, it is possible to avoid a situation in which the vehicle cannot travel according to the travel schedule due to a large shift in the estimation of the self position due to wheel wear or slip, sensor measurement error, or the like.
Furthermore, even when there is a large layout change since the creation of the global map, it is difficult to specify the self-position and it is possible to avoid being unable to run.

(4) Features of Embodiments The above embodiments have the following configurations and functions.
The autonomous mobile body 1 (an example of an autonomous mobile body) includes a travel unit 25 (an example of a travel unit), a travel control unit 24 (an example of a travel control unit), and an obstacle information acquisition unit 21 (an obstacle information acquisition unit). An example), a map creation unit 22 (an example of a map creation unit), and a travel command generation unit 23 (an example of a travel command generation unit). The traveling unit 25 includes a plurality of motors 93 and 94 (an example of an actuator) to which control amounts are input. The traveling control unit 24 generates a control amount to be input to the motors 93 and 94 based on the traveling command (for example, step S43 and step S180). The obstacle information acquisition unit 21 acquires position information of obstacles existing around each predetermined control cycle (for example, step S44 and step S171). The map creation unit 22 creates a local map (an example of a local map) indicating the presence or absence of an obstacle in the peripheral area of the main body 1a based on the position information of the obstacle (for example, step S45 and step S172). When the local map satisfies a predetermined condition, the travel command generation unit 23 outputs a forced stop travel command to the travel control unit 24 (for example, steps S42, S48, and S175).

In this autonomous mobile body 1, a local map located around the main body 1 a is created based on the obstacle position information acquired by the obstacle information acquisition unit 21, and passing points that are difficult to reliably estimate its own position are continuous. When a predetermined condition is satisfied, as in the case where a position where it is difficult to reliably estimate the self-position is set as the travel start position, the travel command generation unit 23 generates a forced stop travel command. As a result, during autonomous traveling of the autonomous mobile body 1, it is possible to prevent traveling out of the traveling schedule due to wheel wear and slip, sensor measurement errors, and the like.
(5) Other Embodiments Although one embodiment of the present invention has been described above, the present invention is not limited to the above embodiment, and various modifications can be made without departing from the scope of the invention. In particular, a plurality of embodiments and modifications described in this specification can be arbitrarily combined as necessary.

(5-1) Environment Map Update / Delete Processing The control unit 10 can manage the environment map with a two-layer structure of a static environment map and a dynamic environment map. The static environment map can be restored based on the map restoration data created in the teaching travel mode. The dynamic environment map is superimposed with the dynamic environment map at time (t) and the local map based on the position information of the obstacle in the vicinity of the own position acquired by the obstacle information acquisition unit 21 at time (t + 1). Can be created by combining.

For example, the obstacle existence probability in each grid of the dynamic environment map at time (t) is set to DynamicMap (t), and the obstacle existence probability in each grid of the static environment map is set to StaticMap. Further, at time (t), the obstacle existence probability in each grid of the difference map indicating the obstacle arrangement change from the static environment map is defined as DiffMap (t).
Furthermore, if the parameter that determines the range of values of each grid of the dynamic environment map is P1,
DynamicMap (t) = StaticMap · P1 + DiffMap (t)
Can be calculated as

The control unit 10 estimates the self-position using the obstacle information acquired by the obstacle information acquisition unit 21 and the surrounding information of the own machine and the static environment map restored from the environment map restoration data. Create a local map of the system.
As the dynamic environment map at time (t + 1), the obstacle existence probability of each grid is calculated using the dynamic environment map at time (t) and the local map at time (t + 1). When the obstacle existence probability of each grid in the local map of the absolute coordinate system is LocalMap (t + 1),
DynamicMap (t + 1) = DynamicMap (t) + LocalMap (t + 1)
Can be calculated as

The control unit 10 obtains the difference between the dynamic environment map at time (t + 1) and the static environment map, and calculates DiffMap (t + 1) that is the obstacle existence probability of each grid of the difference map at time (t + 1). (T + 1) = DynamicMap (t + 1) −StaticMap · P1
And update the dynamic environment map using this.

FIG. 18 is an explanatory diagram of a dynamic environment map update process. In the example shown in FIG. 18, a 4 × 4 grid near the self-position is targeted, (a) is a dynamic environment map at time (t), (b) is a static environment map, and (c) is a time. The difference map in (t) is shown.
As shown in FIG. 18, the value obtained by multiplying the obstacle existence probability of each grid in the difference map DiffMap (t) at time (t) by the parameter P1 to the obstacle existence probability of each grid in the static environment map StaticMap. And the dynamic environment map DynamicMap (t) at time (t) is calculated. The obstacle existence probability of each grid is a value in the range of −1.0 to 1.0, and the parameter P1 is set to “0.5”.
From this, the obstacle existence probability of each grid of the dynamic environment map calculated by DynamicMap (t) = StaticMap · P1 + DiffMap (t) is as shown in FIG.

FIG. 19 is an explanatory diagram of likelihood calculation at the time of self-position estimation. In the example shown in FIG. 19, a 4 × 4 grid in the vicinity of the self-position is targeted, (a) is a local map in the absolute coordinate system, (b) is a static environment map, and (c) is a dynamic environment map. It is.
FIG. 19A shows the obstacle existence probability of each grid after conversion to the local map of the absolute coordinate system based on the obstacle position information around the own machine acquired by the obstacle information acquisition unit 21. .

  In the local map of the absolute coordinate system, when there are N grids whose obstacle existence probability is “1.0”, the coordinates of the i-th grid are (gx_occupy (i), gy_occupy (i)). The obstacle existence probability in the coordinates (gx_occupy (i), gy_occupy (i)) of the static environment map is set to StaticMap (gx_occupy (i), gy_occupy (i)), and the coordinates of the dynamic environment map (gx_occupy (i), gy_occupy) Assume that the obstacle existence probability in (i)) is DynamicMap (gx_occupy (i), gy_occupy (i)).

In the local map of the absolute coordinate system, for each grid where the obstacle existence probability is “1.0”, either the static environment map or the dynamic environment map having the higher obstacle existence probability is used. The likelihood calculation is performed using the following formula.


Thus, in the example shown in FIG. 19, the likelihood can be calculated as s = (1.0 + 1.0 + 0.3 + 0.3) /4=0.65.

In this way, even when the global map is composed of two layers, a static environment map and a dynamic environment map, the problem is that the circular route does not close because the dynamic environment map after passing is deleted. it can.
In addition, even when the location information of an obstacle has changed when the static environment map is restored using the environment map restoration data created in the teaching travel mode, the difference map is used. It becomes possible to detect a change in the arrangement of the obstacle. Accordingly, the autonomous mobile body 1 can avoid contact with an obstacle in response to a layout change on the travel route or the presence of the mobile body.

  The present invention can be applied to a cleaning robot, a transfer robot, and other autonomous mobile bodies that perform autonomous traveling.

DESCRIPTION OF SYMBOLS 1 Autonomous mobile body 1a Main body 11 Control part 21 Obstacle information acquisition part 22 Map preparation part 23 Travel command generation part 24 Travel control part 25 Travel part 26 Teaching data preparation part 27 Memory | storage part 31 Distance sensor 32 Operation part 33 Display part 40 Cleaning robot 42 Throttle 43 Cleaning unit user interface 51 Control board 52 Breaker 53 Front laser range sensor 54 Rear laser range sensor 55 Bumper switch 56 Emergency stop switch 57 Speaker 63 SLAM processing unit 64 Non-volatile memory 72 Self-position estimation unit 73 Map matching unit 81 arithmetic unit 82 determination unit 83 movement control unit 91 motor drive unit 92 motor drive unit 93 motor 94 motor 95 encoder 96 encoder 101 scan area 103 scan area 105 mask area 107 detection grid 109 intermediate grid 11 Unknown grid 115 First scanning area 117 Second scanning area 131 Scanning line 132 Origin 133 Detection grid 134 Intermediate grid 135 Intermediate grid 136 Intermediate grid 137 Intermediate grid 138 Unknown grid 901 Traveling path 911 First passage point 912 Second Passing point 913 third passing point 914 fourth passing point 915 fifth passing point 921 local map 922 local map 923 local map 924 local map 925 local map

Claims (6)

The body,
A traveling unit having a plurality of actuators to which control amounts are input;
A travel control unit that generates a control amount to be input to the actuator based on a travel command;
For each predetermined control cycle, an obstacle information acquisition unit that acquires position information of obstacles existing around,
Based on the position information of the obstacle, a map creation unit that creates a local map indicating the presence or absence of an obstacle in the peripheral area of the main body,
When the local map satisfies a predetermined condition, a travel command generation unit that outputs a travel command for forced stop to the travel control unit;
Equipped with a,
While performing manual travel from the travel start position to the travel end position by an operator's manual operation, while creating a travel schedule that is not a set of passing time and passing point data at the passing time, and reproducing the travel schedule, Reappearance traveling mode to autonomously travel from the travel start position to the travel end position ,
The traveling instruction generator, the local map in the teaching mode of travel you determine whether a predetermined condition is satisfied, the autonomous moving body.   When the local map in which the ratio of obstacles in the surrounding area is equal to or lower than a predetermined value is continuous for a predetermined travel distance or a predetermined travel time, the travel command generation unit The autonomous mobile body according to claim 1, wherein the autonomous mobile body determines that the condition is satisfied.   When the local map in which the obstacle present in the surrounding area has a simple shape with a small shape change is continuous for a predetermined travel distance or a predetermined travel time, the travel command generation unit is configured so that the local map is a predetermined shape. The autonomous mobile body according to claim 1, wherein the autonomous mobile body determines that the condition is satisfied. A storage unit for storing an environmental map of the surrounding environment;
A self-position estimating unit for estimating a self-position on the environmental map;
Further comprising
The travel command generation unit compares the environment map around the self-location estimated by the self-position estimation unit with the local map, and a state that is different beyond a predetermined ratio is a predetermined travel distance or a predetermined The autonomous mobile body according to any one of claims 1 to 3, wherein when the running time continues, it is determined that a predetermined condition is satisfied.   In the local map created by the map creation unit before the start of traveling, if the percentage of obstacles existing in the peripheral area is less than or equal to a predetermined value, the obstacles existing in the peripheral area are simply changed in shape. 5. The driving command generation unit according to claim 4, wherein the driving command generation unit determines that the local map satisfies a predetermined condition when it is a shape or when it differs from an environmental map around a self-position by a predetermined ratio. Autonomous mobile body. The autonomous mobile body according to claim 1 , wherein the travel command generation unit determines whether or not the local map satisfies a predetermined condition in the reproduction travel mode.