JP2013225253A - Autonomous mobile device, autonomous mobile method, and program for autonomous mobile device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an autonomous mobile device capable of planning a route so that a movement load of each person and a movement load of an autonomous mobile device are minimized when it passes by a plurality of persons.SOLUTION: A route change planning part 103 plans one or more ways of course change for a plurality of persons. An attribute determination part 108 determines an attribute on mobility of a person. A load evaluation part 104 evaluates a total movement load of each person and an autonomous mobile device with respect to each way of course change on the basis of an attribute, a location and a speed as to mobility of each person. A target person determination part 105 selects a way of course change which exhibits a smallest total movement load, and determines persons performing and not performing course change. A movement route determination part 106 determines a moving route of the autonomous mobile device on the basis of a selected way of course change. A control part 107 controls the autonomous mobile device so that it travels on a determined moving route.

Description

  The present invention relates to an autonomous mobile device whose movement is controlled by a control device, an autonomous mobile method, and a program for the autonomous mobile device.

  In recent years, research and development of service robots that operate in a space where people exist has been conducted. For example, there are an article transport robot that carries an article on behalf of a person, a robot that patrols and guards a facility, or a robot that performs cleaning. There is a scene where a robot that works while moving as described above encounters a person while moving. At this time, the robot needs to pass the person safely.

  As a conventional technique, there is a technique for generating an avoidance path of a robot so that the robot avoids a moving body represented by a person (Non-Patent Document 1). In this prior art, a velocity vector (speed and direction) of a moving obstacle such as a moving object is detected from sensor information acquired by a sensor provided in the robot. To generate an avoidance path, assume that the moving obstacle maintains constant linear motion. The search for the avoidance route is set in a three-dimensional space-time space, and the moving obstacle is expressed as an oblique cylinder (see FIG. 32A). The shortest avoidance path (broken line type) that does not intersect with all the oblique cylinders and reaches the target point is derived, and this path is set as the avoidance path of the robot. FIG. 32B shows the path of the moving obstacle and the avoidance path of the robot when the search space is viewed from above.

JP 2009-110495 A JP-A-7-225612

Takashi Tsubouchi (Department of Electronic Information Engineering, University of Tsukuba), Tomohide Nanahana (Faculty of Engineering, The University of Tokyo), Taku Arimoto (Faculty of Engineering, The University of Tokyo) Journal of the Robotics Society of Japan 12 (7), pp.1029-1037, 1994

  Since the prior art assumes that the moving obstacle maintains constant linear motion, a situation in which a solution serving as an avoidance path cannot be obtained occurs. For example, as shown in FIG. 33, when two people (a wheelchair user and a healthy person) run or walk in parallel in the passage and occupy the passage, there is no space for the robot to pass through. In this case, the conventional technique cannot find a route that does not intersect the oblique cylinder of the moving obstacle. As an action that the robot takes when the avoidance route is not found, the robot stops in the middle of the passage until the state of the moving obstacle (person) changes and the avoidance route is found, for example. In this case, since the robot is stopped in the middle of the passage, the passage time of the robot is lost in addition to blocking the passage space of the person and preventing the passage of the person.

  As a countermeasure when the avoidance route is not found so as not to obstruct people's traffic, the robot stops at a fixed wall (for example, the wall on the left side of the passage) and stops until the person passes the wall on the opposite side Or proceed along the wall where the robot approaches. In the former, a time loss of the robot occurs. In the latter case, for example, if there is a person (for example, a wheelchair user) who cannot change the direction as soon as a healthy person is on the side where the robot is approaching, the person is caused to make a load (see FIG. 34). . It takes more time for healthy people to change direction. Even if the robot is close to the person, if the person is still changing direction, the robot may stop and the robot must wait until the person is on the other side .

  In contrast to the prior art in which it is assumed that the moving obstacle maintains a constant linear motion, Patent Document 1 and Patent Document 2 consider that a person changes the course. In Patent Document 1, based on the variation in position of one obstacle (person) over time, an area where the obstacle (person) is predicted to move when a predetermined time has elapsed is set as a virtual obstacle area, Control the robot to avoid that area. In Patent Document 2, since there is uncertainty in the speed sensed with the accuracy of the sensor, it is predicted that a moving object that moves at a constant speed may branch at a certain point in time. A probability distribution in each branch direction is obtained.

  However, in Patent Document 1 and Patent Document 2, the movement load when a person changes the course is not evaluated, and when the robot passes by a plurality of persons, how the course change is predicted for each person, It is not considered whether the movement load of each person and the robot is the smallest (the most efficient overall).

  An object of the present invention is to enable autonomous movement to plan a route so that the total total movement load of each person's movement load and the movement load of an autonomous mobile device when passing by a plurality of persons is minimized. An apparatus, an autonomous movement method, and a program for an autonomous mobile device are provided.

  In order to solve the above-described problems, an aspect of the present invention is configured as follows.

According to one aspect of the present invention, when the autonomous mobile device and a plurality of people pass each other, a route change planning unit that generates a route change for each person and plans one or more route change methods;
An attribute determination unit that determines an attribute relating to the ease of movement of the person;
The movement of the autonomous mobile device based on the position of each person relative to the autonomous mobile device, the speed of each person, and the ease of changing the direction according to the attribute relating to the ease of movement of the human determined by the attribute determining unit A load evaluation unit that calculates and evaluates a total mobile load that is the sum of the load and the mobile load of each person;
Based on the evaluation information in the load evaluation unit, the total of the total load of each person's movement load and the movement load of the autonomous mobile device is determined based on the one or more course change methods planned by the course change planning unit. A method for selecting a course change with the smallest moving load, and a target person determination unit for determining a person who changes the course and a person who does not change the course among the plurality of persons;
Based on the information determined by the subject determination unit, a movement route determination unit that determines a movement route of the autonomous mobile device,
An autonomous mobile device is provided that includes a travel control unit that controls the autonomous mobile device so that the autonomous mobile device travels along the travel route determined by the travel route determination unit.

  These general and specific aspects may be realized by a system, a method, a computer program, and any combination of the system, method, and computer program.

  When an autonomous mobile device passes by multiple people, the route is planned based on the position, speed, and attributes of each person so that the total movement load between each person and the autonomous mobile device is minimized. Therefore, the most efficient passing can be realized for a plurality of people as a whole.

It is a front view of the autonomous mobile device in a 1st embodiment. It is a side view of the autonomous mobile device in a 1st embodiment. It is a block diagram which shows the structure in an autonomous mobile apparatus. It is a flowchart which shows the procedure of operation | movement of an autonomous mobile apparatus. It is a figure which shows an environment map. It is a figure which shows the database in an environment map. It is explanatory drawing which shows a mode that the LRF sensor measured the driving | running | working environment and the obstruction. It is explanatory drawing which shows the data which the LRF sensor acquired. It is a figure which shows the example of the determination result of an attribute determination part in a table format. It is a figure which shows the processing flow in a course change plan part. It is a figure which shows the example of the scene where an autonomous mobile apparatus encounters two people. It is a figure which shows the example of the scene where an autonomous mobile apparatus encounters two persons with a table format. It is a figure which shows the example which excludes the course pattern of the person who is impossible by environmental restrictions. It is a figure which shows the example which omits the impossible person's course pattern by the conditions for passing. It is a figure which shows the example of the course pattern produced | generated with respect to the person. It is a figure which shows the example of the course pattern produced | generated with respect to the person. It is a figure which shows the example which produces | generates the combination of all the person's patterns in a table form. It is a figure which shows the example of the combination C1 of a course pattern. It is a figure which shows the example of the combination C4 of a course pattern. It is a figure which shows the example of the empty space produced when a person takes S pattern. It is a figure which shows the example of the empty space produced when a person takes R pattern. It is a figure which shows the example of the empty space produced when a person takes L pattern. It is a figure which shows the determination method of the presence or absence of an avoidance path | route in the way of a course change. It is a figure which shows the determination method of the presence or absence of an avoidance path | route in the way of a course change. It is a figure which shows the determination method of the presence or absence of an avoidance path | route in the way of a course change. It is a figure which shows the determination method of the presence or absence of an avoidance path | route in the way of a course change. It is a figure which shows the determination method of the presence or absence of an avoidance path | route in the way of a course change. It is a figure which shows the determination method of the presence or absence of an avoidance path | route in the way of a course change. It is a figure which shows the determination method of the presence or absence of an avoidance path | route in the way of a course change. It is a figure which shows the flow which subdivided the process of step S303. It is a figure which shows the example which calculates | requires the position of an autonomous mobile device and each person performed for every fixed time interval INT _A in the process of step S303. It is a figure which shows the experiment for calculating | requiring the delay time by the direction change for every attribute of a person. It is a figure which shows the delay time by the direction change in a healthy person. It is a figure which shows the delay time by the direction change in a wheelchair user. It is a figure which shows the delay time by the direction change in the healthy person whose driving speed is quicker than a wheelchair user. It is a figure which shows the example which performed the process of step S401. It is a figure which shows the example which performed the process of step S402. It is a figure which shows how to return the person who changes course after passing the autonomous mobile device. It is a figure which shows the example of the planned autonomous mobile device and the path | route of each person in the way of one course change. It is a figure which shows the example of the planned autonomous mobile device and the path | route of each person in the way of one course change. It is a figure which shows the method of calculating the delay time of the person who changes course. It is a figure which shows the example of the movement load in the method of each course change. It is a figure which shows the example of the scene where an autonomous mobile apparatus encounters two people. It is a figure which shows the example of the scene where an autonomous mobile apparatus encounters two persons with a table format. It is a figure which shows the example of the planned autonomous mobile device and the path | route of each person in the way of one course change. It is a figure which shows the example of the planned autonomous mobile device and the path | route of each person in the way of one course change. It is a figure which shows the example of the movement load in the method of each course change. It is a figure which shows the example of the conventional method when an avoidance path | route is not found with the prior art described in the nonpatent literature 1. It is a figure which shows the example of the path | route of an autonomous mobile apparatus and each person at the time of performing by the conventional method, and those movement loads. It is a figure which shows the example of the path | route of an autonomous mobile apparatus and each person at the time of performing by the conventional method, and those movement loads. It is a figure which shows the example of the scene where an autonomous mobile apparatus encounters two people. It is a figure which shows the example of the scene where an autonomous mobile apparatus encounters two persons with a table format. It is a figure which shows the example of the autonomous mobile device, each person's path | route, and those movement load at the time of performing by the conventional method. It is a figure which shows the example of the autonomous mobile device, each person's path | route, and those movement load at the time of performing by the conventional method. It is a figure which shows the outline | summary of the prior art described in the nonpatent literature 1. FIG. It is a figure which shows the example of the avoidance path | route which the robot produced | generated when the prior art described in the nonpatent literature 1 was applied. It is a figure which shows the example of the condition where an avoidance path | route is not found with the prior art described in the nonpatent literature 1. It is a figure which shows the problem which arises when the robot at the time of an avoidance path not being found with the prior art described in the nonpatent literature 1 advances along the fixed wall. It is a figure which shows the example of the scene where an autonomous mobile device encounters three people.

  Embodiments according to the present invention will be described below in detail with reference to the drawings.

  DESCRIPTION OF EMBODIMENTS Hereinafter, various embodiments of the present invention will be described before detailed description of embodiments of the present invention with reference to the drawings.

According to the first aspect of the present invention, when the autonomous mobile device and a plurality of people pass each other, a course change planning unit that generates a course change of each person and plans one or more course change methods;
An attribute determination unit that determines an attribute relating to the ease of movement of the person;
The movement of the autonomous mobile device based on the position of each person relative to the autonomous mobile device, the speed of each person, and the ease of changing the direction according to the attribute relating to the ease of movement of the human determined by the attribute determining unit A load evaluation unit that calculates and evaluates a total mobile load that is the sum of the load and the mobile load of each person;
Based on the evaluation information in the load evaluation unit, the total of the total load of each person's movement load and the movement load of the autonomous mobile device is determined based on the one or more course change methods planned by the course change planning unit. A method for selecting a course change with the smallest moving load, and a target person determination unit for determining a person who changes the course and a person who does not change the course among the plurality of persons;
Based on the information determined by the subject determination unit, a movement route determination unit that determines a movement route of the autonomous mobile device,
An autonomous mobile device is provided that includes a travel control unit that controls the autonomous mobile device so that the autonomous mobile device travels along the travel route determined by the travel route determination unit.

According to the second aspect of the present invention, further comprising a human detection unit for detecting a person passing by the autonomous mobile device,
The course change planning unit creates a pattern for changing the course and a pattern that does not change the course for each of a plurality of persons detected by the person detection unit, and all the people detected by the person detection unit The autonomous mobile device according to the first aspect of planning a combination of patterns in the above is provided.

  According to the third aspect of the present invention, the route change planning unit represents the person and the autonomous mobile device as oblique cylinders, performs route search in space-time, and changes the route in each route change method. The movement path of the autonomous mobile device is generated with respect to the inclined cylinder of the person who does not, and then the movement route of the person who changes the course with respect to the inclined cylinder of the autonomous mobile device is generated according to the first or second aspect. An autonomous mobile device is provided.

  According to the fourth aspect of the present invention, the load evaluation unit includes a delay time as an extra time required when each person changes the course with respect to the case where each person does not change the course, and the autonomous The first to third evaluations for evaluating the total movement load with a total delay time that is a sum of a delay time as an extra time required when the autonomous mobile device changes the route with respect to the case where the mobile device does not change the route. An autonomous mobile device according to any one aspect is provided.

According to a fifth aspect of the present invention, the apparatus further comprises a person detection unit that detects a person passing by the autonomous mobile device and the speed of the person,
The load evaluation unit is configured to evaluate the movement load of the person by changing the ease of the person's direction change based on the speed of the person detected by the person detection unit in one person attribute. The autonomous mobile apparatus as described in any one aspect of -4 is provided.

According to the sixth aspect of the present invention, further comprising a human detection unit for detecting a person passing by the autonomous mobile device,
The said load evaluation part provides the autonomous mobile apparatus as described in any one of the 1st-5th aspect which evaluates the said person's movement load with the number of the persons who are detected by the said person detection part and change course.

According to the seventh aspect of the present invention, the route change planning unit generates a route change for each person when the autonomous mobile device and a plurality of people pass each other, and plans one or more ways to change the route. ,
The attribute determination unit determines the attribute related to the ease of movement of the person,
Based on the position of each person relative to the autonomous mobile device, the speed of each person, and the ease of changing the direction according to the attribute relating to the ease of movement of the person determined by the attribute determination unit by the load evaluation unit, Evaluate the total mobile load, which is the sum of the mobile mobile device's mobile load and each person's mobile load,
Based on the evaluation information in the load evaluation unit, the subject person determination unit determines the movement load of each person and the movement load of the autonomous mobile device from the one or more course change methods planned by the course change planning unit. Selecting the course change method with the smallest total movement load of the above and determining the person who changes the course and the person who does not change the course among the plurality of persons,
Based on the information determined by the subject determination unit by the travel route determination unit, determine the travel route of the autonomous mobile device,
The travel control unit controls the autonomous mobile device so that the autonomous mobile device travels along the travel route determined by the travel route determination unit.
Provide an autonomous movement method.

According to an eighth aspect of the present invention, there is a program for an autonomous mobile device for controlling an autonomous mobile device that moves by autonomously determining a movement route,
Computer
A route change planning unit that generates a route change of each person when the autonomous mobile device and a plurality of people pass each other, and plans one or more ways of changing the route;
An attribute determination unit that determines an attribute relating to the ease of movement of the person;
The movement of the autonomous mobile device based on the position of each person relative to the autonomous mobile device, the speed of each person, and the ease of changing the direction according to the attribute relating to the ease of movement of the human determined by the attribute determining unit A load evaluation unit that calculates and evaluates a total mobile load that is the sum of the load and the mobile load of each person;
Based on the evaluation information in the load evaluation unit, the total of the total load of each person's movement load and the movement load of the autonomous mobile device is determined based on the one or more course change methods planned by the course change planning unit. A method for selecting a course change with the smallest moving load, and a target person determination unit for determining a person who changes the course and a person who does not change the course among the plurality of persons,
Based on the information determined by the subject determination unit, a movement route determination unit that determines a movement route of the autonomous mobile device,
There is provided an autonomous mobile device program for causing a function of the travel control unit to control the autonomous mobile device so that the autonomous mobile device travels along the travel route determined by the travel route determination unit.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

(First embodiment)
FIG.1 and FIG.2 shows the external appearance of the autonomous mobile device 1 in 1st Embodiment of this invention. As shown in FIGS. 1 and 2, a cylindrical autonomous mobile device (an example of a robot) 1 includes a left wheel 12a, a right wheel 12b, a front wheel 12c, and a rear wheel 12d that are independent of each other below the main body 1a. Thus, the robot is configured to be autonomously movable by being arranged so as to be rotatable forward and backward. Such an autonomous mobile device 1 includes an LRF (Laser Range Finder) sensor 2 as an example of an environment observation unit in front of the upper portion of the main body 1a, and a distance image sensor 3 on the upper surface of the main body 1a. Furthermore, the main body 1a includes an autonomous mobile device control unit 1C, and the autonomous mobile device control unit 1C autonomously determines a movement route and controls the movement.

  The left wheel 12a and the right wheel 12b are connected to the left and right motors 12aM and 12bM, respectively, and the left and right motors 12aM and 12bM are independently rotated forward and reversely under the control of the autonomous mobile device control unit 1C. The autonomous mobile device 1 travels forward and backward through the passage 6. By changing the rotational speed of the left wheel 12a and the right wheel 12b, the autonomous mobile device 1 can be bent in the left-right direction with respect to the front-rear direction.

  Details of the LRF sensor 2 and the distance image sensor 3 will be described later.

  FIG. 3 shows a block diagram of the autonomous mobile device 1 in the first embodiment.

  The autonomous mobile device 1 includes an autonomous mobile device information acquisition unit 111, a human detection unit (human information acquisition unit) 101, a determination unit 102, a course change planning unit 103, and a load evaluation unit as the autonomous mobile device control unit 1C. 104, a target person determination unit 105, a movement route determination unit 106, a travel control unit 107, an attribute determination unit 108, an avoidance route planning unit 109, and a normal route planning unit 112. In addition to this, the autonomous mobile device 1 includes an environment map storage unit 110. However, the information stored in the environment map storage unit 110 is transferred to the environment information acquisition unit 110G via an input / output interface and a communication line. It may be acquired from a database outside the autonomous mobile device 1 and stored in the environment map storage unit 110.

  The autonomous mobile device information acquisition unit 111 acquires the position and speed of the autonomous mobile device 1. That is, the autonomous mobile device information acquisition unit 111 includes the self-position estimation information of the autonomous mobile device 1 from the human detection unit 101, the environment map 51 stored in the environment map storage unit 110, the left wheel 12a, and the right wheel 12b. Based on the information from the encoders 12aE and 12bE of the left and right motors 12aM and 12bM, the position and speed of the autonomous mobile device 1 in the environment in which the autonomous mobile device 1 is traveling are acquired.

  In the autonomous mobile device information acquisition unit 111, in order to acquire the speed of the autonomous mobile device 1, odometry information possessed by the autonomous mobile device 1 is used. That is, in the autonomous mobile device information acquisition unit 111, based on the information from the encoders 12aE and 12bE of the left and right motors 12aM and 12bM of the left wheel 12a and the right wheel 12b, the rotation angles of the left wheel 12a and the right wheel 12b The movement distance and direction are obtained by integrating the rotation angular velocity.

  The environment map storage unit 110 records in advance an environment map 51 including the shape of the environment in which the autonomous mobile device 1 moves (such as the position of the wall 51w). The environment map 51 stores a path on which the autonomous mobile device 1 travels and obstacles that exist in the environment on which the autonomous mobile device 1 travels. That is, the environment map storage unit 110 stores in advance an environment map 51 of a place where the autonomous mobile device 1 and the person 5 coexist and the autonomous mobile device 1 moves, and the autonomous mobile device information acquisition unit 111 and the human detection unit 101. The data of the environment map 51 is provided. The obstacles in the environment map 51 can be classified into two types: a moving obstacle (for example, a person 5) and a fixed obstacle (for example, a wall 51w). The coordinates of the wall 51w that is a fixed obstacle and forms the passage 6 are used. Is included in the environment map 51.

  FIG. 5 shows an example of the environment map 51. In order to store in the environment map storage unit 110, in the environment map 51, for example, each corner 51f of the wall 51w (51a, 51b, 51c, 51d, 51e, 51g, 51h, 51i) is represented by points P1, P2, P3. , P4. The coordinates of a position such as a corner 51f such as the walls 51a, 51b, 51c, 51d, 51e, 51g, 51h, 51i are determined, and the walls 51a, 51b, 51c, 51d, 51e, 51g, 51h, 51i are divided into two. Represented by a line connecting two points. For example, in FIG. 5, the wall 5g is indicated by a line connecting P2 and P3.

  FIG. 6 shows an example of the database in the environment map shown above. In FIG. 6, xy coordinates of the points P1, P2, P3, and P4 and information such as a line connecting the points P2 and P3 are recorded.

  The normal route planning unit 112 is based on information on the autonomous mobile device 1 and the passage 6, that is, information from the determination unit 102 and the autonomous mobile device information acquisition unit 111, and the autonomous mobile device 1 when the person 5 is not detected. Plan your route. A route is planned from the current position of the autonomous mobile device 1 toward a predetermined destination, and the planned route information is output to the travel control unit 108.

  The person detection unit 101 detects the person 5 and acquires information about the person 5. When the person 5 is not detected, information for operating the normal route planning unit 112 is output to the normal route planning unit 112. On the other hand, when the person 5 is detected, information for operating the determination unit 102 is output to the determination unit 102.

  In the human detection unit 101, in order to obtain the position and speed of the person (person who passes the autonomous mobile apparatus 1) 5 in the environment where the autonomous mobile apparatus 1 is traveling, in the first embodiment, FIG. 2 information from the LRF sensor 2 and information from the environment map storage unit 110 and the timer 101T are used. The information of the LRF sensor 2 is collated with the information of the environment map 51 in the environment map storage unit 110 by the human detection unit 101, and the position of the autonomous phase device 1 in the absolute coordinate system 54 is obtained by the human detection unit 101 (see FIG. 5). reference). Details are described below.

  As shown in FIG. 7, when the autonomous mobile device 1 is traveling, the LRF sensor 2 irradiates a plurality of laser beams 11 around the autonomous mobile device 1 at predetermined time intervals, and the surrounding obstacles. And measure the distance. In the first embodiment, the LRF sensor 2 has a measurable range of 270 degrees as an example, and irradiates the laser beam 11 at an interval of 0.25 degrees as an example. Each of the laser lines 11 at each angle measures the distance to the obstacle at that angle. That is, reference numeral 80 in FIGS. 7 and 8 indicates a spot where the laser line 11 hits a known obstacle (in this example, the wall 51w), and reference numeral 81 indicates a spot where the laser line 11 hits the person 5. . The distance from the irradiation port of the laser beam 11 to these spots is measured by the LRF sensor 2 as the distance from the surrounding obstacle. This measurement result is the environmental information. FIG. 8 shows data acquired by the LRF sensor 2.

  One LRF sensor 2 as an example of an environment observation unit observes environment information around the autonomous mobile device 1 in advance or every predetermined time, for example, based on information from the timer 101T, and an environment map storage unit 110. Here, the environment information includes two types of information of a moving obstacle (for example, a person 5) and a fixed obstacle (for example, a wall 51w) around the autonomous mobile device 1, but the environment map 51 includes a fixed obstacle. Includes only objects (for example, wall 51w). In addition, as an environment observation part, only one sensor may be sufficient, and in order to improve measurement accuracy, you may use two or more sensors as sensor fusion. An example of the environment observation unit is not limited to the LRF sensor 2, and a sensor such as a millimeter wave sensor, an ultrasonic sensor, or a stereo camera can also be used.

  FIG. 8 shows data acquired by the LRF sensor 2. The human detection unit 101 can acquire the shape of the environment around the current location of the autonomous mobile device 1 from the data acquired by the LRF sensor 2. The environment shape acquired by the LRF sensor 2 is collated by the human detection unit 101 with the environment map 51 recorded in advance in the environment map storage unit 110 as the shape of the environment in which the autonomous mobile device 1 operates.

  FIG. 5 shows the environment map 51. In order to store the information in the environment map 51 in the database, the coordinates of a place such as a corner of the wall are obtained. For example, in FIG. 5, the points P1, P2, P3, and P4 are shown. The wall is represented by a line connecting two points. For example, in FIG. 5, the line connecting P2 and P3 is shown. An example of the database in the environment map 51 is shown in FIG.

  The human detection unit 101 repeatedly translates and rotates the environment map 51 with respect to the shape of the environment acquired by the LRF sensor 2, and the human detection unit 101 determines the most matching place on the environment map 51. look for. The most matching place is a position in the traveling environment of the autonomous mobile device 1.

  Furthermore, the person detection unit 101 detects the person 5 around the autonomous mobile device 1 based on information from the LRF sensor 2 and acquires the position, speed, and traveling direction of the person 5. In the first embodiment, the LRF sensor 2 of FIGS. 1 and 2 is used as an example of the environment observation unit as an information source input to the human detection unit 101. In addition, as an environment observation part as an information source with respect to the human detection part 101, only one sensor may be sufficient, and in order to improve a measurement precision, two or more sensors may be used as a sensor fusion. In addition, the environmental observation person can also use a sensor such as a distance image sensor, a millimeter wave sensor, an ultrasonic sensor, or a stereo camera.

  A method for detecting the person 5 by the person detection unit 101 using the LRF sensor 2 of FIGS. 1 and 2 will be described.

  Based on the information acquired by the LRF sensor 2, the travel position of the autonomous mobile device 1 in the environment in which the autonomous mobile device 1 is traveling is estimated by the human detection unit 101 as described above. This estimated information is output from the human detection unit 101 to the autonomous mobile device information acquisition unit 111. When the position of the autonomous mobile device 1 is estimated by the human detection unit 101, distance data that does not partially match the environment map 51 is generated from the plurality of laser lines 11.

  As shown in FIG. 7, there is a laser line 11 that hits a person 5 and appears as a point group 81 of unknown obstacles that are not in the environment map 51 as shown in FIG. As the autonomous mobile device 1 travels, the human detection unit 101 determines whether the point group 81 of the unknown obstacle is always at the same position in the absolute coordinate system 54.

  When the point block 81 of the unknown obstacle is always at the same position, the human detection unit 101 determines that it is a fixed obstacle. Otherwise, the human detection is performed that the point group 81 of the unknown obstacle is a moving obstacle. Judgment is made by the unit 101. After determining that the moving obstacle is the person 5, the human detection unit 101 considers the width of the moving obstacle in order to determine whether or not the moving obstacle is the person 5. The human detection unit 101 estimates the width of the moving obstacle from the distance data of the LRF sensor 2. As an example, it is assumed that the wheelchair user 5b has a width of about 80 cm and the healthy person 5a has a width of about 30 cm as viewed from the side. Then, by referring to the stored information and the moving obstacle detected by the LRF sensor 2 with the human detection unit 101, the person 5 and the human detection unit 101 can change the one having a width of 30 to 80 cm. Estimate (judge).

  When the person 5 is detected by the person detection unit 101, the person detection unit 101 can acquire the position, speed, and traveling direction of the person 5 as information about the person 5. From the position of the autonomous mobile device 1 and the distance from the person 5 measured by the LRF sensor 2, the position of the person 5 is estimated by the person detection unit 101. The person detection unit 101 obtains the speed of the person 5 as a time derivative of the position of the person 5. A vector at the speed of the person 5 is acquired by the person detection unit 101 as the traveling direction of the person 5.

  Based on the information such as the position of the person 5 from the person detection unit 101 and the information from the distance image sensor 3, the attribute determination unit 108 is based on the movement ability of the person 5 (ease of movement when the person 5 moves). An attribute (that is, an attribute related to human mobility) is determined. The attribute (person movement attribute) based on the movement ability is an attribute (information regarding the direction change of the person 5) due to the difference in ease of movement for each movement direction (traveling direction) of the person 5. For example, as attributes based on movement ability (person movement attributes), there are a healthy person 5a who can immediately move in all directions, and a wheelchair user 5b which is difficult to change direction and takes time. In addition, the movement attribute of the person 5 includes, for example, a crutch user, a person walking with an infusion device, a walking assist device user, an operator walking while moving a hand-held cart. FIG. 9 shows an example of attributes based on the movement ability of the person 5.

  In the first embodiment, the distance image sensor 3 is used as an example of the person movement attribute acquisition unit of the attribute determination unit 108. For example, using the method described in Non-Patent Document 2, the attribute determination unit 108 determines whether the detected person is a healthy person or a wheelchair user. In Non-Patent Document 2, a feature amount (for example, the size of a person's region, the highest point of a person, the center of gravity, a circumscribed cuboid, etc.) is extracted from the distance image obtained by the distance image sensor 3 using the attribute determining unit 108. The attribute determination unit 108 classifies a person's posture using a support vector machine (SVM).

Norihiro Takahashi (Nara Institute of Science and Technology) Graduate School of Information Science and Technology, Advanced Science and Technology Graduate School, Naokazu Yokoya (Graduate School of Information Science, Nara Institute of Science and Technology) “Personal Classification by SVM from Overhead Image Using Distance Image Sensor” IEICE Technical Research report. PRMU, pattern recognition / media understanding 107 (384), 47-52, 2007-12-06 Based on information on posture classification of people performed in this way, whether person 5 is a healthy person or a wheelchair user, etc. Is determined by the attribute determination unit 108.

  The attribute determination unit 108 can also use a stereo camera.

  The determination unit 102 determines whether or not the autonomous mobile device 1 needs to take a safety measure. If the safety measure does not need to be taken, information from the human detection unit 101 and the autonomous mobile device information acquisition unit 111 is obtained. Based on this, the presence / absence of an avoidance route of the autonomous mobile device 1 for passing by a plurality of people 5 is determined.

In order to determine whether the autonomous mobile device 1 needs to take safety measures, the distance between the autonomous mobile device 1 detected by the LRF sensor 2 and obtained through the human detection unit 101 and the person 5 is viewed. If the determination unit 102 determines that the distance between the autonomous mobile device 1 and the person 5 is equal to or less than a certain threshold value R _stop , the determination unit 102 determines that the autonomous mobile device needs to take safety measures. The safety measures taken by the autonomous mobile device 1 mean, for example, that the autonomous mobile device 1 is urgently stopped, or that the autonomous mobile device 1 moves backward to a certain distance. As an example, when the radius of the person 5 is 0.3 m and a safety distance of 0.2 m is given, the threshold value R _{stop is set} to 0.8 m in the first embodiment.

If the determination unit 102 determines that the distance between the autonomous mobile device 1 and the person 5 is not equal to or less than a certain threshold R _stop , it means that the autonomous mobile device 1 does not need to take safety measures. 102 determines whether there is an avoidance route. As a method for determining whether or not there is an avoidance route in this case, first, it is assumed that the person 5 and the autonomous mobile device 1 maintain a uniform linear motion. The search for the avoidance route is set in a three-dimensional space-time, and the person 5 and the autonomous mobile device 1 are expressed as an oblique cylinder including the moving direction. Then, based on the diameter of the cylinder including the safety area of the autonomous mobile device 1, the diameter of the cylinder including the safety area of the person 5, and the passage width stored in the environment map storage unit 110, the autonomous mobile device 1 The presence / absence of a passage width of the autonomous mobile device 1 is determined by calculating the presence / absence of a passing width (see Non-Patent Document 1). When the determination unit 102 determines that there is an avoidance route, the determination unit 102 sends information to the avoidance route planning unit 109, such as information regarding the autonomous mobile device 1, the person 5, the passage 6, and the like, that is, the determination unit 102, the attribute determination unit 108, Information from the autonomous mobile device information acquisition unit 111 is output. When the determination unit 102 determines that there is no avoidance route, the determination unit 102 sends information about the autonomous mobile device 1, the person 5, the passage 6, etc., that is, the determination unit 102, the attribute determination unit 108, Information from the autonomous mobile device information acquisition unit 111 is output.

  The avoidance route planning unit 109, when the determination unit 102 determines that there is an avoidance route, based on the information from the determination unit 102, the avoidance route planning unit 109 along the route that can secure the width for passage of the autonomous mobile device 1 1 avoidance route is planned (see Non-Patent Document 1). The avoidance route planned by the avoidance route planning unit 109 is output to the travel control unit 107.

  When the determination unit 102 determines that there is no avoidance route, the route change planning unit 103 is based on information from the determination unit 102, the attribute determination unit 108, and the autonomous mobile device information acquisition unit 111. 5a, 5b,...) For each person 5 and a plan for one or more course changes. In each course change method, the course change planning unit 103 plans each route between the autonomous mobile device 1 and each person 5 in the course change method based on the position, speed, and attribute of each person 5. .

  FIG. 10 shows a process flow of the course change planning operation in the course change planning unit 103.

  11A and 11B show a scene in which the autonomous mobile device 1 encounters two people 5a and 5b in the passage 6 sandwiched between the walls 51w1 and 51w2 on both sides. The first person 5a and the second person 5b are referred to as a first person 5a and a second person 5b, respectively. The first person 5a is a healthy person, and the second person 5b is a wheelchair user. The processing flow of the course change planning unit 103 shown in FIG. 10 will be specifically described using the scenes shown in FIGS. 11A and 11B.

  First, a course pattern of each person 5a, 5b is prepared by the course change planning unit 103 (step S301).

  Specifically, the course change planning unit 103 classifies the course pattern of the person 5 into the following three.

      -The course is not changed (S pattern).

      ・ Change the course to the right (R pattern).

      -Change the course to the left (L pattern).

  When the person 5 changes the course to the right, the person 5 changes the course to the right side of the autonomous mobile device 1. Similarly, when the person 5 changes the course to the left, the person 5 changes the course to the left side of the autonomous mobile device 1.

  In the three route patterns, the impossible route patterns are omitted depending on the environmental restrictions and the conditions for passing each other.

  Here, the environment restriction is an object that restricts movement, such as the wall 51w of the passage 6 in the environment where the person 5 is traveling. FIG. 12A shows an example in which the course pattern of the person 5 that is impossible due to environmental restrictions is omitted. In FIG. 12A, since the person 5a is traveling along the right wall 51w1, the R pattern cannot be moved, and is omitted from the course pattern that the person 5a can take.

  The condition for passing is a condition regarding whether or not the autonomous mobile device 1 can pass when the person 5 takes a certain route pattern. FIG. 12B shows an example in which the course pattern of the person 5 that is impossible due to the conditions for passing each other is omitted. In FIG. 12B, the person 5 travels in the middle of the passage 6. When the person 5 passes through the S pattern, the autonomous mobile device 1 cannot pass, and thus the route pattern that the person 5 can take is omitted.

  In the course change planning unit 103, the process of step S301 is applied to the scene shown in FIGS. 11A and 11B.

  FIG. 13A shows the course pattern of the first person 5a generated by the course change planning unit 103 when the process of step S301 is applied to the first person 5a by the course change planning unit 103. Since the first person 5a is traveling along the right wall 51w1, there are two course patterns that can be taken by the first person 5a: an L pattern and an S pattern.

  FIG. 13B shows the course pattern of the second person 5b generated by the course change planning unit 103 when the process of step S301 is applied to the second person 5b by the course change planning unit 103. Since the first person 5a is traveling along the left wall 51w2, there are two course patterns that can be taken by the second person 5b: an R pattern and an S pattern.

  After preparing a course pattern for each person 5a, 5b in the course change planning unit 103, the course change planning unit 103 generates a combination of patterns of all the people 5a, 5b (step S302).

  FIG. 14A, FIG. 14B, and FIG. 14C show the case where the process of step S302 is applied to the scene shown in FIG. 11A and FIG. 11B. The route change planning unit 103 generates four route pattern combinations C1 to C4 for the first person 5a and the second person 5b.

  Here, the route change planning unit 103 omits combinations where the avoidance route of the autonomous mobile device 1 is not found. In order for the route change planning unit 103 to determine whether an avoidance route is found, the route change planning unit 103 uses the concept of an empty space for the autonomous mobile device 1 to pass through. The concept will be described below.

Each agent (autonomous mobile device 1 or person 5) travels at a certain safe distance from the wall 51w when traveling along the passage 6. Each agent 1, 5 travels at a certain safe distance from other agents 5, 1 even when passing by other agents 5, 1. _Let δ _wall and δ _{between_agent} be the minimum safe distance to the wall 51w and the minimum safe distance to the other agents 5 and 1, respectively. Let R _r and R _h be the radius of the autonomous mobile device 1 and the radius of the person 5 when two-dimensionally approximated by a circle, respectively. The minimum required free space _{dr_min} for the autonomous mobile device 1 to pass through can be defined by the following formula (1).

・・・・・・式（１）

・ ・ ・ ・ ・ ・ Formula (1)

In the first embodiment, as an example, the minimum safe distances δ _wall , δ _{between_agent} , and radii R _r and R _h are 0.1 m, 0.2 m, 0.3 m, and 0.3 m, respectively. Therefore, the minimum required free space _{dr_min} is 0.9 m.

An empty space generated when the person 5 takes each course pattern will be described. The width of the passage 6 D, left wall 51w2 and right wall 51w1, respectively _{x Leftwall} the x-coordinate of the _{x Rightwall.} x _h (t ₀ ) and x _h (t) are the x coordinate at the current position of the person 5 and the x coordinate at the position where the person 5 may exist depending on the route pattern taken. In the first embodiment, as an example, the x coordinates x _leftwall and x _rightwall of the left wall 51w2 and the right wall 51w1 are set to −0.90 m and 0.90 m. 11A and 11B, the positions of the first person 5a and the second person 5b shown in the table of FIG. 11B are the current positions x _h1 (t ₀ ) and x _h2 ( t ₀ ).

FIG. 15A shows an empty space that occurs when the person 5 takes the S pattern. Since the person 5 does not change the course, the x-coordinate x _h (t) at a position where the person 5 may exist depending on the taken course pattern is expressed by the following formula (2).

・・・・・・式（２）

・ ・ ・ ・ ・ ・ Formula (2)

The maximum empty space d _{h_max} generated by the x coordinate x _h (t) is expressed by the following equation (3).

・・・・・・式（３）

・ ・ ・ ・ ・ ・ Formula (3)

FIG. 15B shows the empty space that occurs when the person 5 takes the R pattern. Since the person 5 changes the course to the right side of the autonomous mobile device 1, it is assumed that the autonomous mobile device 1 travels along the left wall 51w2, and the x coordinate x _h (t) is expressed by the following equation (4). .

・・・・・・式（４）

・ ・ ・ ・ ・ ・ Formula (4)

A space d _h vacant on the left side of the person 5 generated by the x coordinate x _h (t) is expressed by the following formula (5).

通路６の幅Ｄ、左壁５１ｗ２と右壁５１ｗ１のｘ座標ｘ_{ｌｅｆｔｗａｌｌ}、ｘ_{ｒｉｇｈｔｗａｌｌ}では、下記式（６）の関係式があるため、

Width D of the passage 6, the left wall 51w2 and right wall 51w1 of x coordinate _{x Leftwall,} the _{x Rightwall,} since there is a relation of the following formula (6),

・・・・・・式（６）

・ ・ ・ ・ ・ ・ Formula (6)

  The above equation (6) can be rewritten as the following equation (7).

・・・・・・式（７）

・ ・ ・ ・ ・ ・ Formula (7)

The maximum vacant space _{dh_max} is _expressed by the following formula (8).

・・・・・・式（８）

・ ・ ・ ・ ・ ・ Formula (8)

FIG. 15C shows the empty space that occurs when the person 5 takes the L pattern. Since the person 5 changes the course to the left side of the autonomous mobile device 1, it is assumed that the autonomous mobile device 1 travels along the right wall 51w1, and the x coordinate x _h (t) is expressed by the following equation (9). .

・・・・・・式（９）

・ ・ ・ ・ ・ ・ Formula (9)

A space d _h vacant on the right side of the person 5 by the x coordinate x _h (t) is expressed by the following formula (10).

・・・・・・式（１０）

・ ・ ・ ・ ・ ・ Formula (10)

  The above equation (10) can be rewritten as the following equation (11).

・・・・・・式（１１）

・ ・ ・ ・ ・ ・ Formula (11)

The maximum vacant space _{dh_max} is _expressed by the following equation (12).

・・・・・・式（１２）

・ ・ ・ ・ ・ ・ Formula (12)

The maximum free space _{dh_max} route change planning unit 103 calculates the combination of patterns in a plurality of people 5 and verifies whether or not the following formula (13) is satisfied by the route change planning unit 103.

・・・・・・式（１３）

・ ・ ・ ・ ・ ・ Formula (13)

When the above formula (13) is satisfied, it means that the route change planning unit 103 has an empty space for the autonomous mobile device 1 to pass through and an avoidance route is found. Otherwise, the route change planning unit 103 determines that no avoidance route is found.

  A method of determining whether there is an avoidance route for each combination shown in FIGS. 14A, 14B, and 14C in the route change planning unit 103 will be described. FIG. 14B and FIG. 14C are diagrams showing examples of route pattern combinations C1 and C4, respectively.

FIG. 16A shows how to determine the combination C <b> 1 in the course change planning unit 103. Since the first person 5a is an S pattern and the second person 5b is an S pattern, three empty spaces d _h 1 are formed between the left wall 51w2, the first person 5a, the second person 5b, and the right wall 51w1. , D _h 2 and d _h 3 are generated. Here, as an example, if the radius of the first person 5a and the second person 5b is 0.3 m, the empty spaces d _h 1, d _h 2 and d _h 3 are 0 in the scenes shown in FIGS. 11A and 11B, respectively. .15m, 0.30m, and 0.15m. Therefore, the maximum free space _{dh_max} is 0.30 m. Since the maximum free space _{dh_max} is smaller than the minimum required free space _{dr_min} , the route change planning unit 103 determines that an avoidance route is not found in the combination C1.

FIG. 16B and FIG. 16C show the determination method in the combination C2. The second person 5b has an R pattern, and a range x _h2 (t) illustrated in FIG. 16B represents an x coordinate at a position where the second person 5b may exist by the R pattern. FIG. 16C shows a case where the maximum free space _{dh_max} occurs. In the scene shown in FIGS. 11A and 11B, the maximum empty space _{dh_max} is 1.05 m. Since the maximum free space _{dh_max} is larger than the minimum required free space _{dr_min} , the route change planning unit 103 determines that an avoidance route is found in the combination C2.

FIG. 16D and FIG. 16E show the determination method in the combination C3. The first person 5a has an L pattern, and a range x _h1 (t) illustrated in FIG. 16D represents an x coordinate at a position where the first person 5a may exist due to the L pattern. FIG. 16E shows a case where the maximum free space _{dh_max} occurs. In the scene shown in FIGS. 11A and 11B, the maximum empty space _{dh_max} is 1.05 m. Since the maximum free space _{dh_max} is larger than the minimum required free space _{dr_min} , the route change planning unit 103 determines that an avoidance route is found in the combination C3.

FIG. 16F and FIG. 16G show the determination method in the combination C4. The first person 5a and the second person 5b have an L pattern and an R pattern, respectively. The ranges x _h1 (t) and x _h2 (t) shown in FIG. Represents the x-coordinate at a potential location. FIG. 16G shows a case where the maximum free space _{dh_max} occurs. In the scene shown in FIGS. 11A and 11B, the maximum empty space _{dh_max} is 0.40 m. Since the maximum free space _{dh_max} is smaller than the minimum required free space _{dr_min} , the route change planning unit 103 determines that an avoidance route is not found in the combination C4.

  The course change planning unit 103 omits the combinations C1 and C4 and leaves the combinations C2 and C3 based on the above determination.

  Next, the route change planning unit 103 generates a route between the autonomous mobile device 1 and each person 5a, 5b in each combination (step S303). In the process of step S303, the course change planning unit 103 further divides the process here into two using the concept of the oblique cylinder described in Non-Patent Document 1 (step S401 and step S402 in FIG. 17). See).

  First, the route change planning unit 103 generates a route of the autonomous mobile device 1 for an oblique cylinder of a person who does not change the route (S pattern) (step S401).

  Next, a route of the person (L pattern, R pattern) whose course is changed with respect to the oblique cylinder of the autonomous mobile device 1 is generated by the course change planning unit 103 (step S402).

Performs the process of step S303 to a diversion planning section 103 every predetermined time interval INT _A, the autonomous mobile apparatus 1 and each person 5a of each fixed time interval INT _A, determine the position of 5b in diversion planning unit 103 (FIG. 18 See). In the first embodiment, the constant time interval INT _{A is set} to 0.2 s.

  When the routes of the persons 5a and 5b are generated in step S303, the time taken for the persons 5a and 5b to change direction is also taken into consideration. Since the time required for the direction change is different for each person's movement attribute, the direction change plan unit 103 reproduces the direction change for each attribute based on the attribute of each person 5a, 5b obtained by the attribute determination unit 108. Is planned by the route change planning unit 103.

An experiment shown in FIG. 19A was performed in advance in order to examine the characteristics of time required for the direction change for each attribute by the course change planning unit 103. There are two types of subjects: healthy people and wheelchair users. As shown in FIG. 19A, the subject walks so as to travel straight toward the point A, and when the point A is reached, the direction is changed at three kinds of angles (0 degrees, 45 degrees, 90 degrees), I had to walk from point a to 2m destination point _{(B 1, B 2, B} 3). For each angle, a healthy person and a wheelchair user were each experimented 10 times (total of 60 times). At this time, it was measured travel time from point A to _{B 1,} B 2, _{B 3} points. The difference between the case where the angle is changed by 0 degrees (the direction is not changed) and the case where the direction is changed by the angles (45 degrees and 90 degrees) is set as a delay time due to the direction change.

  FIG. 19B shows the experimental results in the subject. From FIG. 19B, the following relational expression (14) was obtained.

・・・・・・式（１４）

・ ・ ・ ・ ・ ・ Formula (14)

  Here, Δθ is a direction change angle, and Δτ is a time for changing the direction of the direction change angle Δθ. When the above equation (14) is rewritten, the following equation (15) is obtained.

・・・・・・式（１５）

・ ・ ・ ・ ・ ・ Formula (15)

Here, K _health represents a delay time due to a direction change unique to a healthy person.

  Similarly, FIG. 19C shows the experimental results for wheelchair users. From FIG. 19C, the following relational expression (16) was obtained.

・・・・・・式（１６）

・ ・ ・ ・ ・ ・ Formula (16)

  Here, Δθ is a direction change angle, and Δτ is a time for changing the direction of the direction change angle Δθ. When the above equation (16) is rewritten, the following equation (17) is obtained.

・・・・・・式（１７）

・ ・ ・ ・ ・ ・ Formula (17)

Here, K _wheelchair represents a delay time due to a direction change unique to a wheelchair user.

Comparing the delay time _{K wheelchair} of the delay time of the healthy subject _{K healthy} and wheelchair users in diversion planning unit 103, a delay time due to the direction inversion transformation in wheelchair users is larger than the delay time by the direction rolling transformation in healthy subjects I understand that. The delay time due to this change in direction represents the ease of changing the direction for each attribute.

  Note that even in the same attribute of one person, the ease of changing the direction of the person 5 may change based on the speed of the person 5 detected by the person detection unit 101. For example, in a healthy person, the faster the traveling speed, the more difficult it is to change direction, so the delay time due to the change in direction increases. FIG. 19B shows the direction change delay time in a healthy person whose average running speed is approximately 0.9 m / s. FIG. 19D shows the direction change delay time in a healthy person having an average running speed of approximately 2.2 m / s. As shown in FIG. 19D, the delay time due to the direction change unique to the healthy person is 0.0057 [s / deg].

  The process shown in FIG. 17 is applied to the scene shown in FIGS. 11A and 11B. For example, consider a case where the first person 5a (healthy person) is a person who changes the course, and the second person 5b (wheelchair user) is a person who does not change the course.

FIG. 20A shows a time when the course change planning unit 103 performs the process of step S401. The course change planning unit 103 obtains the position of the autonomous mobile device 1 after a fixed time interval INT _A seconds with respect to the oblique cylinder of the person who does not change the course (second person 5b). At the same time, the course change planning unit 103 obtains the position of the person (second person 5b) who does not change the course after a certain time interval INT _A seconds.

FIG. 20B shows a time when the course change planning unit 103 performs the process of step S402. The course change planning unit 103 obtains the position of the person (first person 5a) whose course is changed after a fixed time interval INT _A seconds with respect to the oblique cylinder of the autonomous mobile device 1. The diagonal cylinder of the autonomous mobile device 1 is slanted between the position of the autonomous mobile device 1 at that time and the goal of the autonomous mobile device 1 (the destination where the autonomous mobile device 1 goes straight when there is no obstacle). A circular path is generated by the course change planning unit 103.

When determining the position of the person 5a whose course is to be changed after a certain time interval INT _A seconds, the time for changing the direction at an angle in the course change is determined using the delay time due to the direction change for each attribute. Obtained at 103. In course change planning unit 103, if the time for turning at an angle in the course change is a constant time interval INT _A above, the position of the person 5a to divert after a certain time interval INT _A seconds Current Location Only the posture is updated.

In that case, it means that the person 5a who changes the course is in the middle of changing direction. If the time for turning at the angle in the course change is within a certain time interval INT _A , the turning is completed, and the time obtained by subtracting the time for turning at the angle in the course change from the constant time interval INT _A The person 5a who changes the course proceeds.

With the progress of the person 5a who changes the course in the remaining time, the course change planning unit 103 updates the position after a certain time interval INT _A seconds.

In the process of step S303, when the person 5a whose course is changed passes the autonomous mobile device 1, the route of the person 5a whose course is changed is returned so as to return to the original route of the person 5a who changes the course. FIG. 21 shows how the person 5a who changes the course turns. A point A in FIG. 21 is a coordinate position when the person 5a starts a course change, and a point B in FIG. 21 is a coordinate position when the person 5a passes the autonomous mobile device 1. v _h represents the speed of a person. The direction of v _{h is} the original direction of the person 5a before the course is changed, and the value h when the AB vector is projected in that direction is obtained by the following equation (18). Here, the angle α is an angle formed by the original direction of the person 5a and the direction of the vector AB.

・・・・・・式（１８）

・ ・ ・ ・ ・ ・ Formula (18)

  Point C in FIG. 21 is the coordinate position of the destination of the return, and is calculated by the following equation (19).

・・・・・・式（１９）

・ ・ ・ ・ ・ ・ Formula (19)

In the process of step S303, if the person 5a whose course has been changed comes before the person 5b whose course has not been changed and the relative distance at that time is within a certain threshold R _stop , the person 5b whose course has not been changed is separated by a certain time interval. Stop for INT _A seconds. As an example, it is assumed that the radius of the person 5 is 0.3 m and a safety distance of 0.2 m is provided, and the threshold R _stop is 0.8 m in the first embodiment.

FIG. 22A shows a result of performing the processing shown in FIG. 17 for each constant time interval INT _A seconds on the scene shown in FIGS. 11A and 11B. FIG. 22A is a planned route in the combination C3 shown in FIG. 14A. On the other hand, the planned route in the combination C2 shown in FIG. 14A is shown in FIG. 22B.

  The load evaluation unit 104, in each course change method, the position of each person 5 with respect to the autonomous mobile device 1, the speed of each person 5, and the ease of changing the direction according to the attribute of the person (depending on the attribute relating to the ease of movement of the person) The mobility load of the autonomous mobile device 1 and the mobility load of each person 5 are evaluated based on the ease of direction change. Each route change method uses information planned by the route change planning unit 103. The position of each person 5 and the speed of each person 5 use information acquired by the person detection unit 101 and input via the determination unit 102 and the course change planning unit 103. The ease of changing the direction according to the attribute of the person uses information acquired by the attribute determination unit 108 and input via the course change planning unit 103. When evaluating by, the load evaluation part 104 evaluates by the total movement load which is the sum total of the movement load of the autonomous mobile device 1 and the movement load of each person 5 as an example. In the first embodiment, as an example, a delay time is used as an index of the mobile load. It should be noted that other indicators such as the number of people 5 who change course when the autonomous mobile device 1 passes by a plurality of people 5 may be used as an indicator of the movement load.

  FIG. 35 shows an example in which the number of people 5 whose course is changed is used as an index. Here, it is assumed that all persons are healthy persons. When the autonomous mobile device 1 moves to the left side, the route of the first person 5c is changed. On the other hand, when the autonomous mobile device 1 moves to the right side, the route is changed between the second person 5d and the third person 5e. Thus, since the movement load of the person 5 tends to increase as the number of the persons 5 who change the course increases, if all the persons 5 have the same attribute, to easily calculate the movement load, The number of people 5 who change the course may be used.

  The load evaluator 104 calculates the delay time from the start of the course change to the return to the original path by the person 5a who changes the course as shown in the following equation (20).

・・・・・・式（２０）

・ ・ ・ ・ ・ ・ Formula (20)

Here, ΔT, ΔT _A , and ΔT _B represent the delay time of the person 5a who changes the course, the delay time due to the direction change, and the increase in the travel time of the path after the course change (see FIG. 23).

In the load evaluation unit 104, the delay time [Delta] T _A by turning determined by the following equation (21).

・・・・・・式（２１）

・ ・ ・ ・ ・ ・ Formula (21)

Here, M is the number of times the person 5a who changes the course has changed the direction, and τ _j is a delay time due to the j-th direction change (where j is a positive integer).

Moving time increment [Delta] T _B of the path taken after the course change, represents the difference between the moving time when moving time and human 5b when the human 5b is diverted is not diverted, determined by the following equation (22).

・・・・・・式（２２）

・ ・ ・ ・ ・ ・ Formula (22)

Here, t ₀ is the time when people 5a began to change course. t ₁ ′ is the time when the person 5a returns to the original route when the delay time due to the direction change is not taken into consideration. t ₁ is the time when the person 5a reaches the position when the person 5a does not change the course with respect to the position where the person 5a returns to the route.

  The delay time of all the persons 5 and the delay time of the autonomous mobile device 1 are summed as shown in the following formula (23) to obtain a total movement load that is the sum of the person 5 and the autonomous mobile device 1.

・・・・・・式（２３）

・ ・ ・ ・ ・ ・ Formula (23)

  FIG. 24 shows the movement load (as delay time) in the course changing method (combination of course patterns) C2 and C3 shown in FIG. 14A. It can be seen that the way of changing routes (combination of route patterns) C2 has a higher movement load than the way of changing routes (combination of route patterns) C3. This is because the autonomous mobile device 1 changes the course of a wheelchair user who is difficult to change direction, and the total delay time, which is the sum of the delay time of the person 5 and the delay time of the autonomous mobile device 1, increases.

  In the case shown in FIG. 11A and FIG. 11B, the healthy person is in front of the wheelchair user (side closer to the autonomous mobile device 1), but the person who changed the direction of the healthy person (the healthy person and the chair user) The total movement load that is the sum of the movement load of 5 and the movement load of the autonomous mobile device 1 is reduced.

  However, if priority is given to those who are difficult to change direction, the total movement load is not necessarily reduced. As such a case, the example of FIG. 25A and FIG. 25B is shown. Here, the first person 5a is a healthy person (a delay time parameter K = 0.004 [s / deg] due to a specific direction change), and the second person 5b is less likely to change direction than a normal person (a specific type The parameter K of delay time due to the direction change is 0.0120 [s / deg]. Even in this case, the healthy person (first person 5a) is in front of the person (second person 5b) who is hard to change direction (the side closer to the autonomous mobile device 1). However, in FIGS. 25A and 25B, both the first person 5 a and the second person 5 b are already closer to the autonomous mobile device 1 than in the case shown in FIGS. 11A and 11B.

  FIG. 26A and FIG. 26B are generated routes in the respective route changing methods, and FIG. 27 shows delay times corresponding to the respective routes.

  In the first course changing method shown in FIG. 26A, the direction of the first person 5a is greatly changed. That is, the first person 5a is largely changed in the right direction as viewed from the first person 5a, and after the avoidance of the autonomous mobile device 1, the first person 5a is greatly changed in the left direction. As a result, the first person 5a prevents the second person 5b from proceeding, and a large delay time (1.4000 s) occurs in the second person 5b.

  On the other hand, in the method of changing the second route shown in FIG. 26B, the second person 5b who is in the back (distant from the autonomous mobile device 1) as seen from the autonomous mobile device 1 is turned. . That is, first, the second person 5b is greatly changed in the left direction as viewed from the second person 5b, and after the autonomous mobile device 1 is avoided, the second person 5b is greatly changed in the right direction. As a result, the delay time of the first person 5a does not occur, and the total delay time is smaller in the second route change method than in the first route change method. Although the second person 5b is less likely to change the direction than the first person 5a, the total movement load is greater when the second person 5b is rerouted than when the first person 5a is rerouted. Is small.

  The subject determination unit 105 selects a route having the smallest total movement load from all the route change methods planned by the route change planning unit 103 based on the evaluation result of the load evaluation unit 104. Based on the course change method selected by the subject determination unit 105, the subject determination unit 105 determines the person who changes the course and the person who does not change the course. For example, in the scenes shown in FIGS. 11A and 11B, the course change method (course pattern combination) C <b> 3 is selected by the subject determination unit 105 based on the evaluation results shown in FIG. 24. Therefore, the first person 5a (healthy person) is determined by the person who changes the course and the target person determination unit 105, and the second person 5b (wheelchair user) is determined by the person who does not change the course and the target person determination unit 105.

  The movement route determination unit 106 determines the route of the autonomous mobile device 1 based on the information determined by the subject determination unit 105, that is, the course change method selected by the subject determination unit 105. The route determined by the movement route determination unit 106 is output to the travel control unit 107.

  The travel control unit 107 controls the left and right motors 12aM and 12bM of the autonomous mobile device 1 so that the autonomous mobile device 1 travels along the route determined or planned by the travel route determination unit 106 or the avoidance route planning unit 109. Control each one.

  FIG. 4 shows a processing flow of the operation of the autonomous mobile device 1 according to the first embodiment.

  First, using the information from the LRF sensor 2 and the information from the environment map storage unit 110 and the timer 101T, the human detection unit 101 detects the person 5 in the path 6 where the autonomous mobile device 1 travels (step S210). ).

  Next, based on the information detected by the human detection unit 101, the human detection unit 101 determines whether there is a person (step S211). When the human detection unit 101 determines that there is no person 5 in the passage 6, the normal route planning unit 112 plans the normal route of the autonomous mobile device 1 (step S213). After the normal route is planned by the normal route planning unit 112 in step S213, the process proceeds to step S208. On the other hand, when the person detection unit 101 detects that there are a plurality of people 5 in the passage 6 at step S211, the person detection unit 101 acquires information on the position and speed of each person 5 (step S201).

  Next, based on the information detected by the human detection unit 101 and the information from the distance image sensor 3, the attribute determination unit 108 determines the movement attribute (attribute based on the movement ability) of each person 5 detected by the human detection unit 101. Judgment is made (step S202).

  Next, the determination unit 102 determines whether or not the autonomous mobile device 1 needs to take safety measures for the person 5 (step S212).

  When the determination unit 102 determines that it is necessary to take a safety measure in step S212, as an example of the safety measure, the autonomous mobile device 1 travels in an emergency stop or retreats to a certain distance. A signal for giving an instruction to the control unit 107 is generated (step S214). After generating a signal for safety measures in step S214, the process proceeds to step S208.

  If the determination unit 102 determines that it is not necessary to take safety measures in step S212, the presence / absence of an avoidance route of the autonomous mobile device 1 is determined based on information from the human detection unit 101 and the autonomous mobile device information acquisition unit 111. The determination unit 102 determines (step S203). That is, the determination unit 102 first calculates the diameter of the cylinder including the safety area of the autonomous mobile device 1 calculated based on the information from the autonomous mobile device information acquisition unit 111 and the safety of the person 5 detected by the human detection unit 101. Find the diameter of the cylinder including the region. Thereafter, in the determination unit 102, the diameter of the cylinder including the safety area of the autonomous mobile device 1 and the diameter of the cylinder including the safety area of the person 5 in the passing with the plurality of persons 5 detected by the person detection unit 101. Then, using the passage width stored in the environment map storage unit 110, the presence / absence of a width for passage of the autonomous mobile device 1 is calculated, and the determination unit 102 determines the presence / absence of an avoidance route of the autonomous mobile device 1.

  When the determination unit 102 determines that there is an avoidance route of the autonomous mobile device 1 in step S203, the autonomous mobile device along the route that can secure the width for passage of the autonomous mobile device 1 based on the information from the determination unit 102 One avoidance route is planned by the avoidance route planning unit 109 (step S209). Thereafter, the process proceeds to step S208.

  When the determination unit 102 determines that there is no avoidance route of the autonomous mobile device 1 in step S203, the route change planning unit 103 is based on information from the determination unit 102, the attribute determination unit 108, and the autonomous mobile device information acquisition unit 111. Thus, the combinations of the course changes of the person 5 among the plurality of persons 5 are listed and planned (step S204).

  Next, based on the information from the course change planning unit 103, based on the attribute of each person 5 acquired by the attribute determination unit 108 and the position and speed of each person 5 acquired by the person detection unit 101, the course change planning unit For each combination of route changes planned in 103, the movement load of each person 5 in the route change and the movement load of the autonomous mobile device 1 are totaled and evaluated as a total movement load by the load evaluation unit 104 (step) S205).

  Next, based on the evaluation result in the load evaluation unit 104, the subject determination unit 105 selects the combination having the smallest total movement load, and the subject determination unit 105 determines the person who changes the course and the person who does not change the course ( Step S206).

  Next, the movement route determination unit 106 determines the movement route of the autonomous mobile device 1 based on the combination selected by the subject determination unit 105 (step S207).

  Next, the safety measure generated by the determination unit 102 so that the autonomous mobile device 1 travels along the route determined or planned by the normal route planning unit 112, the movement route determination unit 106, or the avoidance route planning unit 109. The left and right motors 12aM and 12bM of the autonomous mobile device 1 are controlled by the travel control unit 107 so as to perform temporary stop or reverse based on the temporary stop or reverse signal (step S208).

  Here, the case where the autonomous mobile device 1 is controlled by the method shown in FIG. 4 (hereinafter referred to as the proposed method) according to the first embodiment is compared with the case where it is controlled by the conventional method. FIG. 28 shows an example of a control method in the case where an avoidance route is not found by the conventional technique described in Non-Patent Document 1. The autonomous mobile device 1 advances toward a predetermined wall side, for example, the left wall 51w2. When approaching the wall side, the autonomous mobile device 1 moves to a predetermined position (for example, 1 m ahead from the current position of the autonomous mobile device 1). If it passes with all the target persons 5 (when the person 5 detected by the person detection unit 101 disappears), it proceeds toward the destination of the autonomous mobile device 1.

  In the scenes shown in FIGS. 11A and 11B, FIGS. 29A and 29B show the results when autonomous movement of the autonomous mobile device is executed by the conventional method and the proposed method, respectively. 29A and 29B, the upper diagram is a diagram of the route between the autonomous mobile device and each person, and the lower tabular diagram is the mobility load of the autonomous mobile device and the mobility load of each person. It is a figure which shows the example of. As shown in FIG. 29A, the autonomous mobile device 1 travels toward the left wall 51w2 that is always determined. Since there is a wheelchair user 1b on the left wall side, in the conventional method, it is necessary to change the direction of the wheelchair, and the total movement load increases. On the other hand, in the proposed method, the movement route is determined by the movement route determination unit 106 after the movement load is evaluated by the load evaluation unit 104 based on the position, speed, and attribute of the person 5b, so the total movement load is reduced. Therefore, in the scenes shown in FIGS. 11A and 11B, the difference in delay time between the conventional method and the proposed method is large as shown in FIGS. 29A and 29B.

  On the other hand, FIGS. 30A and 30B show a scene where the difference in delay time between the conventional method and the proposed method is small. The situation of the scene shown in FIGS. 30A and 30B is almost similar to the scene shown in FIGS. 11A and 11B, except that the healthy person 5a is on the left wall 51w2 side and the wheelchair user 5b. Is on the right wall 51w1 side. FIGS. 31A and 31B show the results when the autonomous mobile device 1 is executed by the conventional method and the proposed method in the scenes shown in FIGS. 30A and 30B, respectively. 31A and 31B, the upper diagram is a diagram of the route between the autonomous mobile device and each person, and the lower tabular diagram is the mobility load of the autonomous mobile device and the mobility load of each person. It is a figure which shows the example of.

Note that the processing flow shown in FIG. 4 is performed at regular time intervals INT _B , and the movement route is re-planned based on the position, speed, and attribute of each person and the position and speed of the autonomous mobile device 1 at that time. . In the first embodiment, as an example, the constant time interval INT _B is set to 1.0 s.

  In addition, when the person 5 does not react as expected when traveling on the travel route planned by the autonomous mobile device 1, the autonomous mobile device 1 takes safety measures. For example, when the person 5 approaches the autonomous mobile device 1 instead, the autonomous mobile device 1 performs an emergency stop or the autonomous mobile device 1 moves back to a certain distance.

  Although the present invention has been described based on the above embodiment or modification, it is needless to say that the present invention is not limited to the above embodiment or modification. The following cases are also included in the present invention.

  Each autonomous mobile device controller 1C is specifically a computer system including a microprocessor, ROM, RAM, a hard disk unit, a display unit, a keyboard, a mouse, and the like. A computer program is stored in the RAM or hard disk unit. Each device achieves its functions by the microprocessor operating according to the computer program. Here, the computer program is configured by combining a plurality of instruction codes indicating instructions for the computer in order to achieve a predetermined function.

For example, each component can be realized by a program execution unit such as a CPU reading and executing a software program recorded on a recording medium such as a hard disk or a semiconductor memory. In addition, the software which implement | achieves the autonomous mobile apparatus control part 1C in the said embodiment or modification is the following programs. In other words, this program
A route change planning unit that generates a route change of each person when the autonomous mobile device 1 and the plurality of people 5 pass each other, and plans one or more ways of changing the route;
An attribute determination unit that determines an attribute relating to the ease of movement of the person;
The movement of the autonomous mobile device based on the position of each person relative to the autonomous mobile device, the speed of each person, and the ease of changing the direction according to the attribute relating to the ease of movement of the human determined by the attribute determining unit A load evaluation unit that calculates and evaluates a total mobile load that is the sum of the load and the mobile load of each person;
Based on the evaluation information in the load evaluation unit, the total of the total load of each person's movement load and the movement load of the autonomous mobile device is determined based on the one or more course change methods planned by the course change planning unit. A method for selecting a course change with the smallest moving load, and a target person determination unit for determining a person who changes the course and a person who does not change the course among the plurality of persons,
Based on the information determined by the target determination unit, the autonomous mobile device travels along the travel route determined by the travel route determination unit and the travel route determination unit that determines the travel route of the autonomous mobile device Thus, there is a program for an autonomous mobile device for causing the autonomous mobile device to function as a travel control unit that controls the autonomous mobile device.

  The program may be executed by being downloaded from a server or the like, and a program recorded on a predetermined recording medium (for example, an optical disk such as a CD-ROM, a magnetic disk, or a semiconductor memory) is read out. May be executed.

  Further, the computer that executes this program may be singular or plural. That is, centralized processing may be performed, or distributed processing may be performed.

  In addition, it can be made to show the effect which each has by combining arbitrary embodiment or modification of the said various embodiment or modification suitably.

  An autonomous mobile device, an autonomous mobile method, and an autonomous mobile device program according to an aspect of the present invention provide a route so that a total movement load between each person and the autonomous mobile device is minimized when passing by a plurality of people. It has functions to plan and control to travel on the route, and is useful as a transport robot, a cleaning robot, etc. in a hospital. The present invention can also be applied to a guidance robot, a security robot, a mobile robot on which a person rides, etc. outside a hospital.

DESCRIPTION OF SYMBOLS 1 Autonomous mobile device 1a Main body part 1C Autonomous mobile device control part 2 LRF sensor 3 Distance image sensor 5 Person 5a First person, Healthy person 5b Second person, Wheelchair user 11 Laser line 12a-12d Wheel 12aM, 12bM For wheel drive Motor 12aE, 12bE Wheel rotation detection encoder 51 Environmental map 54 Absolute coordinate system 80 in environmental map Spot 81 where a laser beam hits a known obstacle 101 Spot 101 where a laser beam hits an unknown obstacle Human detection unit 101T Timer 102 Determination unit 103 Course Change Planning Unit 104 Load Evaluation Unit 105 Target Person Determination Unit 106 Travel Route Determination Unit 107 Travel Control Unit 108 Attribute Determination Unit 109 Avoidance Route Planning Unit 110 Environmental Map Storage Unit 110G Environmental Information Acquisition Unit 111 Autonomous Mobile Device Information Acquisition Unit 112 Normal route planning department

Claims (8)

A route change planning unit that generates a route change of each person when the autonomous mobile device and a plurality of people pass each other, and plans one or more ways of changing the route;
An attribute determination unit that determines an attribute relating to the ease of movement of the person;
The movement of the autonomous mobile device based on the position of each person relative to the autonomous mobile device, the speed of each person, and the ease of changing the direction according to the attribute relating to the ease of movement of the human determined by the attribute determining unit A load evaluation unit that calculates and evaluates a total mobile load that is the sum of the load and the mobile load of each person;
Based on the evaluation information in the load evaluation unit, the total of the total load of each person's movement load and the movement load of the autonomous mobile device is determined based on the one or more course change methods planned by the course change planning unit. A method for selecting a course change with the smallest moving load, and a target person determination unit for determining a person who changes the course and a person who does not change the course among the plurality of persons;
Based on the information determined by the subject determination unit, a movement route determination unit that determines a movement route of the autonomous mobile device,
An autonomous mobile device comprising: a travel control unit that controls the autonomous mobile device such that the autonomous mobile device travels along the travel route determined by the travel route determination unit. A human detection unit for detecting a person passing by the autonomous mobile device;
The course change planning unit creates a pattern for changing the course and a pattern that does not change the course for each of a plurality of persons detected by the person detection unit, and all the people detected by the person detection unit The autonomous mobile device according to claim 1, wherein a combination of patterns is planned.   The route change planning unit expresses the person and the autonomous mobile device as slanted cylinders, performs a route search in space and time, and in each way of changing the route, The autonomous mobile device according to claim 1, wherein after generating a travel route of the mobile device, a travel route of a person whose course is changed with respect to the oblique cylinder of the autonomous mobile device is generated.   The load evaluation unit is configured to provide a delay time as an extra time when each person changes the route with respect to the case where each person does not change the route and a case where the autonomous mobile device does not change the route. The autonomous movement according to any one of claims 1 to 3, wherein the total movement load is evaluated by a total delay time obtained by adding a delay time as an extra time when the autonomous mobile device changes a route. apparatus. A human detection unit that detects a person passing by the autonomous mobile device and the speed of the person;
The load evaluation unit evaluates the movement load of the person by changing the ease of the person's direction change based on the speed of the person detected by the person detection unit in the attribute of one person. The autonomous mobile apparatus as described in any one of 1-4. A human detection unit for detecting a person passing by the autonomous mobile device;
The autonomous mobile device according to any one of claims 1 to 5, wherein the load evaluation unit evaluates the movement load of the person by the number of persons detected by the person detection unit and changing the course. The route change planning unit generates a route change for each person when the autonomous mobile device and a plurality of people pass each other, and plans one or more ways to change the route,
The attribute determination unit determines the attribute related to the ease of movement of the person,
Based on the position of each person relative to the autonomous mobile device, the speed of each person, and the ease of changing the direction according to the attribute relating to the ease of movement of the person determined by the attribute determination unit by the load evaluation unit, Evaluate the total mobile load, which is the sum of the mobile mobile device's mobile load and each person's mobile load,
Based on the evaluation information in the load evaluation unit, the subject person determination unit determines the movement load of each person and the movement load of the autonomous mobile device from the one or more course change methods planned by the course change planning unit. Selecting the course change method with the smallest total movement load of the above and determining the person who changes the course and the person who does not change the course among the plurality of persons,
Based on the information determined by the subject determination unit by the travel route determination unit, determine the travel route of the autonomous mobile device,
The travel control unit controls the autonomous mobile device so that the autonomous mobile device travels along the travel route determined by the travel route determination unit.
Autonomous movement method. A program for an autonomous mobile device for controlling an autonomous mobile device that moves by autonomously determining a movement route,
Computer
A route change planning unit that generates a route change of each person when the autonomous mobile device and a plurality of people pass each other, and plans one or more ways of changing the route;
An attribute determination unit that determines an attribute relating to the ease of movement of the person;
The movement of the autonomous mobile device based on the position of each person relative to the autonomous mobile device, the speed of each person, and the ease of changing the direction according to the attribute relating to the ease of movement of the human determined by the attribute determining unit A load evaluation unit that calculates and evaluates a total mobile load that is the sum of the load and the mobile load of each person;
Based on the evaluation information in the load evaluation unit, the total of the total load of each person's movement load and the movement load of the autonomous mobile device is determined based on the one or more course change methods planned by the course change planning unit. A method for selecting a course change with the smallest moving load, and a target person determination unit for determining a person who changes the course and a person who does not change the course among the plurality of persons,
Based on the information determined by the subject determination unit, a movement route determination unit that determines a movement route of the autonomous mobile device, and the autonomous mobile device along the movement route determined by the movement route determination unit A program for an autonomous mobile device for causing the autonomous mobile device to function as a travel control unit that controls the autonomous mobile device to travel.

Images