WO2016067640A1

WO2016067640A1 - Autonomous moving device

Info

Publication number: WO2016067640A1
Application number: PCT/JP2015/052391
Authority: WO
Inventors: 清水　一寿; 翔貴吉野
Original assignee: シャープ株式会社
Priority date: 2014-10-28
Filing date: 2015-01-28
Publication date: 2016-05-06
Also published as: JP2016085689A; JP5902275B1

Abstract

The purpose of the present invention is to reduce computational volume for virtually measuring distances to obstacles upon a grid map. Provided is an autonomous moving device, a grid map (GM2) whereof comprising a plurality of grid elements. Each grid element retains obstacle information which represents an obstacle (OB), adjacent region information which represents an adjacent region (R2) which is adjacent to the obstacle (OB), or distant region information which represents a distant region (R3) which is removed from the obstacle (OB).

Description

Autonomous mobile device

The present invention relates to an autonomous mobile device such as a cleaning robot that can move autonomously.

In recent years, autonomous mobile devices such as autonomous mobile robots that can autonomously move within a limited area inside a building or outdoors based on surrounding environmental information are being developed. Such an autonomous mobile device enables autonomous movement by creating a movement route from the current position to a specific target point on a movement map stored in advance with respect to the movement area. For this reason, in general, an autonomous mobile device has a function of recognizing its own position in a movement area.

In order to control such an autonomous mobile device, as a technique for allowing the autonomous mobile device to recognize its own position, for example, a reflector whose position is known in advance is installed in the surrounding environment and a laser range finder provided in the autonomous mobile device. A method is known in which the self-position is calculated based on the principle of triangulation by measuring the reflector.

However, the above-described method has a problem that a complicated work is required to install the reflector, and a large-scale facility work is required for self-position estimation. For this reason, there is a need for an infrastructureless self-position estimation method that does not require extensive construction work.

As a method for estimating the self-location without requiring large-scale equipment construction, a method for estimating the self-location by comparing the map information of the moving area with the acquired environmental information has been proposed. For example, the autonomous mobile device described in Patent Document 1 recognizes its own position by comparing environmental information acquired by a laser range finder with map information. As such a self-position estimation method, a method using a grid map is often used. The grid map is a map showing a moving region in a simulated manner, and presence information such as walls and obstacles is registered on the map (for example, whether or not a predetermined area on the map is a wall is 0). Or it is registered with one digital information).

Japanese Patent Gazette “Japanese Unexamined Patent Publication No. 2009-223757 (Published October 1, 2009)”

In such a self-position estimation method using the grid map, the autonomous mobile device compares the data measured by the autonomous mobile device in the actual environment with the data virtually measured by the autonomous mobile device on the grid map. Estimate the actual self-position. If the autonomous mobile device is equipped with a laser range finder, the actual position of the autonomous mobile device can be determined by comparing the data acquired by the laser range finder with the data virtually measured on the grid map. It is possible to calculate with high accuracy.

Here, in the calculation of data virtually measured by the autonomous mobile device on the grid map, the distance measurement from the autonomous mobile device to the surrounding obstacles by the actual measuring device is simulated. If the autonomous mobile device is equipped with a laser range finder as described above, the laser light for distance measurement from the virtual position of the own aircraft virtually set on the grid map for self-position estimation calculation by the grid map The distance from the virtual position of the aircraft to the obstacle is determined by extending the straight line simulating the obstacle to the surrounding obstacles and determining whether there is a collision between the straight line and the obstacle by a program calculation process based on the collision judgment. Is calculated. In such a calculation, a straight line is extended from the position of the aircraft by one grid element constituting the grid map, and a program calculation process is performed based on a collision determination between the grid element grid and an obstacle. If it is determined by the collision determination that there is no collision and there is no obstacle, the straight line is extended by the next one cell, and the determination of the collision between the grid element 1 cell and the obstacle is performed again.

For this reason, it is necessary to repeat a large amount of program calculation processing based on collision determination just by virtually measuring the distance to an obstacle for a straight line simulating one laser beam. Furthermore, for the self-position estimation calculation, it is necessary to repeat the distance measurement based on this virtual straight line for the multiple virtual self-positions (self-position candidate points) arranged as the virtual position of the own aircraft. The calculation amount of the program calculation process is enormous. For this reason, in order for the autonomous mobile device to perform the self-position estimation calculation as described above while moving in the real space, there is a problem that an arithmetic device with high calculation performance is required.

An object of the present invention is to provide an autonomous mobile device that can reduce the amount of calculation for virtually measuring the distance to an obstacle on a grid map.

In order to solve the above problems, an autonomous mobile device according to an aspect of the present invention includes a sensor that detects environmental information related to an obstacle, and an obstacle based on the environmental information detected by the sensor. A generation unit that generates a grid map; and an estimation unit that calculates a virtual distance from the self-position candidate point to the obstacle on the grid map generated by the generation unit, and estimates the position of the grid. The map includes a plurality of grid elements, and each grid element represents obstacle information representing the obstacle, neighboring area information representing an neighboring area adjacent to the obstacle, and a remote area away from the obstacle. Any of the remote area information is retained.

According to one aspect of the present invention, there is an effect that it is possible to reduce the amount of calculation for virtually measuring the distance to the obstacle on the grid map.

It is a figure which shows the relationship between the autonomous mobile device which concerns on Embodiment 1, and a movement area. It is a top view which shows typically the structure of the said autonomous mobile apparatus. It is a figure for demonstrating the structure of the grid map mounted in the said autonomous mobile apparatus. It is a block diagram which shows the structure of the calculating part provided in the said autonomous mobile apparatus. It is a figure for demonstrating the grid map produced | generated by the grid map production | generation part provided in the said calculating part. It is the figure which showed typically the measurement state by the laser range finder provided in the said autonomous mobile apparatus. It is the figure which showed the state which is implementing the virtual calculation of the distance to an obstruction by the said grid map. It is the figure which showed the state which is implementing the virtual calculation of the distance to an obstruction with the grid map which concerns on Embodiment 2. FIG. It is the figure which showed the state which is implementing the virtual calculation of the distance to an obstruction with the grid map which concerns on Embodiment 3. FIG. It is the figure which showed the state which is implementing the virtual calculation of the distance to an obstruction with the grid map which concerns on Embodiment 4. FIG.

Hereinafter, embodiments of the present invention will be described in detail.

[Embodiment 1]
The autonomous mobile device according to the first embodiment of the present invention includes a laser range finder as a sensor that acquires environmental information indicating the position and shape of an object existing in a moving area, and uses the moving map related to the moving area, It is an autonomous mobile device that can move while recognizing its own position based on environmental information acquired by a laser range finder.

The autonomous mobile device according to Embodiment 1 of the present invention will be described with reference to the drawings. The autonomous mobile device in the present embodiment is a device that autonomously moves on a plane by driving a pair of wheels.

FIG. 1 is a diagram illustrating a relationship between the autonomous mobile device 1 and the movement area 102 according to the first embodiment. Referring to FIG. 1, the vehicle-type autonomous mobile device 1 moves in a limited movement area 102 (area surrounded by a broken line) on the floor 101. In FIG. 1, it is assumed that an obstacle is not particularly installed in the moving area 102 on the floor 101, and the autonomous mobile device 1 can arbitrarily move in the moving area 102. .

FIG. 2 is a plan view schematically showing the configuration of the autonomous mobile device 1. As shown in FIG. 2, the autonomous mobile device 1 is an opposing two-wheeled mobile device including a box-shaped main body 201, a pair of

opposing wheels

202 a and 202 b, and a caster 203. These

wheels

202a and 202b and casters 203 support the main body 201 horizontally. Further, in the main body 201,

motors

204a and 204b for driving the

wheels

202a and 202b,

encoders

205a and 205b for detecting the rotation speeds of the

wheels

202a and 202b, and

wheels

202a and 202b, respectively, are driven. Are provided, and a calculation unit 206 that transmits the control signal to the

motors

204a and 204b, and a battery 207 for supplying power to these components. A storage area 206a such as a memory provided as a storage unit provided inside the calculation unit 206 has a program for controlling the movement speed, movement direction, movement distance, and the like of the autonomous mobile device 1 based on the control signal. The shape and size of the moving area 102 are stored as map information.

Furthermore, a laser range finder 2 for recognizing environmental information in the direction (front) in which the autonomous mobile device 1 moves is provided on the front surface of the main body 201. The laser range finder 2 omits a detailed structure, but receives a light source for irradiating laser light at a predetermined spread angle with respect to the front of the main body 201 and reflected light of the laser light emitted from the light source. And a light receiving portion. Object detection (sensing) based on the so-called TOF (Time （of flight) principle of detecting the position of an obstacle that reflected the laser light based on the angle of irradiation with the laser light and the time required for reflection. ) Is performed. Details of the procedure for obtaining environmental information (position and shape of an object to be sensed) by the laser range finder 2 will be described later.

In addition, the autonomous mobile device 1 includes a self-position acquisition unit for acquiring its own position. This self-position acquisition unit includes the

encoders

205a and 205b and the calculation unit 206 described above. That is, by calculating the rotational speed of the

wheels

202a and 202b detected by the

encoders

205a and 205b in the calculation unit 206, information such as the moving speed and distance of the autonomous mobile device 1 is obtained. The self position (odometry position) of the autonomous mobile device 1 is calculated.

And the position information as a self-position candidate obtained from these self-position acquisition units is appropriately corrected based on environment information obtained by a laser range funder 208 described later. Details of the method for correcting the position information will be described later.

The autonomous mobile device 1 configured in this way independently controls the drive amount of the pair of

wheels

202a and 202b, thereby moving straight, curving (turning), moving backward, and rotating on the spot (the midpoint of both wheels). Movement such as turning around the center. Then, the autonomous mobile device 1 determines the movement route to the designated destination in the movement area 102 in accordance with a command from a PC (Personal Computer) or the like (not shown) provided outside that designates the movement location. The destination is reached by creating autonomously and moving to follow the movement route. In addition, although autonomous creation of a movement route is performed in the arithmetic unit 15, description is abbreviate | omitted here.

Next, a grid map as a movement map created based on the shape of the movement area 102 and an object (an obstacle such as a wall) included in the movement area 102 stored in the storage area 206a will be described. In the storage area 206a provided inside the arithmetic unit 206, the shape of the entire moving area 102 on the floor 101 is set as an outer frame, and is arranged in the horizontal direction and the vertical direction at a constant interval m (for example, 5 cm) inside. A grid map obtained by virtually describing grid lines connecting lattice points is stored.

FIG. 3 illustrates an example of the grid map described above. The grid map GM1 depicts grid lines 303 that connect the lattice points arranged in the horizontal direction and the vertical direction at a fixed interval m in the outer frame 302 simulating the shape of the moving region 102 in the horizontal direction and the vertical direction. Is. Then, by the grid element GE surrounded by the grid line 303, the location corresponding to the autonomous position of the autonomous mobile device 1, the movement end point that is the destination, and the movement direction of the autonomous mobile device 1 at the movement end point are specified. Is done. Note that the interval m can be appropriately changed according to conditions such as the curvature of the autonomous mobile device 1 and the accuracy of recognizing the absolute position. The interval m can also be used as a threshold value when it is determined that a slip has occurred. Further, when it is known that fixed objects such as walls and obstacles exist in the moving area 102, the storage area is stored in a form in which the positions of these known objects are registered in advance on the grid map GM1. Assume that it is stored in 206a.

FIG. 4 is a block diagram illustrating a configuration of the calculation unit 206 provided in the autonomous mobile device 1. The calculation unit 206 includes a grid map generation unit 3 (generation unit) and a self-position estimation unit 4 (estimation unit). The grid map generation unit 3 generates a grid map showing obstacles based on environmental information detected by the laser range finder 2 (sensor). The self-position estimation unit 4 calculates a virtual distance from the self-position candidate point to the obstacle on the grid map generated by the grid map generation unit 3 and estimates its own position.

Here, distance information from an obstacle in each grid element of the grid map will be described. FIG. 5 is a diagram for explaining the grid map GM2 generated by the grid map generation unit 3 provided in the calculation unit 206. The grid map GM2 is divided into an area where the obstacle OB exists and an area where the autonomous mobile device 1 can travel. The distance information from the obstacle OB is added to the grid elements of the grid map GM2 by the following procedure.
(1) Each grid element is pre-written with obstacle information indicating whether or not the obstacle OB is in the obstacle region R1. For example, when the grid element is the obstacle region R1 where the obstacle OB exists, the distance “0” (obstacle information) is written in the grid element and is not the obstacle region R1 where the obstacle OB exists. In this case, “−1” is written in the grid element.
(2) For one of the grid elements in which “−1” is written, if the adjacent grid element is the obstacle region R1 where the obstacle OB exists, the distance “1” (adjacent region information) to the grid element Will be overwritten.
(3) The above (2) is repeated for all grid elements in which “−1” is written.
(4) For one of the grid elements in which “−1” is written, if the distance “1” is written in the adjacent grid element, the distance “2” (remote area information) is overwritten in the grid element. The
(5) The above (4) is repeated for all grid elements in which “−1” is written.
(6) Similarly, if the distance “n−1” is written in the adjacent grid element for one of the grid elements in which “−1” is written, the distance “n” is overwritten in the grid element. (N is an integer greater than or equal to 3), and this is repeated for all grid elements in which “−1” is written.
(7) The above (6) is repeated until all grid elements of the grid map are overwritten with the distance “0” or more.

In the example shown in FIG. 5, the distance “0” representing the obstacle information is written in the obstacle region R1 corresponding to the grid element where the obstacle OB exists. Then, the distance “1” of the adjacent area information representing the adjacent area R2 adjacent to the obstacle OB is written in the adjacent area R2 corresponding to the grid element adjacent to the obstacle area R1. Grid elements other than the grid elements corresponding to the obstacle region R1 and the adjacent region R2 are included in the remote region R3 away from the obstacle OB. The remote area R3 includes a first remote adjacent area R4 adjacent to the adjacent area R2 from the opposite side of the obstacle OB, and a second remote adjacent area R5 adjacent to the first remote adjacent area R4 from the opposite side of the obstacle OB. Including.

In the first remote adjacent area R4, the remote area information distance “2” representing the remote area away from the obstacle OB is written. The distance “3” of the remote area information is written in the second remote adjacent area R5. The distance value of the remote area information increases as the distance from the obstacle OB increases, and corresponds to the number of other grid elements existing between the obstacle OB.

The work of giving information related to the distance from the obstacle OB to the grid map GM2 only needs to be performed once immediately after the movement map is created by the autonomous mobile device 1, and thereafter the same grid map is used. can do. Note that incidental information such as the type of the obstacle OB and the entry prohibition area of the autonomous mobile device 1 may be given to each grid element GE. However, in order to simplify the description, description of these pieces of information is omitted.

FIG. 6 is a diagram schematically showing a measurement state by the laser range finder 2 provided in the autonomous mobile device 1. Next, a method for acquiring environmental information on the front surface of the autonomous mobile device 1 using the laser range finder 2 described above will be briefly described. First, when the reflected light from the obstacle OB of the actual light beam RL by the laser irradiated from the laser range finder 2 of the autonomous mobile device 1 is received by the laser range finder 2, the autonomous mobile device 1 when irradiated with the actual light beam RL. The point at which the irradiated actual light beam RL is reflected from the self-position, the irradiation direction of the actual light beam RL irradiated from the laser range finder 2, and the time from receiving the actual light beam RL to receiving the reflected light. Identified.

The data measured by the laser range finder 2 is output as a polar coordinate system defined by the measured distance Rd from the laser range finder 2 and the angle φ in the irradiation direction of the actual light beam RL. For example, when the resolution when measuring 270 degrees forward of the traveling direction of the autonomous mobile device 1 is 1.0 degree, 271 data are measured by one scan. Such a sensing result by the laser range finder 2 is stored in the storage area 206a.

FIG. 6 shows a state in which an obstacle OB is provided in front of the autonomous mobile device 1 in the movement area. As shown in FIG. 6, the autonomous mobile device 1 irradiates the front surface with a real light RL from a laser range finder 2. At this time, an obstacle OB is included in the sensing area of the laser range finder 2. In such a case, the distance to the obstacle OB sensed by the laser range finder 2 is stored in the storage area 206a. A grid map is generated based on the environmental information acquired in this way and stored in the storage area 206a.

FIG. 7 is a diagram showing a state in which the virtual calculation of the distance to the obstacle OB is performed by the grid map GM2. Next, as a technique for estimating the current self-position from among the self-position candidates calculated by the odometry method using the environment information acquired by the laser range finder 2, this embodiment is called MCL (Monte-Carlo-Localization). Adopt a method. This is a technique for expressing the probability expression of the robot's location using position candidate points called particles P (self-position candidate points).

Here, a calculation method using distance information from the grid element to the obstacle OB according to the present embodiment will be described. In the present embodiment, a distance calculation method for performing distance calculation by irradiating a virtual light beam from a particle P that is a candidate for the self position in the MCL calculation toward the surroundings will be described.

First, the self-position estimation unit 4 provided in the calculation unit 206 extends a straight line from the position of the particle P along the measurement direction of the laser range finder 2 on the grid map GM2. That is, the virtual ray VL is virtually depicted from the position of the particle P on the grid map GM2.

Here, referring to FIG. 7, if the distance calculation is performed by extending the virtual ray VL from one particle P, the virtual ray VL extended at a predetermined angle from the grid element including the particle P is used as the grid element. Pass in order. Here, regarding the grid map GM2, the horizontal direction is the x direction and the vertical direction is the y direction. In order for the virtual ray VL to pass through the grid elements of the grid map GM2, the inclination of the virtual ray VL is known in advance, so that when the virtual ray VL is advanced in the x direction or the y direction by one square, another virtual ray VL is obtained. If the grid of the grid element of the virtual ray VL is advanced in the axis, that is, the x-direction, the y-direction is determined. If the grid is advanced in the y-direction, the inclination of the virtual ray VL is determined whether the grid is advanced in the x-direction. Based on the calculation method. This is a technique called Bresenham's straight line drawing algorithm.

While the value of the distance written in the grid element through which the virtual ray VL passes is 2 or more, it is determined that there is no possibility that the grid element collides with the obstacle OB. Become. Then, it is considered that the collision determination process may be performed after the distance value written to the grid element through which the virtual ray VL passes becomes 1.

For example, the distance “5” is written in the grid element including the particle P. Since “5” that is “2” or more is written in the grid element in the seventh row from the top and the third column from the right through which the virtual ray VL from the particle P passes, the collision determination process is not performed. . Similarly, “5” is written in the grid element in the sixth row from the top and the third column from the right through which the virtual ray VL passes, and in the grid element in the fifth row from the top and the third column from the right, “4” is written, and “3” is written in the grid element in the fourth row from the top and the third column from the right. Then, “3” is written in the grid element in the fourth row from the top and the second column from the right through which the virtual ray VL passes, and “3” is written in the grid element in the third row from the top and the second column from the right. 2 "is written. For this reason, collision determination processing is not performed for these grid elements in which “5” to “2” are written. Since “1” is written in the grid element in the second row from the top and the second column from the right through which the virtual ray VL passes, the collision determination process is performed.

As described above, the collision determination process is not performed for the grid element in which “2” or more is written, and the collision determination process is performed for the grid element in which “1” is written. It is possible to reduce the amount of calculation for virtually measuring the distance between the two on the grid map.

When the value corresponding to the distance from the obstacle OB is not written in each grid element of the grid map, the value of the grid element is called every time the virtual ray VL advances one grid line on the grid map, A collision determination process for determining whether or not the grid element is an obstacle OB is required. According to the present embodiment, the virtual ray VL is displayed in the grid without performing the reading of the value of the grid element and the collision determination process. Can pass through the element.

As described above, while the distance value written in the grid element is 2 or more, the virtual ray VL can pass through the grid element without performing the collision determination, but the distance value written in the grid element. After the value becomes 1, the virtual ray VL moves the grid element by one cell, and the grid element after the movement determines the collision with the obstacle OB. If the grid element is an obstacle OB, the length of a straight line extending from the particle P to the obstacle OB can be obtained from the position of the grid element and the position of the grid element including the particle P. Thus, distance data between the particle P imitating measurement with one virtual ray VL and the obstacle OB can be calculated. Thereafter, the distance measurement between each particle and the obstacle OB is repeated in the same manner.

Thus, the distance measurement between the particle P and the obstacle OB according to the present embodiment is executed. Thereafter, the calculation of MCL continues, but the description is omitted here.

The calculation method according to the present embodiment does not need to perform the collision determination process with the obstacle OB for each grid element of the grid map, so that the calculation can be performed at high speed. In particular, in MCL, the distance calculation to the obstacle OB by the virtual ray VL is repeated for a plurality of angles for each of several hundred particles. For this reason, even if only one collision determination process is omitted, in one self-position estimation calculation, the reduction effect of the calculation process is large. Furthermore, since the calculation of MCL is repeatedly executed while the autonomous mobile device 1 moves, the autonomous mobile device 1 can be operated in a shorter cycle if the calculation processing load of the MCL is reduced and the self-position estimation time per time is shortened. Self-position estimation becomes possible, and the accuracy of the actual movement position can be improved.

[Embodiment 2]
The following will describe another embodiment of the present invention with reference to FIG. For convenience of explanation, members having the same functions as those described in the embodiment are given the same reference numerals, and descriptions thereof are omitted.

In the second embodiment, the value N corresponding to the distance from the obstacle OB in the grid element including the particle P in the calculation of the process in which the virtual ray VL passes through each grid element from the particle P in the grid map GM2. Based on the above, the calculation start position when the virtual ray VL sequentially passes through the grid elements is changed according to the following procedure.
(1) The coordinate of the position of the grid element containing the particle P is set to (xi, yi).
(2) The angle of the virtual ray VL emitted from the particle P is known, and its value is φ. (3) When the coordinates of the calculation start position SP related to the collision determination process when the virtual ray VL passes through the grid element are (xj, yj),
xj = N × cosφ, yj = N × sinφ
To obtain the coordinates of the calculation start position SP related to the collision determination process.

As described above, the calculation start position SP related to the collision determination process on the grid map is skipped from the particle P to a position close to the obstacle OB as described above. For this reason, the collision determination process is not necessary for the grid element between the particle P and the start position SP, and may be performed only for the grid element between the start position SP and the obstacle OB. Therefore, the amount of calculation for the virtual ray VL to perform the collision determination process on the grid element is further reduced.

[Embodiment 3]
An autonomous mobile device according to Embodiment 3 will be described. FIG. 9 is a diagram illustrating a state in which the virtual calculation of the distance to the obstacle OB is performed by the grid map GM3 according to the third embodiment. For convenience of explanation, members having the same functions as those described in the embodiment are given the same reference numerals, and description thereof is omitted.

In the third embodiment, information corresponding to the distance from the obstacle OB is given by the following procedure for generating the grid map GM3 that gives the information corresponding to the distance from the obstacle OB to each grid element.
(1) Each grid element is pre-written with obstacle information indicating whether or not the obstacle OB is in the obstacle region R1. For example, when the grid element is the obstacle region R1 where the obstacle OB exists, the distance “0” (obstacle information) is written in the grid element and is not the obstacle region R1 where the obstacle OB exists. In this case, “−1” is written in the grid element.
(2) List the positions of grid elements in which the distance “0” is written.
(3) If “−1” is written for the upper, lower, left, and right grid elements of each grid element in which the distance “0” is written in (2) above, the distance “1” is written.
(4) List the positions of grid elements in which the distance “1” is written.
(5) Repeat the above steps (3) to (4) while increasing the distance information by 1 until there is no grid element to write the distance information.

In this way, the information corresponding to the distance from the obstacle OB can be written in all the grid elements without repeatedly performing the determination process for all the grid elements of the grid map as in the first embodiment. Reduction can be achieved.

[Embodiment 4]
An autonomous mobile device according to Embodiment 4 will be described. FIG. 10 is a diagram illustrating a state in which the virtual calculation of the distance to the obstacle OB is performed by the grid map GM2 according to the fourth embodiment. For convenience of explanation, members having the same functions as those described in the embodiment are given the same reference numerals, and description thereof is omitted.

In the fourth embodiment, how many times the collision determination need not be performed based on the information corresponding to the distance from the obstacle OB is processed as follows.
(1) The distance information written in the grid element at the position of a certain particle P is N1. (2) The collision determination is not performed for (N1 / 2) grid elements through which the virtual ray VL passes.
(3) The distance information written in the grid element that has arrived without making a collision determination in (2) above is N2.
(4) The collision determination is not performed for (N2 / 2) grid elements from which the virtual ray VL further passes.
(5) Repeat (3) to (4) above.

For example, in the example shown in FIG. 10, the distance “5” is written in the grid element of the particle P. Then, the collision determination is not performed for (N1 / 2 = 5/2 = 2) grid elements (but rounded down after the decimal point) through which the virtual ray VL passes. Next, the distance “N2 = 5” is written in the grid element of the arrival point P2 that has arrived without making a collision determination.

Thereafter, the collision determination is not performed for (N2 / 2 = 5/2 = 2) grid elements (but rounded down after the decimal point) through which the virtual ray VL further passes from the arrival point P2. In the above, the distance “N2 = 3” is written in the grid element of the arrival point P3 that has arrived without making a collision determination.

Then, (N2 / 2 = 3/2 = 1) grid elements (however, rounded down after the decimal point) where the virtual ray VL further passes from the arrival point P3 is not subjected to collision determination. In the above, the distance “N2 = 2” is written in the grid element of the arrival point P4 that has arrived without making a collision determination. Next, the calculation related to the collision determination process is performed from the arrival point P4. Accordingly, the collision determination process is unnecessary for the grid element between the particle P and the arrival point P4.

By performing the processing in this way, the collision determination of the straight line extended from the particle P is performed by a simpler method without performing the processing such as the trigonometric function that is complicated as in the second embodiment (FIG. 8). be able to.

[Example of software implementation]
The calculation unit 206 (particularly the grid map generation unit 3 and the self-position estimation unit 4) of the autonomous mobile device 1 may be realized by a logic circuit (hardware) formed in an integrated circuit (IC chip) or the like, or a CPU It may be realized by software using (Central Processing Unit).

In the latter case, the arithmetic unit 206 includes a CPU that executes instructions of a program that is software that implements each function, a ROM (Read Only Memory) in which the program and various data are recorded so as to be readable by a computer (or CPU), or A storage device (these are referred to as “recording media”), a RAM (Random Access Memory) for expanding the program, and the like are provided. And the objective of this invention is achieved when a computer (or CPU) reads the said program from the said recording medium and runs it. As the recording medium, a “non-temporary tangible medium” such as a tape, a disk, a card, a semiconductor memory, a programmable logic circuit, or the like can be used. The program may be supplied to the computer via an arbitrary transmission medium (such as a communication network or a broadcast wave) that can transmit the program. The present invention can also be realized in the form of a data signal embedded in a carrier wave in which the program is embodied by electronic transmission.

[Summary]
The autonomous mobile device 1 according to the first aspect of the present invention includes a sensor (laser range finder 2) that detects environmental information related to an obstacle OB, and an obstacle based on environmental information detected by the sensor (laser range finder 2). A generation unit (grid map generation unit 3) that generates grid maps GM1 to GM3 showing the object OB, and self-position candidate points on the grid maps GM1 to GM3 generated by the generation unit (grid map generation unit 3) An estimation unit (self-position estimation unit 4) that calculates a virtual distance from (particle P) to the obstacle OB and estimates its own position, and the grid maps GM1 to GM3 include a plurality of grid elements GE. Each grid element GE includes obstacle information representing the obstacle OB and an adjacent area R2 adjacent to the obstacle OB. And band information, and holds one of the remote area information representative of the remote regions R3 away from the obstacle OB.

According to the above configuration, since the grid map has the remote area information representing the remote area, when the virtual ray virtually passes through the remote area corresponding to the remote area information, the calculation related to the collision determination process Need not be implemented. As a result, it is possible to reduce the amount of calculation for virtually measuring the distance to the obstacle on the grid map in order to estimate the position of itself.

The autonomous mobile device 1 according to aspect 2 of the present invention is the above-described aspect 1, wherein the remote area R3 includes the first remote adjacent area R4 that is adjacent to the adjacent area R2 from the opposite side of the obstacle OB, and the first remote adjacent area R4. A second remote adjacent region R5 adjacent to the remote adjacent region R4 from the opposite side of the obstacle OB may be included.

According to the above configuration, the second remote adjacent area is farther from the obstacle OB than the first remote adjacent area. For this reason, information corresponding to the distance from the obstacle OB can be written in the grid elements corresponding to the first remote adjacent area and the second remote adjacent area.

In the autonomous mobile device 1 according to the aspect 3 of the present invention, in the aspect 1, the remote area information may include information corresponding to the number of other grid elements existing between the obstacle OB.

According to the above configuration, information corresponding to the distance from the obstacle OB can be written in the grid element corresponding to the remote area.

In the autonomous mobile device 1 according to the aspect 4 of the present invention, in the aspect 1, the environment information detected by the sensor (laser range finder 2) includes information indicating an actually measured distance to the obstacle OB, and the estimation unit (Self-position estimation unit 4) may calculate the virtual distance based on the obstacle information, the adjacent area information, and the remote area information held by each grid element.

According to the above configuration, since the calculation related to the collision determination process is not performed in the remote area corresponding to the remote area information, the calculation amount for virtually measuring the distance to the obstacle on the grid map is reduced. Can be reduced.

In the autonomous mobile device 1 according to the aspect 5 of the present invention, in the aspect 1, the estimation unit (self-position estimation unit 4) has the virtual ray VL that collides with the obstacle OB from the self-position candidate point (particle P). For each grid element that passes, the grid element and the obstacle that read out any one of the obstacle information held by each grid element, the adjacent area information, and the remote area information and read the adjacent area information Calculation for collision determination with the OB may be performed.

According to the above configuration, since the calculation for collision determination between the grid element and the obstacle OB is not performed on the grid element from which the remote area information is read, the distance between the obstacle and the grid map It is possible to reduce the amount of calculation for virtually measuring above.

In the autonomous mobile device 1 according to aspect 6 of the present invention, in the aspect 1, the sensor may be a laser range finder 2.

According to the above configuration, environmental information related to the obstacle can be detected with a simple configuration.

The calculation unit of the autonomous mobile device according to each aspect of the present invention may be realized by a computer. In this case, the calculation unit is realized by the computer by operating the computer as each unit included in the calculation unit. An arithmetic unit control program for an autonomous mobile device to be recorded and a computer-readable recording medium on which the program is recorded also fall within the scope of the present invention.

The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope shown in the claims, and embodiments obtained by appropriately combining technical means disclosed in different embodiments. Is also included in the technical scope of the present invention. Furthermore, a new technical feature can be formed by combining the technical means disclosed in each embodiment.

The present invention can be used for an autonomous mobile device such as a cleaning robot that can move autonomously.

1 Autonomous mobile device 2 Laser range finder (sensor)
3 Grid map generator (generator)
4 Self-position estimation unit (estimation unit)
OB Obstacle GM1 Grid map GM2 Grid map GM3 Grid map P Particle (Self-position candidate point)
P2 Arrival point P3 Arrival point P4 Arrival point GE Grid element R1 Obstacle area R2 Adjacent area R3 Remote area R4 First remote adjacent area R5 Second remote adjacent area RL Real light VL Virtual light Rd Actual distance Vd Virtual distance SP Start position

Claims

A sensor for detecting environmental information related to obstacles;
A generating unit that generates a grid map showing obstacles based on environmental information detected by the sensor;
An estimation unit that calculates a virtual distance from the self-position candidate point to the obstacle on the grid map generated by the generation unit, and estimates the position of the self,
The grid map includes a plurality of grid elements;
Each grid element holds any one of obstacle information representing the obstacle, adjacent area information representing an adjacent area adjacent to the obstacle, and remote area information representing a remote area away from the obstacle. An autonomous mobile device characterized by that.
The remote area includes a first remote adjacent area adjacent to the adjacent area from the opposite side of the obstacle, and a second remote adjacent area adjacent to the first remote adjacent area from the opposite side of the obstacle. Item 4. The autonomous mobile device according to Item 1.
2. The autonomous mobile device according to claim 1, wherein the remote area information includes information corresponding to the number of other grid elements existing between the obstacles.
The environmental information detected by the sensor includes information representing the measured distance to the obstacle,
The autonomous mobile device according to claim 1, wherein the estimation unit calculates the virtual distance based on the obstacle information, the adjacent area information, and the remote area information held by each grid element.
The estimation unit, for each grid element through which a virtual ray colliding with the obstacle from the self-position candidate point has passed, the obstacle information held by each grid element, the adjacent area information, the remote area information, The autonomous mobile device according to claim 1, wherein a calculation for determining a collision between the grid element from which the adjacent area information is read and the obstacle is performed.
The autonomous mobile device according to claim 1, wherein the sensor is a laser range finder.