JP6599143B2 - Autonomous mobile robot - Google Patents

Description

  The present invention relates to an autonomous mobile robot that autonomously moves in a moving space.

  Conventionally, an autonomous mobile robot that autonomously moves and acquires an image of a target object such as an intruder who has entered a monitoring area has been proposed.

  For example, in Patent Document 1, a plurality of movement candidate positions are set at peripheral positions separated from a target object by a predetermined separation distance, and a movement candidate position close to the current position is selected from these movement candidate positions as a movement target. An autonomous mobile robot that is set as a position, generates a flight path from the current position to the movement target position, and is flight-controlled along the flight path is disclosed.

  Further, in Patent Document 2, the target object can be viewed from various angles as much as possible by making it easy to set a movement candidate position that can be imaged from a direction in which the target object has not been captured by the camera. An autonomous mobile robot that enables imaging is disclosed.

JP 2014-142828 A JP 2014-119901 A

  However, since the above-described conventional technology does not consider concealment (occlusion) by an obstacle existing in the movement space when setting the movement target position, the set movement target position may be hidden by the obstacle. It is not always a position where the target object can be imaged.

  Accordingly, the present invention has an object of enabling a follow-up flight while imaging a target object by setting a position where a predetermined function such as imaging can be performed on the target object as a movement target position. And

One aspect of the present invention is an autonomous mobile robot including a functional unit that autonomously moves in a moving space and exhibits a predetermined function with respect to a target object, and a plurality of partial spaces that constitute the moving space Spatial information indicating the position in the moving space for each, attribute information indicating whether or not it becomes a failure of the function for each partial space and the degree of functional failure that indicates the degree of failure of the function , A storage unit that stores a target object position indicating the position of the target object, position estimation means for estimating the current self position and posture of the autonomous mobile robot, and a movement target position to set the movement from the self position Route search means for calculating a movement route to a target position, and movement control means for controlling the autonomous mobile robot to move along the movement route, wherein the route search means includes the spatial information and the attribute. Affection Depending on the evaluation value obtained by the evaluation function using the dysfunction of the positional relationship between the position and the desired object position of the subspace to be the failure of the function obtained from the function is exhibited easily the An autonomous mobile robot is characterized in that processing is performed such that a position in a moving space is more easily set to the target movement position .

  Here, the partial space is preferably a unit space obtained by dividing the moving space by a predetermined unit size.

  The storage unit further stores functional condition information indicating an effective range of the functional unit, and the route search means uses the functional condition information and the target target position to determine whether the target target is the effective target. It is preferable to set the movement target position at a position within the range.

  Further, the storage unit further stores a feature position indicating a relative position from the target object position with respect to a feature part of the target object, and the route search means includes the target object position, the feature position, and It is preferable that the movement target position is set at a position where the function can be performed with respect to the characteristic part.

  ADVANTAGE OF THE INVENTION According to this invention, the position which can exhibit the function of an autonomous mobile robot can be provided as a movement target position, and the autonomous mobile robot which can move on the movement path | route which goes to the said movement target position can be provided.

It is a figure which shows the structure of the autonomous mobile robot in embodiment of this invention. It is a figure which shows the structure of the autonomous mobile robot system in embodiment of this invention. It is a functional block diagram which shows the structure of the autonomous mobile robot in embodiment of this invention. It is a figure which shows the example of the spatial information in embodiment of this invention. It is a figure which shows the example of the attribute information in embodiment of this invention. It is a figure which shows the graph structure of the voxel used for the production | generation of the movement path | route in embodiment of this invention. It is a figure for demonstrating the setting method of the movement target position in embodiment of this invention. It is a figure for demonstrating the setting method of the movement target position in embodiment of this invention. It is a figure for demonstrating the setting method of the movement target position in embodiment of this invention. It is a figure explaining the calculation process of the local target in embodiment of this invention.

  The autonomous mobile robot 1 according to the embodiment of the present invention is a quad-rotor type small unmanned helicopter as shown in the general view of FIG. The scope of application of the present invention is not limited to a quad-rotor type small unmanned helicopter, but can be similarly applied to a single-rotor type small unmanned helicopter and a traveling robot that moves autonomously.

  As shown in the system configuration diagram of FIG. 2, the autonomous mobile robot 1 communicates with the external security center 100 and the management device 102, and uses a predetermined moving object existing in the monitoring space (moving space) as the target object M. It is configured to track and to exhibit a predetermined function for the target object M. The moving object that is the target object M is, for example, a person (such as a bandit) or a vehicle that has entered the monitoring area. In the present embodiment, a function for imaging the target object M will be described as an example of the predetermined function. However, the function is not particularly limited, and a sound is emitted from the target object M or a warning by light emission is given. It may be a function such as performing.

  The security center 100 and the management device 102 are connected via an information communication network 110 such as the Internet so that information can be transmitted. The autonomous mobile robot 1 and the management device 102 are connected so as to be able to transmit information by wireless communication or the like.

  The security center 100 communicates with the autonomous mobile robot 1 via the management device 102 and receives a captured image of the target object M captured by the autonomous mobile robot 1. The security center 100 may have a function of performing image processing on a captured image and issuing a warning to an administrator (not shown) who is monitoring an abnormality at the security center 100. Further, it has a function of receiving information on the position (coordinates) of the target object M from the management device 102 and managing the target object M and the captured image captured by the autonomous mobile robot 1 in association with each other. Also good.

The management apparatus 102 includes a fixed target object detection sensor 104 (104a, 104b,...) Installed on the ground or a wall surface, and detects the position of the target object M. The target object detection sensor 104 can be a laser sensor, for example. The laser sensor scans the laser two-dimensionally at a certain angular sample interval, thereby detecting objects (obstacles) existing within a detection range in a horizontal plane at a certain height from the ground (or floor surface). Is obtained as polar coordinate values. The laser sensor scans the exploration signal, which is laser light radially, receives the exploration signal reflected back from the object, calculates the distance to the object from the time difference between transmission and reception, and installs the laser sensor Then, the polar coordinate value of the position of the object is obtained from the coordinates in which the search signal is transmitted and the calculated distance, and the three-dimensional orthogonal coordinate value (X _t , Y _t , Z _t ) is obtained from the polar coordinate value. The management apparatus 102 transmits the position of the object obtained by the target object detection sensor 104 to the autonomous mobile robot 1 as the position of the target object M. When the autonomous mobile robot 1 receives the position of the target object M, the autonomous mobile robot 1 calculates its own movement path based on the position, and moves along the movement path. The management apparatus 102 tracks the target object M over the detection ranges of a plurality of laser sensors by verifying the identity of the target object M existing in an area where the detection ranges of the laser sensors overlap. That is, even if the target object M existing in the detection range of the laser sensor 104a moves to the detection range of the laser sensor 104b, the management apparatus 102 can determine that the target object M is the same object. .

  Hereinafter, the configuration and functions of the autonomous mobile robot 1 will be described with reference to the general view of FIG. 1 and the functional block diagram of FIG.

  As shown in FIG. 1, the autonomous mobile robot 1 has four rotors (propellers) 2 (2a to 2d) on one plane. Each rotor 2 is rotated using a motor 4 (4a to 4d) driven by a battery (secondary battery: not shown). Generally, in a single rotor type helicopter, the azimuth angle is maintained by canceling the counter torque generated by the main rotor with the moment generated by the tail rotor. On the other hand, in a quad-rotor type helicopter such as the autonomous mobile robot 1, counter-torque is canceled by using a rotor 2 that rotates in different directions in front and rear and left and right. Then, by controlling the number of rotations (fa to fd) of each rotor 2, various movements and adjustments of the posture can be performed. For example, when it is desired to rotate the airframe in the yaw direction, a difference may be given to the rotational speeds of the front and rear rotors 2a and 2c and the left and right rotors 2d and 2b.

  The imaging unit 3 includes a two-dimensional image sensor having an optical system such as a lens and a two-dimensional array element such as a CCD or CMOS with predetermined pixels (for example, 640 × 480 pixels), and captures captured images of the flight space for a predetermined time. This is a so-called color camera that is acquired at intervals. In the present embodiment, the imaging unit 3 is installed in the housing portion so that its optical axis captures the front direction of the autonomous mobile robot 1 and is obliquely below by a predetermined depression angle θ from the horizontal plane (XY plane). It is installed to image the space at a viewing angle φ. The acquired captured image is output to the control unit 7 described later, and is stored in the storage unit 8 by the control unit 7 or transmitted to the management apparatus 102 via the communication unit 9 described later.

  The distance detection sensor 5 is a sensor that detects a distance between an obstacle existing around the autonomous mobile robot 1 and the autonomous mobile robot 1 and acquires a relative position of the obstacle present in the sensor detection range. is there. In the present embodiment, a microwave sensor is provided as the distance detection sensor 5. The microwave sensor emits a microwave to space and detects the reflected wave, thereby detecting an object around the autonomous mobile robot 1 and obtaining a distance to the object. The distance detection sensor 5 is provided, for example, toward the front of the autonomous mobile robot 1 and can be used to measure the distance to the target object M. In addition, the distance detection sensor 5 can be provided, for example, downward in the lower part of the autonomous mobile robot 1 and used to measure the distance (altitude) from the ground. Further, the distance detection sensor 5 is provided, for example, in the direction of the optical axis of the imaging unit 3 and can be used to measure the distance to the target object M that is the imaging object of the imaging unit 3.

  The position detection sensor 6 is a sensor for acquiring the current position of the autonomous mobile robot 1. The position detection sensor 6 receives radio waves (navigation signals) transmitted from navigation satellites (artificial satellites) such as GNSS (Global Navigation Satellite System). The position detection sensor 6 receives navigation signals transmitted from a plurality of navigation satellites (artificial satellites) and inputs them to the control unit 7.

  Note that the position detection sensor 6 may acquire information for obtaining a self-position by a known conventional technique using another sensor such as a laser scanner, a gyro sensor, an electronic compass, or an atmospheric pressure sensor.

The communication unit 9 is a communication module for performing wireless communication with the management apparatus 102 through, for example, a wireless LAN or a mobile phone line. In the present embodiment, a captured image acquired by the imaging unit 3 is transmitted to the management device 102 by the communication unit 9, and the captured image is transmitted from the management device 102 to the security center 100, so that a security guard or the like can invade remotely. It is possible to monitor the person. In addition, the communication unit 9 receives the position (coordinates: X _t , Y _t , Z _t ) of the target object M from the management device 102, thereby enabling setting of a movement route as described later.

  The storage unit 8 is an information storage device such as a ROM (Read Only Memory), a RAM (Random Access Memory), and an HDD (Hard Disk Drive). The storage unit 8 stores various programs and various data, and inputs / outputs such information to / from the control unit 7. Various types of data include target object position 81, spatial information 82, attribute information 83, moving space graph information 84, separation condition 85, various parameters 86 and other information used for each process of control unit 7, and outputs of sensors and the like. Values, captured images, and the like.

The target object position 81 is position information (coordinates: X _t , Y _t , Z _t ) of the target object M received from the management apparatus 102. In the present embodiment, the target object position 81 is a foot position of the target object M, that is, a position where the target object M is in contact with the ground (floor surface). When receiving the position of the target object M via the communication unit 9, the control unit 7 stores the target object position 81 in the storage unit 8. The management apparatus 102 indicates the posture of the target object M (for example, the front direction of the target object M) based on the time change of the position of the target object M (coordinates: X _t , Y _t , Z _t ). Information) may be estimated and stored as the target object position 81.

  The space information 82 is information representing the structure of obstacles in the flight space by dividing the flight space into a plurality of voxels (unit spaces) as a voxel space, and is set in advance by an administrator or the like and stored in the storage unit 8. Information. In the present embodiment, the flight space is equally divided into voxels of a predetermined unit size (for example, 15 cm × 15 cm × 15 cm), and as shown in FIG. 4, the voxel ID that is an identifier for each voxel and the voxel in the flight space Are stored as spatial information 82 in association with each other.

  The attribute information 83 is information indicating the attribute of each voxel. The attributes include “movement failure attribute” and “functional failure attribute”. In the present embodiment, as shown in FIG. 5, the voxel ID that is the identifier and the attribute of each voxel are associated with each other and stored as attribute information 83.

  The movement obstacle attribute is an attribute indicating whether or not the movement of the autonomous mobile robot 1 becomes an obstacle. For example, a voxel located at an obstacle such as a building is set as a movement obstacle attribute as a space where the autonomous mobile robot 1 cannot move. In addition, voxels located in a space near the obstacle are also set as movement obstacle attributes so as not to approach the obstacle more than necessary. In the present specification, the voxel located in the obstacle is referred to as “occupied voxel”, the voxel located in the space near the obstacle is referred to as “proximity voxel”, and the other voxels are referred to as “free voxels”.

  The functional failure attribute is an attribute indicating whether or not the functional failure of the autonomous mobile robot 1 is a failure. For example, when the autonomous mobile robot 1 has an imaging function, a voxel having translucency does not have a functional failure attribute, but a voxel having a light shielding property has a functional failure attribute.

  For example, a light-blocking obstacle such as a general wall or roof is a movement obstacle attribute and a function obstacle attribute. On the other hand, an obstacle having translucency such as a glass window, a net, and a fence is a movement obstacle attribute, but is not a functional obstacle attribute. Further, fog and smoke are functional disorder attributes, but are not movement disorder attributes.

  Each voxel is registered in the storage unit 8 in association with a voxel cost value indicating the degree of occupancy so that it can be used when a travel route is generated by the route search means 73 described later. The voxel cost value takes a maximum value in the occupied voxel and is set to be smaller as the distance from the occupied voxel increases. For example, it is preferable to calculate the voxel cost value from a calculation formula of voxel cost value = exp {−λ · (distance from occupied voxel)}. Here, λ is a parameter obtained by experiment. In order to prevent the autonomous mobile robot 1 from going out of the predetermined movable space in the moving space, the voxel serving as the boundary between the movable space and the outside is regarded as the occupied voxel, and the voxel cost of the voxel is concerned. It is preferable to set the value as the maximum value.

The movement space graph information 84 is three-dimensional graph information created based on the space information 82. Specifically, based on the spatial information 82, it is information having a graph structure in which the center position of each voxel is a node and a line segment connecting nodes adjacent to the node is an edge. FIG. 6 is a diagram for explaining the graph structure of the moving space graph information 84, in which some (27) voxels in the flight space are cut out. In FIG. 6, each cube represented by a symbol B represents a voxel. Further, a sphere filled with black or hatching at the center position of the voxel B is a node, and a line segment displayed by a dotted line connecting the nodes is an edge. It is assumed that a distance cost C _{M (} described later) obtained based on the distance between adjacent nodes is set as the edge weight in the moving space graph information 84.

  The separation condition 85 is information indicating a positional relationship to be maintained between the autonomous mobile robot 1 and the target object M when flying following the target object M. The separation condition 85 is preset by an administrator of the autonomous mobile robot 1 or the like. When the predetermined target object M is monitored using the autonomous mobile robot 1, it is necessary to approach the target object M and acquire a more detailed captured image. However, in order not to be attacked by the target object M that is hostile such as an intruder, it is necessary to be separated by a certain distance or more. Therefore, the autonomous mobile robot 1 of the present embodiment can acquire a detailed captured image of the target object M (the target object M is located within the imaging range of the imaging unit 3), and the target object M The separation condition 85 is obtained in advance so as to be a distance at which it is difficult to receive an attack, and is stored in the storage unit 8, and controlled to follow the flight so as to satisfy the separation condition 85. The separation condition 85 is set, for example, as a separation distance of 3 m from the target object M and a separation altitude of 3 m from the target object M in the plane. Further, the separation condition 85 may be a condition indicating a range having a certain width.

  The various parameters 86 are other parameters necessary for exerting control and functions of the autonomous mobile robot 1. The various parameters 86 include, for example, imaging condition information (functional condition information) and the like. The imaging condition information is information representing the field of view of the imaging unit 3. As the imaging condition information, the depression angle θ and the viewing angle φ of the imaging unit 3 are stored. The imaging condition information is appropriately set by an administrator or the like based on the characteristics of the imaging unit 3, the installation angle, and the like. In addition, when the camera control apparatus which can change the imaging direction of the imaging part 3 is mounted with respect to the autonomous mobile robot 1, you may make it acquire the information of depression angle (theta) from the said camera control apparatus. When the imaging unit 3 that can change the viewing angle φ is used, information on the viewing angle φ may be acquired from the imaging unit 3. The imaging condition information described above is used when a predetermined function to be exerted on the target object M is an imaging function by the imaging unit 3, but when a function other than the imaging function is targeted, imaging condition information is used. Instead, it can be functional condition information. The function condition information includes a condition for defining a range where the function can be sufficiently exhibited. For example, in the case of a function such as emitting a sound to the target object M or giving a warning by light emission, the function condition information includes information that can specify a range in which sound and light reach effectively. For example, the radiation direction θ ′ of sound or light and the half width (radiation angle at which the radiation energy becomes −3 dB) φ ′, which is an index indicating the directivity, may be set as the function condition information.

  The control unit 7 is composed of a computer including a CPU and the like. As a series of processes for performing position estimation processing, speed estimation processing, route search processing, and movement control processing (route tracking control), position estimation means 71, speed estimation means. 72, a route search means 73, and a movement control means 74 are included.

  The position estimation unit 71 performs position estimation processing for estimating the current position (self-position) of the autonomous mobile robot 1 in the flight space based on the output of the position detection sensor 6.

More specifically, the latitude / longitude estimated based on the known known technology based on the navigation signals from the plurality of navigation satellites obtained from the position detection sensor 6 and the altitude obtained from the distance detection sensor 5 The position coordinates (X _s , Y _s , Z _s ) are calculated. Further, the posture YAW is obtained as a self position in response to an output from the position detection sensor 6 such as an electronic compass or a gyro sensor. The self-position estimation method is not limited to this, and the current position of the autonomous mobile robot 1 may be estimated using another method.

The position estimation means 71 includes the estimated self position (coordinates: X _s , Y _s , Z _s and posture YAW) and the position of the target object M received from the management device 102 (coordinates: X _t , Y _t , Z _t). ) To the route search means 73.

  The position estimating unit 71 performs a process of updating the spatial information 82 and the attribute information 83 based on the position of the target object M. Specifically, a cylindrical model (for example, a monitoring target) having a size substantially the same as the size of the target object M that is determined in advance in the voxel space based on the spatial information 82 of the storage unit 8 with the position of the target object M as the center. When the target object M is an intruder, a cylinder model having a bottom radius of 0.3 m and a height of 1.7 m is arranged, and voxels that interfere with the cylinder model are designated as “movement obstacle attribute” and “function obstacle”. The spatial information 82 and the attribute information 83 are updated by setting as an occupied voxel having “attribute”. As will be described later, the autonomous mobile robot 1 is flight-controlled so as not to move to the occupied voxel, but the autonomous mobile robot 1 is updated by updating the spatial information 82 based on the position of the target object M as described above. Contact between the target object 1 and the target object M can be avoided.

The speed estimation means 72 performs processing for estimating the current flight speed (v _x , v _y , v _z , v _yaw ) of the autonomous mobile robot 1 to be used for movement control in the movement control means 74 described later. In the present embodiment, the flight speed is obtained from the time change of the self-position (coordinates: X _s , Y _s , Z _s and attitude YAW) estimated by the position estimation means 71. At this time, it is preferable to estimate the flight speed using an extended Kalman filter in consideration of the influence of measurement errors and the like. In addition, a speed estimation method using the Doppler effect in GNSS may be used.

  The route searching unit 73 calculates a moving route of the autonomous mobile robot 1 using the self position estimated by the position estimating unit 71 and the position of the target object M and various information stored in the storage unit 8. Process. In the present embodiment, the route searching process in the route searching means 73 generates a moving route from the own position to the moving target position set in the vicinity of the target object M.

  In the route search process, first, a process for setting a movement candidate position based on the target object position 81 and the separation condition 85 is performed. In the present embodiment, a process of setting a plurality of movement candidate positions in a surrounding area separated from the target object position 81 by a separation distance and a separation height determined as the separation condition 85 is performed. The movement candidate position is a circular position having a radius of the separation distance (for example, 3 m) stored in the storage unit 8 with the target object M as the center, and a separation altitude (for example, 3 m) predetermined by the administrator or the like. ) Areas are set at equal intervals.

  Fig.7 (a)-FIG.7 (c) are explanatory drawings about the setting of a movement candidate position. FIG. 7A to FIG. 7C are views when a part of the flight space displayed by the voxel at time t is looked down from directly above.

  A black area represented by reference numeral B1 indicates an occupied voxel having a movement failure attribute and a function failure attribute. For example, the voxel indicates a space occupied by a light-shielding object such as a building. A hatched region represented by reference symbol B2 indicates an occupied voxel having a movement failure attribute but not having a functional failure attribute. For example, the voxel indicates a space occupied by a light-transmitting object such as glass, a net, or a fence. An area painted with dots represented by reference numeral B3 indicates a proximity voxel having a movement failure attribute but not having a function failure attribute. A white region (a region where nothing is painted) represented by reference numeral B4 indicates a free voxel having neither a movement failure attribute nor a functional failure attribute.

  First, as shown in FIG. 7A, the route searching means 73 sets a plurality of movement candidate positions Pc at equal angular intervals on a circumference centered on the target object M and having a separation condition 85 as a radius. To do. In the example of FIG. 7A, 72 movement candidate positions Pc are set every 5 °.

  Also, functional condition information indicating the effective range of the imaging unit 3 (functional unit) of the autonomous mobile robot 1 instead of setting the movement candidate position Pc based on the separation condition 85 (fixed value) preset by the administrator. The target object M may be obtained from the autonomous mobile robot 1 in the moving space where the target object M is in the effective range, and the movement candidate position Pc may be set in the position or range.

  For example, if the function of the autonomous mobile robot 1 is imaging, functional condition information indicating imaging conditions such as the angle of view of the imaging unit 3 is read from the storage unit 8, and the functional condition information and the position of the target object M are read. Based on the above, the position or range in the moving space where the target object M falls within the imaging range of the imaging unit 3 is obtained geometrically, and the storage unit 8 is updated with the separation condition 85 indicating the position or range. . For example, the altitude at which the target object M can be imaged (within the imaging range) at a position separated from the target object M by a predetermined horizontal distance (for example, 3 m) is geometrically calculated based on the functional condition information, The separation condition 85 is updated with the value including the horizontal distance and the altitude. Alternatively, the horizontal distance that allows the target object M to be imaged (within the imaging range) at a predetermined altitude position is calculated geometrically based on the functional condition information, and the separation condition is determined based on the value of the horizontal distance and the altitude. 85 is updated. Then, a plurality of movement candidate positions Pc are set at positions that satisfy the updated separation condition 85.

  Further, a feature position indicating a relative position of the feature portion of the target object M from the target object position 81 is stored in the storage unit 8, and a plurality of pieces of information are obtained using the feature position and the target object position 81. The movement candidate position Pc may be set. For example, when the target object M is a person (bandit), the characteristic part is a part that can identify a person (bandon) such as a face, and the characteristic position is approximately 150 cm in the height direction (Z-axis direction) that is the position of the face. May be stored as the position. Then, using the position of the characteristic part obtained from the target object position 81 and the characteristic position, a positional relation for the characteristic part to fall within the effective range of the function of the functional unit (imaging unit 3) is obtained and the positional relation ( For example, a position at a horizontal distance of 3 m and a height of 3 m from the target object position 81 is stored in the storage unit 8 as the separation condition 85, and the movement candidate position Pc is set to a position that satisfies the separation condition 85. . At this time, a position or range where the characteristic part of the target object M falls within the effective range of the functional unit (imaging unit 3) of the autonomous mobile robot 1 is geometrically obtained, and a plurality of movement candidate positions are included in the position or range. Pc may be set. In addition to the above, when the target object M is a vehicle, the characteristic part is a license plate, and the target object position 81 including the posture (front direction) of the target object M and the characteristic position are used. The position of the characteristic part (number plate) may be obtained.

  Next, the route search means 73 excludes the movement candidate position Pc included in the voxel having the movement obstacle attribute from the set movement candidate positions Pc. That is, movement candidate positions Pc included in “occupied voxels” and “proximity voxels” among the set movement candidate positions Pc are excluded from the candidates.

  Furthermore, the route search means 73 selects the movement candidate position Pc from which the target object M can be imaged among the remaining movement candidate positions Pc. First, each of the remaining movement candidate positions Pc is taken as a movement candidate position Pc in which attention is paid in turn, and a straight line connecting the movement candidate position Pc in focus and the target object position 81 is obtained. Next, with reference to the spatial information 82, each voxel existing on the straight line is extracted, and the voxel ID of the extracted voxel is obtained. Then, with reference to the attribute information 83, the attribute of the voxel specified by the obtained voxel ID is checked, and the movement candidate position Pc focused on when there is a voxel that is a function failure attribute is excluded from the candidates. This is because the target object M cannot be imaged from the focused movement candidate position Pc when there is a voxel that is a functional disorder attribute. On the other hand, if there is no voxel that is a functional disorder attribute, the focused movement candidate position Pc is left as a candidate. This is because the target object M can be imaged from the focused movement candidate position Pc. By these processes, the movement candidate position Pc is selected as shown in FIG.

  When the characteristic position of the target object M is set, a straight line connecting the focused movement candidate position Pc and the characteristic position of the target object M is obtained, and whether or not the voxel on the straight line is a functional disorder attribute. It is preferable to determine whether or not.

  Alternatively, the movement candidate position Pc at which the target object M can be imaged may be obtained using a rendering technique from a virtual viewpoint in general three-dimensional computer graphics processing. That is, a voxel whose attribute information 83 is a functional disorder attribute exists as an imaging obstacle (opaque object) when viewed from the virtual viewpoint (imaging unit 3 provided virtually) at the movement candidate position Pc. To do. On the other hand, voxels whose attribute information 83 is not a functional impairment attribute exist as not being an imaging obstacle (transparent) when viewed from the virtual viewpoint (imaging section 3 provided virtually) at the movement candidate position Pc. I will do it. Then, when the three-dimensional model of the target object M installed at the target object position 81 from each of the movement candidate positions Pc is rendered (imaged), it is determined whether or not the target object M can be imaged. And geometrically using the imaging conditions.

  The route search means 73 uses the movement candidate position Pc that is closest to the self-position estimated by the position estimation means 71 among the movement candidate positions Pc remaining in the above processing, that is, the movement candidate position Pc having a smaller linear distance from the self-position. Set to position. Thereby, the movement target position Q is set as shown in FIG.

  If it is determined that the target object M cannot be imaged for all the movement candidate positions Pc, the movement closest to the self position among the movement candidate positions Pc remaining after being excluded as being included in the voxel having the movement obstacle attribute. The candidate position Pc may be set as the movement target position Q.

  That is, in the present embodiment, the route search means 73 considers not only the three-dimensional shape of the obstacle in the flight space but also whether or not it becomes an obstacle to exhibiting the function of the autonomous mobile robot 1, A movement target position Q at which the function of the robot 1 can be exerted is preferentially set.

  Then, the route search means 73 uses the set movement target position Q, the spatial information 82 and the movement space graph information 84 stored in the storage unit 8, and the self position estimated by the position estimation means 71. A movement path generation process for calculating the movement path of the autonomous mobile robot 1 is performed. In the movement path generation process, the spatial information 82 and the movement space graph information 84 are referred to, and the voxel node corresponding to the movement target position (hereinafter referred to as “start node”) to the voxel node corresponding to the self position (hereinafter referred to as “goal node”). The route with the smallest total cost value C to “)” is searched by the A * route search method.

In the A * route search method, the total cost value C (n) at a certain evaluation node n (node whose node ID is n) is expressed by Expression (1).

Here, g (n) is a g cost value in the A * route search method for the evaluation node n, and in this embodiment, the moving distance from the start node to the evaluation node n and the contact with the obstacle Calculated as a cost value considering the danger. That is, g (n) which is the g cost value of the evaluation node n is g (n-1) which is the g cost value in the adjacent node and the distance cost set based on the distance from the adjacent node to the evaluation node. The sum of C _M and the voxel cost C _v (n) at the evaluation node is obtained by g (n) = g (n−1) + C _M + C _v (n). H (n) is an h cost value in the A * route search method for the evaluation node n, and is an estimated cost value of the distance from the evaluation node n to the goal node. In the present embodiment, h (n) is obtained from the straight line distance from the evaluation node n to the goal node. In the A * route search method, the calculation of the total cost value C (n) of the adjacent nodes in order from the start node is repeated, and the route having the smallest total cost value C (n) finally reaching the goal node is obtained. I will explore.

  Hereinafter, the generation of a movement route by the A * route search method will be briefly described. In the A * route search method, first, one or a plurality of nodes (adjacent nodes) adjacent to the start node are set as evaluation nodes. Then, the total cost value is obtained for each evaluation node by the equation (1). Next, when the node with the lowest total cost value among the evaluation nodes is referred to as the attention node, the adjacent node (newly not set as the evaluation node so far) of the attention node is newly added as the evaluation node. The total cost value is similarly obtained for the newly added evaluation node. Subsequently, the evaluation node having the minimum total cost value among all the evaluation nodes is reset as the attention node. In this way, the resetting of the target node based on the total cost value of the evaluation node, the addition of a new evaluation node and the calculation of the total cost value accompanying the resetting are repeated, and the goal node is finally set as the target node When done, the route search ends.

  In this way, the best movement path from the start node as the starting point to the goal node as the end point is generated. The data of the travel route generated by the route search means 73 is set data of the position (x, y, z) of the node serving as the via point, and this information is temporarily stored in the storage unit 8.

  The route search method is not limited to the A * route search method, and other route search methods such as the Dijkstra method may be applied.

  The movement control means 74 uses the movement route calculated by the route search means 73, the self-position estimated by the position estimation means 71, and the flight speed estimated by the speed estimation means 72, so that the autonomous mobile robot 1 Route follow-up control is performed so as to fly along the movement route calculated by the route search means 73. Specifically, a control command value, which is a flight control value at each time, is obtained using the movement route, the self position, and the flight speed, the motor 4 is controlled based on the control command value, and the rotation speed of the rotor 2 is determined. Control.

In the route following control, first, a process of calculating the nearest position (hereinafter referred to as “local target”) that the autonomous mobile robot 1 should target at each time is performed. FIG. 8 is a diagram for explaining the calculation of the local target. In calculating the local target, the movement control means 74 reads the movement route generated by the route search means 73 from the storage unit 8, and has already passed the transit point W _p1 that the autonomous mobile robot 1 is aiming at at the current time. A straight line W connecting the two points with the via point W _p0 is obtained. Then, the movement control means 74 calculates the intersections L _p ′, L _p between the obtained straight line W and the sphere S centered on the self-position of the autonomous mobile robot 1, and the intersection L close to the intended via point W _{p1. Find p} as a local goal. In this way, by performing flight control so that the autonomous mobile robot 1 moves toward the local target at each time, the local target always moves on the moving path toward the moving target position _Po , and the autonomous mobile robot. 1 will fly along the movement path.

Next, in the path following control, a process of calculating control command values u _x , u _y , u _z , uψ for each direction of X, Y, Z, and yaw angle so as to fly toward the calculated local target is performed. . At this time, a control command value is determined so that the difference between the current self position and the position of the local target is small. Specifically, the control command value u = (u _x , u _y , u _z ) in the XYZ-axis directions is calculated based on the self position r = (X _s , Y _s , Z _s ) obtained by the position estimation means 71 and the speed. The speed v = (v _x , v _y , v _z ) estimated by the estimation means 72 is used to obtain by PID control. Each control command value of the XYZ-axis direction _{u = (u x, u y} , u z), the local target r '= when (x, y, z) and the control command value is, u = K _p (r '-R) + K _d · v + K _i · e is calculated. _{Here, K p, K d, K} i is that the gain of the PID control, _{respectively, e = (e x, e} y, e z) is the integral value of the error. On the other hand, the control command value uψ in the yaw angle direction is expressed as follows: ψ ′ is the target angle, ψ is the posture (angle) of the autonomous mobile robot 1 estimated by the position estimation means 71, and v _yaw is the angular velocity estimated by the speed estimation means 72. Then, determined by PD control as equation _{uψ = K p (ψ'-ψ} ) + K d · vψ. In the present embodiment, the target angle ψ ′ is an angle that faces the direction of the target object M, that is, the direction of the position of the target object M.

  In this way, the control unit 7 sequentially repeats the processes in the position estimation unit 71, the speed estimation unit 72, the route search unit 73, and the movement control unit 74 described above. Thereby, the autonomous mobile robot 1 of this Embodiment is a position which satisfy | fills the separation conditions 85 from the target object M, Comprising: A movement target position is sequentially with respect to the position which can exhibit the function of the imaging part 3 (functional part). By updating and sequentially updating the moving route each time, it is possible to follow the target object M while appropriately imaging it.

<Modification 1>
In the above-described embodiment, whether or not the function of the autonomous mobile robot 1 becomes an obstacle is registered as the attribute information 83, and a position where the function cannot be exerted on the target object M is moved based on the attribute. The processing is excluded from the target position Q candidates.

  On the other hand, it is good also as what registers the functional failure degree which shows the grade used as the failure of the function of the autonomous mobile robot 1 as the attribute information 83. FIG. The degree of functional failure is 0% for voxels that do not cause any functional failure of the autonomous mobile robot 1 and 100% for voxels that are completely functionally impaired (completely blocked). What is necessary is just to set so that it may become a high value as the grade to become high. For example, in the case where the function of the autonomous mobile robot 1 is imaging, the range in which each voxel becomes an obstacle to imaging based on the reflectance, transmittance, refractive index, etc. of the voxel for each voxel is in the range of 0 to 100%. You can set in.

  When the functional failure degree is registered as the attribute information 83 as described above, the route search means 73 sets the movement target position Q as follows.

  First, the route search means 73 installs a plurality of movement candidate positions Pc at equiangular intervals on a circumference centered on the target object M and having a separation condition 85 as a radius, as in the above embodiment. Next, the route search means 73 excludes the movement candidate position Pc included in the voxel having the movement obstacle attribute from the set movement candidate positions Pc. That is, movement candidate positions Pc included in “occupied voxels” and “proximity voxels” among the set movement candidate positions Pc are excluded from the candidates. Furthermore, the route search means 73 selects the movement candidate position Pc from which the target object M can be imaged among the remaining movement candidate positions Pc. At this time, first, the remaining movement candidate positions Pc are sequentially considered as movement candidate positions Pc, and a straight line connecting the focused movement candidate position Pc and the target object position 81 (or feature position) is obtained. Next, with reference to the spatial information 82, each voxel existing on the straight line is extracted, and the voxel ID of the extracted voxel is obtained. Then, by referring to the attribute information 83, the attribute of the voxel specified by the obtained voxel ID is checked, and the functional failure degree of the voxels on the straight line connecting the focused movement candidate position Pc and the target object M is accumulated. A value K (e.g., an evaluation value composed of a value obtained by simply adding the degree of functional failure) is calculated. Then, the movement candidate position Pc in which the cumulative value K of the degree of functional failure is equal to or greater than a predetermined reference value is excluded from the candidates. Then, of the remaining movement candidate positions Pc, the movement candidate position Pc that is closest to the self-position estimated by the position estimation means 71, that is, has a shorter linear distance from the self-position, is set as the movement target position.

  Further, an evaluation value is calculated in consideration of the straight line distance L from the self-position of the autonomous mobile robot 1 to each movement candidate position Pc and the cumulative value K of the degree of functional failure for the movement candidate position Pc, and the movement is performed based on the evaluation value. The target position Q may be set. For example, the movement candidate position Pc having the smallest evaluation value calculated by the evaluation function αL + βK consisting of the distance L and the accumulated value K may be set as the movement target position Q. Here, α and β are parameters obtained by experiments. The evaluation function is not limited to this, and may be set as appropriate according to the flight space, function, etc. of the autonomous mobile robot 1.

<Modification 2>
In the above embodiment, as “a plurality of partial spaces constituting the moving space” according to the present invention, the three-dimensional flight space of the autonomous mobile robot 1 is divided into voxels (unit spaces of a predetermined unit size). However, the three-dimensional flight space may be expressed and processed by another three-dimensional data expression method. For example, an object existing in the flight space may be expressed by continuous three-dimensional vector data and processed by three-dimensional CAD or the like.

  Specifically, vector data representing the size and position of an object existing in the flight space is stored as the spatial information 82. For example, the storage unit 8 stores the size and position of the object, which is a combination of coordinates indicating the starting point, and vectors indicating the length, width, and height from the starting point, as the spatial information 82. Further, whether or not each object stored as the spatial information 82 becomes an obstacle to the function of the autonomous mobile robot 1 is stored in the storage unit 8 as attribute information 83 in association with each object.

  The route search means 73 refers to the spatial information 82 and the attribute information 83 for a position that satisfies the separation condition 85 with respect to the target object M and that can perform the function of the autonomous mobile robot 1 with respect to the target object M. And the position is set as the movement target position Q.

  Specifically, first, the route search means 73 sets the movement candidate position Pc on the circumference centered on the target object position 81 and having the separation condition 85 as a radius. Next, the route search means 73 refers to the spatial information 82 and excludes the position where the object exists from the set movement candidate positions Pc on the circumference from the movement candidate positions Pc. Further, the route search means 73 refers to the spatial information 82 and the attribute information 83 to determine whether or not there is an object that becomes a function obstacle on the straight line connecting the remaining movement candidate position Pc and the target object M. And the movement candidate position Pc where an object that becomes a function obstacle exists is excluded from the candidates. A three-dimensional spatial analysis technique may be applied to these processes.

  As described above, the autonomous mobile robot 1 in the above-described embodiment and modifications can search for a route from its own position to the movement target position, and can move autonomously along the searched route. In the present embodiment, the moving target position of the autonomous mobile robot 1 is set at a position where the target object M can be imaged by the obstacle in consideration of concealment (occlusion) by the obstacle. Accordingly, the autonomous mobile robot 1 can follow and fly while imaging the target object M.

  It should be noted that the present embodiment can be similarly applied when the function that the autonomous mobile robot 1 can exhibit is not imaging. For example, when the autonomous mobile robot 1 has a function of emitting a sound to the target object M, an attribute indicating whether each voxel transmits or does not transmit sound is registered in the attribute information 83, and the movement candidate is registered. If there is a voxel that does not transmit sound between the position Pc and the target object M, the movement candidate position Pc may be excluded from the candidates. In addition, when the autonomous mobile robot 1 has a function of issuing a warning by light emission to the target object M, the attribute information 83 registers an attribute indicating whether each voxel transmits light or not, If there is a voxel that does not transmit light between the movement candidate position Pc and the target object M, the movement candidate position Pc may be excluded from the candidates.

  In the above embodiment, the target object M is detected using the target object detection sensor 104 connected to the management device 102. However, the present invention is not limited to this, and the position of the target object M may be estimated by analyzing the captured image acquired by the imaging unit 3. For example, each frame of the captured image is subjected to image processing to extract an image area of the target object M. At this time, a discriminator using an optical flow method or boosting learning (for example, a face detection method using an AdaBoost-based discriminator using Haar-like features), which is a known prior art (see Japanese Patent Laid-Open No. 2006-146551). The image area (person area) of the target object M is extracted using the above. Next, the distance between the target object M and the autonomous mobile robot 1 is estimated based on the position of the extracted image region. Specifically, a correspondence table between the y-coordinate position (in the captured image) and the distance of the top of the image area of the extracted target object M is created in advance for each flight altitude, and the current flight altitude and target object The distance from the autonomous mobile robot 1 is estimated by comparing the y coordinate position of the top of the object M with the correspondence table. However, the present invention is not limited to this, and the distance may be calculated from the size of the head of the extracted target object M. That is, a correspondence table between the size of the head and the distance is prepared in advance, and the distance from the autonomous mobile robot 1 is estimated by comparing the size of the head of the extracted target object M with the correspondence table. May be.

  In addition, a distance image sensor is mounted on the autonomous mobile robot 1 instead of a color camera as the imaging unit 3, and a target object M is extracted by a known moving object extraction technique using a distance image acquired from the distance image sensor. Then, the position of the target object M may be estimated from the distance value between the extracted target object M and the autonomous mobile robot 1 and the self position. Alternatively, the autonomous mobile robot 1 may be equipped with a laser scanner, and the position of the target object M may be estimated using the output value of the laser scanner and its own position.

  Moreover, in the said embodiment, in the control part 7, a series of processes of a position estimation process, a speed estimation process, a route search process, and a movement control process (path follow-up control) are performed. However, the present invention is not limited to this, and a control PC (not shown) may be prepared and the PC may perform a series of these processes. That is, the autonomous mobile robot 1 receives the control command value obtained by the position estimation process, the speed estimation process, the route search process, and the movement control process performed by the PC from the PC by wireless communication or wired communication. You may make it fly to the target position by controlling the rotation speed of the motor 4 based on a value. In this way, by sharing the above-described series of processes using the external PC, it is possible to reduce the CPU processing load of the autonomous mobile robot 1 and thus to suppress battery consumption.

  In the route search process in the above embodiment, the self-position is set as the starting point of the moving path, the moving target position set in the vicinity of the target object M is set as the end point of the moving path, and the moving path from the starting point to the end point. Is generated. However, the present invention is not limited to this, and a movement path from the position near the self position to the movement target position may be generated. For example, a self-position (predicted arrival position) at a future time from the self-position by a predetermined time (for example, 1 second) is predicted from the current self-position and moving speed, and the predicted arrival position is set as the starting point of the movement path. Also good.

  Further, in the route search process in the above embodiment and the modification, the movement candidate position Pc is set at a position that satisfies the separation condition 85, and whether or not each movement candidate position Pc can exhibit a function with respect to the target object M. Is judged. However, the present invention is not limited to this, and by referring to the space information 82 and the attribute information 83, the position or range in the moving space where the function can be performed with respect to the target object M is geometrically calculated. Alternatively, a plurality of movement candidate positions Pc may be set within the range, and the movement target position Q may be selected from these movement candidate positions Pc based on the separation condition 85, the distance from the own position, or the like.

  DESCRIPTION OF SYMBOLS 1 Autonomous mobile robot, 2 (2a-2d) Rotor, 3 Imaging part, 4 (4a-4d) Motor, 5 Distance detection sensor, 6 Position detection sensor, 7 Control part, 8 Storage part, 9 Communication part, 71 Position estimation Means, 72 route search means, 72 speed estimation means, 73 route search means, 74 movement control means, 81 target object position, 82 spatial information, 83 attribute information, 84 movement space graph information, 85 separation condition, 86 various parameters, DESCRIPTION OF SYMBOLS 100 Security center, 102 Management apparatus, 104 Target object detection sensor, 110 Information communication network.

Claims (4)

An autonomous mobile robot having a function unit that autonomously moves in a moving space and performs a predetermined function on a target object,
Wherein the space information indicating a position in the mobile space for each of a plurality of subspaces constituting the moving space, showing the extent to which the disability become whether and said function of said function for each of the subspaces function A storage unit that stores attribute information indicating a degree of obstacle and a target object position indicating a position of the target object;
Position estimation means for estimating the current self-position and posture of the autonomous mobile robot;
Route search means for setting a movement target position and calculating a movement route from the self position to the movement target position;
Movement control means for controlling the autonomous mobile robot to move along the movement path;
Have
The route search means is obtained by an evaluation function using the degree of functional disorder from the positional relationship between the position of the partial space that becomes an obstacle to the function and the target object position obtained from the spatial information and the attribute information . An autonomous mobile robot characterized in that , depending on an evaluation value, processing is performed such that a position in the movement space where the function is easily exhibited is more easily set as the movement target position . The autonomous mobile robot according to claim 1,
The autonomous mobile robot according to claim 1, wherein the partial space is a unit space obtained by dividing the moving space by a predetermined unit size. The autonomous mobile robot according to claim 1 or 2,
The storage unit further stores functional condition information indicating an effective range of the functional unit,
The autonomous mobile robot characterized in that the route search means sets the movement target position at a position where the target object is within the effective range using the functional condition information and the target object position. The autonomous mobile robot according to any one of claims 1 to 3 ,
The storage unit further stores a characteristic position indicating a relative position from the target object position with respect to a characteristic part of the target object,
The autonomous mobile robot characterized in that the route search means sets the movement target position to a position where the function can be performed with respect to the characteristic part based on the target object position and the characteristic position.

Images