JP2020027574A - Autonomous mobile device, method, and three-dimensional modeling system - Google Patents

Abstract

To provide an autonomous mobile device, a method, and a three-dimensional modeling system capable of achieving recovery from a stack state, if any.SOLUTION: According to one embodiment, an autonomous mobile device is capable of autonomous movement with a predetermined space provided as a search space. The autonomous mobile device includes: specification means for specifying a search space; first setting means for setting a strategy indicating the next behavior for a state provided upon movement in the search space; collection means for moving in the search space in accordance with the set strategy and collecting three-dimensional modeling information for generating a three-dimensional model; and second setting means for setting end condition for ending the movement in the search space with the set strategy, and the first setting means switches the strategy to a strategy different from the currently set strategy in a case where the end condition has been satisfied.SELECTED DRAWING: Figure 3

Description

  Embodiments of the present invention relate to an autonomous mobile device, a method, and a three-dimensional modeling system.

In recent years, opportunities for using robots in various fields have been increasing. For example, when generating a three-dimensional model of a city, a room, or the like, information necessary for generating the three-dimensional model is collected by using an autonomously movable robot.
For example, in a space where an object for which a three-dimensional model is to be generated exists, a plurality of autonomously movable robots are moved to collect information necessary for generating a three-dimensional model, and the collected information is A technique for generating a three-dimensional model based on the three-dimensional model has been devised.

JP 2013-109325 A

  However, for multiple robots that can move autonomously, it is necessary to move in an unknown space and collect information, so it takes a long time to complete the information collection, There is a possibility that an inconvenience such as falling into a stuck state where the vehicle cannot advance from the middle may occur.

  A problem to be solved by an embodiment of the present invention is to provide an autonomous mobile device, a method, and a three-dimensional modeling system that can recover from a stuck state even if the stack state has been lost. .

  According to one embodiment, the autonomous mobile device can autonomously move using a predetermined space as a search space. The autonomous mobile device includes a specifying unit that specifies the search space, a first setting unit that sets a policy indicating a next action with respect to a state when moving in the search space, and the search according to the set policy. Collection means for moving the space and collecting three-dimensional modeling information to generate a three-dimensional model; and second setting for setting an end condition for ending the movement of the search space by the set policy. Means, wherein the first setting means switches the measure to a measure different from the currently set measure when the end condition is satisfied.

FIG. 1 is a diagram illustrating a schematic configuration example of a three-dimensional modeling system including an autonomous mobile device according to an embodiment. FIG. 2 is a diagram showing a hardware configuration example of the autonomous mobile device according to the embodiment. FIG. 2 is a diagram showing a functional configuration example of the autonomous mobile device according to the embodiment. The figure for demonstrating the search space and divided area | region set by the autonomous mobile apparatus which concerns on the embodiment. 4 is a flowchart showing an example of the operation of the autonomous mobile device according to the embodiment. The figure showing the example of functional composition of the autonomous mobile device concerning the modification of the embodiment.

  Hereinafter, embodiments will be described with reference to the drawings. The disclosure is merely an example, and the invention is not limited by the contents described in the following embodiments. Variations readily conceivable to those skilled in the art are, of course, within the scope of the disclosure. In order to make the description clearer, in the drawings, the size, shape, and the like of each portion may be schematically illustrated as being changed from the actual embodiment. In a plurality of drawings, corresponding elements are denoted by the same reference numerals, and detailed description may be omitted.

  FIG. 1 shows a schematic configuration example of a three-dimensional modeling system including an autonomous mobile device according to an embodiment. The three-dimensional modeling system shown in FIG. 1 includes an autonomous mobile device 10 and an information processing device 20, and the autonomous mobile device 10 and the information processing device 20 are communicably connected via a network N.

  The autonomous mobile device 10 autonomously moves (runs) in a predetermined space and generates information necessary to generate a three-dimensional model for the predetermined space (hereinafter, referred to as “three-dimensional modeling information”). Are collected to generate the three-dimensional model. In the present embodiment, it is assumed that the autonomous mobile device 10 is a so-called drone. However, the present invention is not limited to this, and any device having a mechanism capable of autonomous movement can be used as the autonomous mobile device 10. Applicable.

The information processing device 20 is a personal computer (PC), a tablet terminal, a smartphone, various wearable terminals, or the like operated by a user, and transmits various information and various requests to the autonomous mobile device 10. Specifically, the information processing device 20 transmits search space information and division information, a condition setting request, a policy setting request, and the like, which will be described later, to the autonomous mobile device 10 in response to a user operation. The user performs an operation using a GUI (Graphical User Interface) or a CUI (Character User Interface) in order to transmit the above-described various information and various requests. Specifically, the user performs an operation of inputting a numerical value, an operation of specifying a predetermined space for moving the autonomous mobile device 10 (in other words, an operation of specifying a predetermined space in which a three-dimensional model is to be generated), various conditions and various measures. Perform an operation to select. It should be noted that the information processing device 20 is preferably equipped with a GPS (Global Positioning System) so that longitude, latitude, and altitude (elevation) can be easily specified when specifying the predetermined space. preferable.
The information processing device 20 receives the information indicating the three-dimensional model generated by the autonomous mobile device 10 and displays the information on a display (not shown).

  FIG. 2 illustrates a hardware configuration example of the autonomous mobile device 10. As shown in FIG. 2, in the autonomous mobile device 10, a bus 11 is connected to a nonvolatile memory 12, a CPU (Central Processing Unit) 13, a main memory 14, a motor 15, a sensor 16, a communication unit 17, and the like. I have. The non-volatile memory 12 and the main memory 14 constitute a storage device of the autonomous mobile device 10.

  The nonvolatile memory 12 stores, for example, an operating system (OS) and various application programs. The application program includes a program (hereinafter, referred to as a “three-dimensional modeling program”) for generating a three-dimensional model for the above-described predetermined space.

The CPU 13 is, for example, a processor that executes various application programs stored in the nonvolatile memory 12. The CPU 13 also controls each unit included in the autonomous mobile device 10. The main memory 14 is used, for example, as a work area required when the CPU 13 executes various application programs.
In FIG. 2, only the nonvolatile memory 12 and the main memory 14 are shown as storage devices included in the autonomous mobile device 10, but the present invention is not limited thereto. Drive) and other storage devices such as SSD (Solid State Drive).

The motor 15 enables the autonomous mobile device 10 to move when driven. The sensor 16 is various devices capable of collecting three-dimensional modeling information necessary for generating a three-dimensional model, and is, for example, a camera, a laser range finder (lightwave range finder), or the like.
The communication unit 17 connects the autonomous mobile device 10 to the network N by wireless (or wired). The communication unit 17 controls transmission and reception of various information and various requests with the information processing device 20 via the network N.

  FIG. 3 shows a main functional configuration of the autonomous mobile device 10 realized by the CPU 13 of the autonomous mobile device 10 executing the above-described three-dimensional modeling program. Note that the three-dimensional modeling program can be stored in a computer-readable storage medium in advance and distributed. This program may be downloaded to the autonomous mobile device 10 via the network N, for example.

  As shown in FIG. 3, the autonomous mobile device 10 includes an input receiving unit 101, a dividing unit 102, an end condition setting unit 103, a searching unit 104, a generating unit 105, an evaluating unit 106, and the like. The storage device configured by the nonvolatile memory 12 and the main memory 14 stores policy information described later.

  In the present embodiment, each of the units 101 to 106 is described as being realized by the CPU 13 executing a three-dimensional modeling program (that is, software), but is not limited thereto. It may be realized by hardware, or may be realized by a combination of software and hardware.

  The input receiving unit 101 receives an input of various information transmitted from the information processing device 20. For example, the input receiving unit 101 specifies a predetermined space (hereinafter, referred to as a “search space”) in which the autonomous mobile device 10 moves in order to collect three-dimensional modeling information necessary for generating a three-dimensional model. Of search information (search space information). The form of the search space information is not particularly limited as long as the information can specify the search space. For example, the search space information may be information indicating the longitude, latitude, and altitude (elevation) of a position corresponding to each vertex of the search space. good. In FIG. 4, it is assumed that the search space is a rectangular parallelepiped. In this case, the search space information may be information indicating the longitude, latitude, and altitude of positions corresponding to the vertices P1 to P8 of the rectangular parallelepiped search space SS.

  When the search space has a rectangular parallelepiped shape as shown in FIG. 4, the search space information may be in the following form. The search space information indicates the longitude, latitude, and altitude of a position corresponding to one of the vertices of the rectangular parallelepiped search space, and the origin when the one point is regarded as the origin of the three-dimensional coordinate system. It may be information indicating the length in the X direction, the length in the Y direction, and the length in the Z direction from. For example, the search space information includes the longitude, latitude, and altitude (in other words, the coordinates of the vertex P1) of the position corresponding to the vertex P1 of the search space SS illustrated in FIG. 4 and the distance from the vertex P1 to the vertex P2 positioned in the X direction. Information indicating the length Lx, the length Ly to the vertex P4 located in the Y direction, and the length Lz to the vertex P5 located in the Z direction (in other words, the length of each side of the rectangular parallelepiped) may be used. .

  When the search space has a rectangular parallelepiped shape as shown in FIG. 4, the search space information may be in the following form. The search space information indicates the longitude, latitude, and altitude of a position corresponding to one of the vertices of the rectangular parallelepiped search space, and when the one point is regarded as the origin of the three-dimensional coordinate system, Information indicating the longitude, latitude, and altitude of the position corresponding to the vertex located farthest away may be used. For example, the search space information includes the longitude, latitude, and altitude of the position corresponding to the vertex P1 of the search space SS illustrated in FIG. 4, and the longitude, latitude, and altitude of the position corresponding to the vertex P7 furthest from the vertex P1. May be indicated.

  In the present embodiment, the description will be made assuming that the search space is a rectangular parallelepiped as shown in FIG. 4, but the shape of the search space is not limited to this, and may be an arbitrary shape such as a sphere. It doesn't matter.

  As illustrated in FIG. 4, the autonomous mobile device 10 divides the search space SS into a plurality of regions DR by a dividing unit 102 described later, and sequentially divides each region (hereinafter, referred to as a “divided region”) DR. Move and collect three-dimensional modeling information for each divided region DR. The input receiving unit 101 also receives input of information (division information) for defining the divided region DR. Specifically, the division information is information indicating the shape of the division area, the size (granularity) of the division area, the degree of overlap with the adjacent division area, and the like.

The size of the divided region may be represented by the number of divided regions included in the search space. For example, if the divided region has a rectangular parallelepiped shape (voxel) as shown in FIG. May be represented by the length of each side. Alternatively, if the shape of the divided region is spherical, the size of the divided region may be represented by the length of the radius. Also, the degree of overlap with an adjacent divided region is represented by, for example, a real number, and a positive real number indicates that a predetermined divided region overlaps with an adjacent divided region by the length of the real number. In the case of a real number, it indicates that the predetermined divided region is separated from the adjacent divided region by the length of the real number, and if the value is zero, it indicates that the predetermined divided region is in contact with the adjacent divided region. Is also good.
The search space information and the division information that have received the input are sent to the division unit 102.

  When receiving the input of the search space information and the division information transmitted from the input reception unit 101, the division unit 102 specifies (designates) the search space SS based on the search space information, and determines the search space SS based on the division information. The search space SS is divided into a plurality of divided regions DR. Further, when receiving the input of the evaluation information sent from the evaluation unit 106 described later, the division unit 102 adjusts the size (size) of each divided region DR based on the evaluation information.

  In response to a condition setting request from the information processing device 20, the end condition setting unit 103 sets a search end condition serving as a trigger for switching a measure set for the search unit 104 described later (in other words, the current A search termination condition for terminating the search according to the set policy is set). The search end condition is a condition set for the divided region DR. Specifically, as the search end conditions, the time at which the divided region DR was searched, the number of the three-dimensional modeling information collected in the divided region DR (the number of samplings), and the search starting at the next moved divided region DR are started. The search start time is an example.

  For example, it is assumed that the search end condition is a time when the divided area DR is searched. In this case, the search unit 104, which will be described later, separates the search strategy from the current strategy when the time for searching the predetermined divided region DR exceeds the time set as the search end condition. Execute the processing to switch to the measure.

  In the present embodiment, a case will be described in which the search end condition is set for the divided region DR. However, the present invention is not limited to this, and the search end condition may be set for the search space SS.

  The search unit 104 controls the operation of the autonomous mobile device 10 based on one of a plurality of pieces of policy information stored in a storage device configured by the nonvolatile memory 12 and the main memory 14, and A search for DR is performed to collect three-dimensional modeling information. The collected three-dimensional modeling information differs depending on the type of the sensor 16 provided in the autonomous mobile device 10. For example, information on an image captured by a camera or a distance measured by a laser range finder (for details, see FIG. , Distance from the autonomous mobile device 10 to a predetermined object existing in the divided area DR).

  The search unit 104 refers to the search end condition set by the end condition setting unit 103, and switches the search policy from the current policy to another policy when the search end condition is satisfied. More specifically, the search unit 104 reads, out of the plurality of pieces of policy information stored in the storage device, policy information different from the policy information corresponding to the current policy, and searches the policy indicated by the read policy information. , Set as a new strategy. At the start of the search, the search unit 104 reads out the policy information according to the policy setting request from the information processing device 20, and sets the read-out policy as a policy at the start of the search. Alternatively, a predetermined policy determined in advance may be set as a policy at the start of the search.

Here, a search strategy set by the search unit 104 will be described.
The policy indicates a search method (rule), and is defined, for example, by the following equations (1) and (2).
a _t = π (s _t ) (1)
a _{t to} π (· | s _t ) (2)
Time above (1), the s _t in equation (2) represents the state of the autonomous mobile apparatus 10 at time t, the above-mentioned a _t is the state of the autonomous mobile apparatus 10 returns when the predetermined state s _t The action at t is shown. According to the above (1), defines the mapping from the state s _t to action a _t, according to the above (2), the probability of performing an action a _t when the state s _t was observed Is defined. That is, given when the above equation (1) defines a strategy when the state s _t is obtained performing an action a _t [pi, and equation (2), the state s _t is obtained It defines the policy π that in accordance with the probability distribution perform the action a _t.

As specific examples of the state s _t described above, the position status and about the autonomous moving device 10, the state relating to the collection of three-dimensional modeling information, and the like as an example. Specific examples of the actions a _t that described above, and it relates to the moving direction of the autonomous mobile apparatus 10 such as forward and backward, right movement, left movement, rotation, etc., it relates to an angular velocity torque in controlling motor 15, Etc. are mentioned as an example.

In the present embodiment, as a means [pi, when the state s _t is obtained, the strategy [pi ₁ to select an action a _t to move the divided region DR have not yet searched, the state s _t is obtained case, it is assumed that the policy π ₂ to select the action a _t to return to the search starting point can be set. In other words, the storage device described above, a strategy information indicating a policy [pi _1, and measures information indicating the strategy [pi ₂ it is assumed that is at least stored. It should be noted that, in the policy π ₁ and the policy π _2, policy of choosing the action a _t for the state s _t (hereinafter, referred to as "action selection policy") and what is different. In the present embodiment, as described above, the policy π ₁ has the action selection policy “move to the divided area DR that has not been searched yet”, whereas the policy π ₂ has the action selection policy “search”. Return to start point "is assumed.

As described above, when the state s _t is obtained, by policy [pi ₁ to select an action a _t to move the divided region DR have not yet searched is set, an unknown spatial efficiently By searching, three-dimensional modeling information can be efficiently collected. On the other hand, when autonomously exploring unknown space according to such an approach [pi _1, there is a risk of falling into the stack state search will not proceed from the middle. There are various possible causes of the stacking state. For example, the case where the number of times of collecting the three-dimensional modeling information is too large for the size of the divided region DR, the case where the user gets lost in a dead end, and the like are exemplified. .

However, as described above, at least two measures π ₁ and π ₂ having different action selection policies are set to be switchable by the search unit 104. For example, the search is performed according to the measure π ₁ and the stack state is set. even it fell, it is possible to act selection policy switches to a different strategy [pi _2, it is possible to achieve recovery from the stack state described above.
Here, it is assumed that the action selection policy of the policy π ₁ is “move to the divided area DR that has not been searched yet” and the action selection policy of the policy π ₂ is “return to the search start point”. However, the present invention is not limited to this, and the action selection policies of the policies π ₁ and π ₂ may be any of the following examples.
(Action selection policy of policy π ₁ )
Move to a divided area DR that has not been searched yet.
・ Move to an area that has not been searched yet.
・ Move using the right method (or the left method).
(Policy π ₂ of action selection policy)
・ Return to the search start point.
・ Move away from the current location.
・ Move using the left method (or the right method).
・ Return the route you have traveled to a certain extent.
Note that the action selection policies of the above-mentioned measures π ₁ and π ₂ can be arbitrarily combined.

In the present embodiment, when the autonomous mobile device 10 has fallen into the stack state, that is, when the search end condition is satisfied, the search unit 104 switches the policy π to recover from the stack state. However, the method of recovering from the stuck state is not limited to this. For example, when the search termination condition is satisfied, the autonomous mobile device 10 has fallen into the stuck state instead of switching the policy π. The user may be notified that the possibility is high. According to this, it is possible for the user to select a policy π preferable for the user after confirming whether or not the user is in a stuck state (in other words, to set a policy π preferable for the user). It is possible to send a policy setting request). Alternatively, after receiving the notification, the user switches the operation mode of the autonomous mobile device 10 from the automatic operation mode to the manual operation mode, and operates the autonomous mobile device 10 to recover from the stack state. Is also possible. Note that the switching of the operation mode may be performed by the search unit 104 in accordance with a user operation.
The three-dimensional modeling information collected by performing the search is sent to the generation unit 105.

The generation unit 105 generates a three-dimensional model based on the three-dimensional modeling information collected in the search performed by the search unit 104. As a method of generating a three-dimensional model, a method of restoring a three-dimensional shape using information on a distance measured by a laser range finder, or a method of using an image (multi-viewpoint image) captured by a camera is used. For restoring a three-dimensional shape. However, the method of generating the three-dimensional model is not limited to these, and the three-dimensional model may be generated by any known method.
The three-dimensional information indicating the generated three-dimensional model is sent to the evaluation unit 106.

When receiving the input of the three-dimensional information transmitted from the generation unit 105, the evaluation unit 106 evaluates whether the three-dimensional model indicated by the received three-dimensional information is a model of sufficient quality ( judge. As a method of evaluating the quality of the three-dimensional model, for example, it is determined whether or not the three-dimensional model indicated by the three-dimensional information has a missing part, and if it is determined that there is a missing part, the method is insufficient. There is a method of estimating that the quality is sufficient, and evaluating that the quality is sufficient when it is determined that there is no missing portion. However, the method of evaluating the quality of the three-dimensional model is not limited to this, and the quality of the three-dimensional model may be evaluated by any known method.
Evaluation information indicating the result of the above-described evaluation is sent to the division unit 102, the end condition setting unit 103, and the search unit 104. The form of the evaluation information is not particularly limited as long as it is information that can specify at least a portion (divided region DR) having insufficient quality. Also, as a result of the above-described evaluation, three-dimensional information indicating a three-dimensional model evaluated to have sufficient quality is sent to the information processing device 20.

  In the present embodiment, the evaluation unit 106 evaluates whether or not the three-dimensional model is a model of sufficient quality. However, the present invention is not limited to this, and the three-dimensional model is a model of sufficient quality. The evaluation as to whether or not there is any information is performed by the user who operates the information processing apparatus 20, and the evaluation unit 106 receives input of evaluation information indicating the result of the evaluation, and divides the evaluation information as necessary into the dividing unit 102, It may be simply sent to the end condition setting unit 103 and the search unit 104. In this case, when receiving the input of the three-dimensional information from the generation unit 105, the evaluation unit 106 constantly transmits the three-dimensional information to the information processing device 20.

  Upon receiving the input of the evaluation information sent from the evaluation unit 106, the dividing unit 102 adjusts the size (size, granularity) of the divided region DR based on the received evaluation information. More specifically, the dividing unit 102 divides the divided region DR so that the granularity of the divided region DR corresponding to the portion having the insufficient quality indicated by the evaluation information becomes finer.

Further, when receiving the input of the evaluation information sent from the evaluation unit 106, the end condition setting unit 103 adjusts the search end condition set for the divided region DR based on the received evaluation information. I do. Specifically, the end condition setting unit 103 adjusts the search time set as the search end condition only in the divided region DR corresponding to the portion of the insufficient quality indicated by the evaluation information so as to lengthen the search time. The adjustment is performed so as to increase the sampling number of the three-dimensional modeling information set as the search end condition only in the divided region DR.
Further, when the search unit 104 receives an input of the evaluation information sent from the evaluation unit 106, based on the received evaluation information, a direction different from the previous search (for example, different from the previous search) is used. The search start point is adjusted so that the search can be started from a direction in which three-dimensional modeling information can be collected from the angle). Note that the search unit 104 adjusts not the search start point but the direction at which to start searching for the divided region DR having insufficient quality (in other words, the direction to enter the divided region DR having insufficient quality). You may do it.

Next, an example of an operation realized by each functional unit of the above-described autonomous mobile device 10 will be described with reference to a flowchart of FIG.
First, when receiving the input of the search space information and the division information sent from the information processing device 20, the input reception unit 101 transfers the search space information and the division information that have received the input to the division unit 102.
When receiving the input of the search space information transferred from the input receiving unit 101, the dividing unit 102 specifies the search space SS based on the search space information that has received the input (step S1). Further, when receiving the input of the division information transferred from the input reception unit 101, the division unit 102 divides the search space SS specified by the process of step S1 into a plurality of divisions based on the division information received. The region is divided into regions DR (step S2).

Subsequently, the end condition setting unit 103 sets a search end condition according to the condition setting request from the information processing device 20 for the divided region DR (step S3). Further, the search unit 104 reads, from the storage device, policy information corresponding to a policy (for example, the policy π ₁ ) according to the policy setting request from the information processing device 20, and sets the policy as a policy for search (step S4).

  The search unit 104 starts a search (step S5), and collects three-dimensional modeling information while sequentially moving the divided regions DR according to the policy set by the process of step S4 (step S6). The search unit 104 determines, at predetermined time intervals, whether or not the search end condition set by the process of step S3 is satisfied (step S7). If it is determined that the search termination condition is not satisfied (NO in step S7), the process of step S9 described below is performed.

On the other hand, as a result of the processing in step S7, when it is determined that the search end condition is satisfied (YES in step S7), the search unit 104 determines that the autonomous mobile device 10 may be in a stuck state. To switch policies. Specifically, the search unit 104 reads policy information corresponding to a policy (for example, the policy π ₂ ) having a different action selection policy from the current policy from the storage device, and adds a new policy corresponding to the read policy information to the new policy. It is set as a measure (step S8).

  Next, the search unit 104 determines whether the search has been completed (Step S9). Whether or not the search has been completed is determined based on, for example, whether or not the three-dimensional modeling information has been collected in all the divided regions DR divided by the process of step S2. However, when the switching of the policy is performed by the process of step S8, it is determined that the search is completed according to the newly set policy even if the three-dimensional modeling information is not collected in all the divided regions DR. May be.

As a result of the processing in step S9, when it is determined that the search is not completed (NO in step S9), the process returns to the above-described step S6, and three-dimensional modeling information is collected in the divided area DR that has not been searched yet. The processing is executed.
On the other hand, when it is determined that the search has been completed (or the policy has been switched) as a result of the processing in step S9 (YES in step S9), the search unit 104 generates the collected three-dimensional modeling information. Output to the unit 105. The generation unit 105 generates a three-dimensional model based on the three-dimensional modeling information sent from the search unit 104 (Step S10). The three-dimensional information indicating the generated three-dimensional model is sent to the evaluation unit 106.

  Upon receiving the input of the three-dimensional information transmitted from the generation unit 105, the evaluation unit 106 evaluates whether or not the quality of the three-dimensional model indicated by the received three-dimensional information is sufficient (Step S10). S11). As a result of this evaluation, when it is evaluated that the quality of the three-dimensional model is not sufficient (NO in step S11), the evaluation unit 106 divides the evaluation information indicating the result of the evaluation into the division unit 102, the end condition setting unit 103 And the search unit 104. Thereafter, the process returns to the above-described steps S2 and S3, and after the size adjustment of the divided region DR, the adjustment of the search end condition, and the adjustment of the search start point are performed, the processes after the step S4 are executed again.

  On the other hand, when the quality of the three-dimensional model is evaluated to be sufficient as a result of the processing in step S11 (YES in step S11), the evaluation unit 106 outputs the three-dimensional information received in the processing in step S11 as information. This is transmitted to the processing device 20 (step S12), and a series of operations here is terminated.

  In the present embodiment, it is assumed that the autonomous mobile device 10 has the above-described respective functional units, but the functions of the generating unit 105 and the evaluating unit 106 are mounted on the information processing device 20. It is good. In this case, three-dimensional modeling information is collected in the autonomous mobile device 10, and a three-dimensional model is generated and evaluated in the information processing device 20.

  According to the embodiment described above, the autonomous mobile device 10 can set the search end condition in order to determine whether or not there is a possibility that the search has fallen into a stuck state where the search does not proceed halfway. And a search unit 104 capable of switching a policy for searching to another policy having a different action selection policy from the current policy when it is determined that there is a possibility of falling into a stuck state. This makes it possible to recover from a stuck state even if it has fallen.

  Further, according to the present embodiment, since the end condition setting unit 103 can set the search end condition for the divided region DR, the autonomous mobile device 10 falls into a stack state with a finer granularity than the search space SS. It is possible to determine whether or not there is a fear that a stack state may have occurred, and even if the state has fallen into a stack state, the search unit 104 can immediately detect that there is a risk of having fallen into a stack state.

  Furthermore, according to the present embodiment, since the dividing unit 102 can adjust the size of the divided region DR based on the evaluation result in the evaluating unit 106, the quality is insufficient in the next search. This makes it possible to focus on the divided area DR. That is, the collection efficiency of the three-dimensional modeling information can be improved.

Similarly, according to the present embodiment, the end condition setting unit 103 can adjust the search end condition based on the result of the evaluation in the evaluation unit 106, so that the next search is evaluated as having insufficient quality. This makes it possible to focus on the divided region DR. That is, the collection efficiency of the three-dimensional modeling information can be improved.
Furthermore, according to the present embodiment, since the search unit 104 can adjust the search start point based on the result of the evaluation by the evaluation unit 106, the divided areas evaluated as having insufficient quality in the next search DR can be searched from another direction (angle), and the collection efficiency of three-dimensional modeling information can be improved.

Hereinafter, a modified example of the present embodiment will be described.
<Modification>
FIG. 6 illustrates a functional configuration example of the autonomous mobile device 10 according to a modification of the present embodiment. This modification is different from the case shown in FIG. 3 in further including a template selection unit 107 and a display unit 108. In the present modified example, it is assumed that a storage device including the nonvolatile memory 12 and the main memory 14 previously stores template information described later in addition to the above-described policy information.

  In response to a template selection request from the information processing device 20, the template selection unit 107 reads template information corresponding to the template selection request from the storage device described above, and sends the template information to the input reception unit 101 (or the division unit 102). Output. The template information is information for defining a divided region DR corresponding to a predetermined object (for example, a bridge, a building, or the like) in the search space SS, and is used when a previously generated three-dimensional model is generated. This is information generated with reference to the area DR.

  According to this, for example, when generating a three-dimensional model for the search space SS in which a bridge exists, the template selecting unit 107 reads out template information corresponding to the bridge, and reads this through the input receiving unit 101. Since the data can be sent to the dividing unit 102, the dividing unit 102 can divide the search space SS into a plurality of divided regions DR in consideration of the existence of the bridge. That is, instead of dividing the search space SS into all the divided regions DR having the same size, a fine divided region DR is set for a portion where a bridge is considered to exist, and a large divided region DR is set for a portion where a bridge is not considered to exist. The divided region DR can be set, and the collection efficiency of the three-dimensional modeling information can be improved.

  The display unit 108 receives input of three-dimensional information indicating the three-dimensional model generated by the generation unit 105, and causes the display (not shown) to display the three-dimensional model indicated by the three-dimensional information, or virtual display using VR. Display in space. The display (not shown) may be provided on the autonomous mobile device 10, or may be, for example, goggles that can be worn by a user who operates the information processing device 20, not the autonomous mobile device 10 or the information processing device 20. It may be a type display or the like. According to this, the user can intuitively check the quality of the three-dimensional model.

  According to the embodiment described above, it is possible to provide an autonomous mobile device, a method, and a three-dimensional modeling system that can recover from a stuck state even if the stuck state occurs.

  Although some embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These new embodiments can be implemented in other various forms, and various omissions, replacements, and changes can be made without departing from the spirit of the invention. These embodiments and their modifications are included in the scope and gist of the invention, and are also included in the invention described in the claims and their equivalents.

  DESCRIPTION OF SYMBOLS 10 ... Autonomous mobile device, 20 ... Information processing device, 101 ... Input reception part, 102 ... Division part, 103 ... End condition setting part, 104 ... Search part, 105 ... Generation part, 106 ... Evaluation part, 107 ... Template selection part , 108... Display unit.

Claims (11)

An autonomous mobile device capable of autonomously moving a predetermined space as a search space,
Designating means for designating the search space;
First setting means for setting a measure indicating a next action for a state when moving in the search space;
Collection means for moving the search space according to the set policy and collecting three-dimensional modeling information to generate a three-dimensional model;
A second setting means for setting an end condition for ending the movement of the search space by the set policy,
The first setting means includes:
An autonomous mobile device that switches the policy to a policy different from a currently set policy when the termination condition is satisfied. The first setting means includes:
The autonomous mobile device according to claim 1, wherein when the termination condition is satisfied, the policy is switched to a policy having a different action selection policy from a currently set policy. Dividing means for dividing the designated search space into a plurality of regions,
The second setting means,
The autonomous mobile device according to claim 1 or 2, wherein the end condition is set for the designated search space or each of the divided areas. The second setting means includes:
The autonomous mobile device according to claim 3, wherein a travel time is set as the end condition for the search space or each of the areas. The second setting means,
The autonomous mobile device according to claim 3, wherein the number of three-dimensional modeling information collected by the collection unit is set as the end condition for the search space or each area. Generating means for generating a three-dimensional model for the search space based on the collected three-dimensional modeling information;
The autonomous mobile device according to claim 3, further comprising: an evaluation unit configured to evaluate whether the generated three-dimensional model is of sufficient quality. The dividing means,
The autonomous mobile device according to claim 3, wherein the search space is re-divided into a plurality of regions when a three-dimensional model of the search space generated based on the collected three-dimensional modeling information is not of sufficient quality. . An autonomous mobile device capable of autonomously moving a predetermined space as a search space,
Designating means for designating the search space;
First setting means for setting a measure indicating a next action for a state when moving in the search space;
Collection means for moving the search space according to the set policy and collecting three-dimensional modeling information to generate a three-dimensional model;
A second setting means for setting an end condition for ending the movement of the search space by the set policy,
The first setting means includes:
An autonomous mobile device that notifies a user when the termination condition is satisfied. The first setting means includes:
9. The autonomous mobile device according to claim 8, wherein an operation mode is switched from a first mode in which autonomous movement is possible to a second mode in which the operation mode is moved in accordance with operation of the user in response to the operation of the user. A method applied to an autonomous mobile device that can autonomously move a predetermined space as a search space,
Specifying the search space;
Setting a measure indicating the next action for the state when moving in the search space;
Moving the search space according to the set strategy and collecting 3D modeling information to generate a 3D model;
Setting an end condition to end moving the search space with the set policy;
Switching the policy to a policy different from the currently set policy if the termination condition is satisfied. An autonomous mobile device capable of autonomously moving using a predetermined space as a search space, and a three-dimensional modeling system in which an information processing device operated by a user is communicably connected,
The autonomous mobile device,
Designating means for designating the search space;
First setting means for setting a measure indicating a next action for a state when moving in the search space;
Transmitting means for moving the search space according to the set policy, collecting three-dimensional modeling information for generating a three-dimensional model, and transmitting the collected three-dimensional modeling information to the information processing apparatus; ,
A second setting means for setting an end condition for ending the movement of the search space by the set policy,
The information processing device,
Generating means for receiving the three-dimensional modeling information transmitted from the autonomous mobile device and generating a three-dimensional model for the search space based on the received three-dimensional modeling information;
Evaluating means for evaluating whether or not the generated three-dimensional model is of sufficient quality;
The first setting means includes:
A three-dimensional modeling system that switches the policy to a policy different from a currently set policy when the termination condition is satisfied.

Images