Robot, optimal goal search device and its program

Abstract

To automatically search for an optimum goal position taking surrounding environmental conditions, which may change with time, and contents of a task into consideration when a robot performs the predetermined task involving movement.SOLUTION: A robot 10 includes: an operating unit 11 that operates to enable a predetermined task involving movement in a preset environment; a detection device 12 that detects positional information and speed information of a moving body and a stationary body existing in surroundings; and a control unit 13 that controls an operation of the operating unit 11, based on a detection result of the detection device 12. The control device 13 has an optimum goal search function that searches for an optimum goal to be a destination of the robot 10 most suitable for performing the task from a plurality of points in the environment. In the optimum goal search function, a plurality of points in the environment are set as goal candidates, a total cost obtained by quantifying and totaling a plurality of elements to be considered in performing the task is calculated for each goal candidate, and a goal candidate lowest in the total cost is determined as an optimum goal.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a robot that searches for an appropriate goal position for natural task execution according to the situation of the surrounding environment, an optimal goal search device, and a program therefor. [Background technology]

[0002] Recently, there are increasing expectations for autonomous mobile robots that can provide services in complex environments where humans exist. Some of these robots move autonomously from a predetermined start position to a goal position by sensing the status of surrounding humans and obstacles and moving while avoiding future interference with these humans and obstacles. Robots that move autonomously in an environment where humans coexist are required to not only unilaterally avoid humans, but also to take into consideration human movements and obstacles and to act cooperatively with humans to effectively avoid interference. That is, when moving in such an environment, in addition to passive actions in response to human actions such as interference avoidance actions and stopping actions, it is necessary for the robot to convey its intention to humans through such actions and actively influence the human's own interference avoidance actions. Therefore, as the results of their own research, the present inventors have proposed various robots that move while planning optimal routes based on the above-mentioned influences (see Patent Document 1, etc.). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2020-46773 A Summary of the Invention [Problem to be solved by the invention]

[0004] However, the above-mentioned path planning is based on the premise that one point set by a human being in advance is set as the goal position. However, in a robot for performing a specific task, the specific task may be performed at a position other than the one point, and the robot may be caused to perform unnatural movements due to changes in the environment over time because of the insistence on such a one-point goal position. For example, when the robot's task is to take a picture of a poster displayed in a hall in a crowd with a camera and present the video at a remote location, the following inconveniences arise. That is, for example, when a point at a predetermined distance from the front of the poster is set as the goal position of the robot, if a human is at the goal position at a certain point, the robot must wait until the human is no longer there. Furthermore, the robot can move to the goal position by encouraging the human to move, but this may cause discomfort to the human. In this regard, if the objective is to perform the task of appreciating (taking a picture of) a poster, it is effective to change the goal position of the robot as long as the poster can be taken. Therefore, for autonomous mobile robots performing specified tasks, a new method is needed to automatically identify an appropriate goal position depending on the location of humans in the environment, their movement status, and the content of the task.

[0005] The present invention has been devised with an eye on these problems, and its purpose is to provide a robot, an optimal goal search device, and a program for the same that can automatically search for an appropriate goal position taking into account the surrounding environmental conditions and the content of the task, which change over time, when the robot performs a specified task involving movement. [Means for solving the problem]

[0006] In order to achieve the above-mentioned object, the present invention provides a robot that mainly comprises an operating unit that operates to perform a predetermined task involving movement within a preset environment, a detection device that detects position information and speed information of moving objects and stationary objects in the vicinity, and a control device that controls the operation of the operating unit based on the detection results of the detection device, in which the control device has an optimal goal search function that searches for an optimal goal that is the destination of the robot that is optimal for performing the task from a plurality of points within the environment, and the optimal goal search function sets a plurality of points within the environment as goal candidates, quantifies a plurality of elements to be considered when performing the task, calculates a combined total cost for each of the goal candidates, and determines the goal candidate with the lowest total cost as the optimal goal. Effect of the Invention

[0007] According to the present invention, when a robot is executing a specified task involving movement, it can automatically search for a flexible and natural destination that takes into account changes in the surrounding environment and considers people and others in the environment, thereby enabling the robot to operate in harmony with surrounding people and others. [Brief description of the drawings]

[0008] [Figure 1] 1 is a block diagram that shows only a schematic overall configuration related to the motion control of a robot according to an embodiment of the present invention. [Diagram 2] (A) is a conceptual diagram of a cost map, (B) is a conceptual diagram similar to (A) showing the distribution of task costs in the environment, (C) is a conceptual diagram similar to (A) showing the distribution of social norm costs in the environment, and (D) is a conceptual diagram similar to (A) showing the distribution of obstacle costs in the environment. [Diagram 3] 13A to 13E are conceptual diagrams for explaining the yielding patterns between a human and a robot in the yielding pattern derivation unit. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0010] 1 is a block diagram showing only the overall configuration related to the motion control of a robot 10 according to this embodiment. The robot 10 of this embodiment functions as an autonomous robot that moves autonomously from a preset start position to a goal position so as to operate so as to be able to execute a predetermined task involving movement within a preset environment.

[0011] Although not shown in the drawings, in this embodiment, the task is described below as an example in which, in a situation where a poster is displayed on a wall in a hall where a number of people are present, a person moves through the hall, stops in front of the poster, and takes an image of the poster with a camera. Note that the present invention is not limited to this task, and can be applied to other tasks.

[0012] As shown in FIG. 1, the robot 10 is provided with an operating unit 11 consisting of mechanisms and devices configured to enable various operations, a detection device 12 that detects environmental information around the robot 10, and a control device 13 that controls the operation of the operating unit 11 based on the detection results of the detection device 12.

[0013] The operating unit 11 includes devices and mechanisms that serve as a power source for autonomously moving within a predetermined range, and has a known structure for autonomously moving the robot 10 at a predetermined speed along a path from a start position to a goal position in response to an operation command from the control device 13. The configuration of this operating unit 11 is not essential to the present invention, so detailed illustrations and explanations will be omitted.

[0014] The detection device 12 includes a position detection unit 15 that detects position information of humans, objects, etc. present around the robot 10, and a speed detection unit 16 consisting of a computer that detects speed information of the humans, objects, etc. based on the detection results of the position detection unit 15. The detection targets here include stationary objects that are stationary in the environment, such as obstacles like walls and humans, and moving objects that are moving in the environment, such as humans, other robots, and animals.

[0015] In this embodiment, it is assumed that there are posters and obstacles such as walls that are the targets of the task in the environment, as well as multiple people who are trying to perform the same task. It is assumed that the movements of these people change over time, such as moving or stopping in front of a poster. Therefore, the stationary objects are assumed to be obstacles such as posters and walls, as well as people who stay at a specified location for a specified time. On the other hand, the moving objects are assumed to be people who move to a destination to perform a task, etc.

[0016] The position detection unit 15 is not particularly limited, but is composed of known sensors including a range sensor such as a known laser range finder and a depth sensor that uses camera information. These sensors acquire the position coordinates of a person, object, etc. at the detection time, and determine whether or not it is the same moving object based on time-series data. Then, for the same moving object, the speed detection unit 16 obtains speed information from the change in the acquired time-series position information.

[0017] In the speed detection unit 16, the speed information of the moving object around the robot 10 is calculated from the position information each time the position information of the moving object around the subject and the robot 10 is acquired by a method already proposed by the present inventors (JP Patent Publication No. 2020-91235), etc. Note that the algorithm for deriving the speed information is not an essential element of the present invention, so a detailed description thereof will be omitted.

[0018] The detection device 12 is not limited to the above-mentioned embodiment, and various sensors, devices, systems, etc. can be applied as long as they can detect position information and speed information of moving objects and stationary objects around the robot 10.

[0019] As described above, the position information and speed information of humans, obstacles, etc. present within a specified range around the robot 10 are detected sequentially by the detection device 12 at specified time intervals and transmitted to the control device 13 at the detection timing, and the following processes are performed in the control device 13 at each detection timing.

[0020] The control device 13 is provided integrally with the robot 10 or separately, and is configured by a computer including a processor such as a CPU and storage devices such as a memory and a hard disk. Programs for causing the computer to function as each of the following means are installed in the computer.

[0021] In the control device 13, multiple candidate goals that are candidates for the goal position of the robot 10 are extracted from multiple points in the environment so that the robot 10 can execute tasks involving behavioral judgment as if it were a human. After planning a path for the robot 10 at each of these candidate goals, an optimal goal that is the optimal goal position taking into account the content of the path is determined, and an operation command is given to the operating unit 11 so that the robot 10 moves along the path to the optimal goal.

[0022] This control device 13 is an optimal goal searching device having an optimal goal searching function that searches for an optimal goal that is the destination of the robot 10 that is optimal for executing a task from multiple points in the environment. As described in detail below, the optimal goal searching function here is configured to set multiple points in the environment as goal candidates, quantify multiple elements to be considered in executing a task for each goal candidate, calculate a total cost by combining them, and determine the goal candidate with the lowest total cost as the optimal goal.

[0023] The factors in this case are the first factor related to the ease of performing the task, the second factor related to social norms regarding manners and the sense of distance from other people, the third factor related to the distance from obstacles such as walls, the fourth factor related to compromise with other people performing the same task, and the fifth factor related to the effort of moving to the goal position.

[0024] Next, a specific configuration of the control device 13 will be described.

[0025] The control device 13 includes a goal candidate extraction means 18 that extracts multiple goal candidates for the robot 10 when executing a task from multiple points in the environment, taking into consideration the first to fourth elements; an optimal goal determination means 19 that determines an optimal goal after planning a path for the robot 10 when each goal candidate is set as a destination, taking into consideration the fifth element; and an operation command means 20 that commands the operating unit 11 to operate so that the robot 10 operates along the path to the determined optimal goal.

[0026] The goal candidate extraction means 18 includes a cost map generation unit 22 that generates a cost map in which a cost distribution of each point in the environment is recorded, the cost distribution being a combination of the first to third elements to be considered depending on the task when each point is set as a goal position, a first goal candidate selection unit 23 that selects a plurality of points that are goal candidates as first goal candidates based on the cost map, and a concession consideration unit 24 that determines a second goal candidate as a goal candidate taking into consideration the fourth element based on relative movement information between a human and the robot 10 that are moving to perform the same task.

[0027] The cost map generation unit 22 includes a task cost calculation unit 26 which corresponds to the first element and calculates a task cost indicating the ease of executing the task at each point; a social norm cost calculation unit 27 which corresponds to the second element and calculates a social norm cost at each point according to the distance from the person and the distance from the object of the task (poster); an obstacle cost calculation unit 28 which corresponds to the third element and calculates an obstacle cost at each point according to the distance from an obstacle such as a wall; a cost summation unit 29 which stores the cost obtained by adding up the task cost, social norm cost, and obstacle cost for each coordinate of each point; a cost addition review unit 30 which reviews an additional cost to be added to the cost in the cost summation unit 29 for the point toward which the moving person is heading; and a map determination unit 31 which calculates a final cost map reflecting the cost of each point based on the processing results in the cost summation unit 29 and the added cost review unit 30.

[0028] In the following processes, the planar range of the environment in which the robot 10 can move is divided into a square grid, and the process is performed in units of cells, with the central coordinates of each cell constituting the grid being used as a representative value. The resulting cost map is conceptually shown in Fig. 2(A). In each map in Fig. 2, the darker the color, the higher the cost. In each figure, the circle mark indicates the location of the human H in the environment, the square mark indicates the location of the robot R in the environment, and the cross mark indicates the location of the poster.

[0029] The task cost calculation unit 26 calculates the task cost for each point in the environment based on how easy it is to execute the task, and the distance and angle from the target object are factors. That is, in the case of a task of looking at a poster, a position close to the poster from the front is easier to see and has high value as a goal, so the task cost is calculated based on this viewpoint. Note that this task cost is set so that the cost is between 0 and 1 inclusive within the range in which the task can be accomplished, and is greater than 1 in the range in which the task cannot be accomplished. Note that this task cost is changed according to the content of the task.

[0030] In the poster appreciation task of this embodiment, since it is important how clear the image of the poster photographed by the robot 10 is, the task cost is determined based on the resolution of the poster when photographed by the camera (not shown) of the robot 10. Here, the vertical resolution of the photographed poster depends only on the shooting distance, whereas the horizontal resolution is also affected by the shooting angle, so the task cost C is determined based on the horizontal resolution according to the following formula: t is calculated for each cell at each point in the environment. Note that the task cost C t The distribution is illustrated in FIG. 2(B).

number

[0031] The social norm cost calculation unit 27 calculates a social norm cost related to the sense of distance from humans and manners. In general, if a destination is a place that is too close to other humans or where other humans are standing, it will make those humans feel uncomfortable. Also, standing too close to a poster makes it difficult for other humans to see the poster, which is considered bad manners, so the social norm cost is calculated as an index that takes these factors into consideration.

[0032] Specifically, here, the distance cost from a human, which is based on the psychological repulsion that arises depending on the distance from the human, and the distance cost from an object that has a Gaussian distribution centered on the poster, are calculated for each cell at each point in the environment, and the sum of these distance costs becomes the social norm cost.

[0033] First, the distance cost to humans, C h The cost of humans that are the object of the cost calculation at this time is calculated by the following formula. Humans moving at a speed of a certain value (for example, 1.5 m / sec) or more are not considered to remain in the same place, and are therefore excluded from the cost calculation here. Based on the position of other stationary humans, the distance cost C h is required.

number

[0034] Next, the distance cost C p is calculated by the following formula:

number

[0035] And the distance cost to humans C h and the distance cost C p The sum of these is the social norm cost C s The social norm cost C s The distribution is illustrated in FIG. 2(C).

[0036] The obstacle cost calculation unit 28 calculates an obstacle cost related to the sense of distance to an obstacle. Generally, it is natural to keep a certain distance from an obstacle to avoid collision with the obstacle, so the obstacle cost is calculated as an index that takes into consideration not getting too close to an immovable obstacle such as a wall or an obstacle that is dangerous when approached. Obstacle cost C w is calculated for each cell at each point in the environment using the following formula, which is based on the psychological repulsion due to the distance from an obstacle, and is standardized to the range from 0 to 1.

number

[0037] The cost totalizer 29 calculates the task cost C t , social norm cost C s , and the obstacle cost C w The total cost is calculated by multiplying each of the cells by a weighting constant and adding them together, and a basic cost map is generated in which the total costs are stored corresponding to each cell. This basic cost map is shown in FIG. 2A, which is a superposition of FIG. 2B to FIG. 2D. In the following explanation, the cost of each cell in the basic cost map is referred to as the "cost within the map."

[0038] The cost addition review unit 30 assumes that a moving human performs the same task as the robot 10, and calculates an additional cost to be added to the in-map cost calculated by the cost totalization unit 29 for cells related to the movement of the human.

[0039] That is, the cost addition review unit 30 includes a predicted goal estimation unit 33 that estimates the predicted goal of a human being recognized as moving, and a cost addition unit 34 that adds an additional cost to the in-map cost of the cell in which the predicted goal of the human is located.

[0040] The predicted goal estimation unit 33 targets each cell existing in the direction of the speed of the person recognized as moving, and for each cell, the cell with the lowest cost among the in-map costs calculated by the cost totalization unit 29 is set as the predicted goal of the person.

[0041] In addition, in this predicted goal estimation unit 33, for example, when the robot 10 is currently stationary at a predetermined position such as in front of a poster, the in-map cost calculated by the cost totalization unit 29 is adjusted. That is, in this case, it is assumed that a human is not heading toward the position of the robot 10, and the in-map cost is calculated by further adding a predetermined cost value within a certain range around the cell of the position of the robot 10.

[0042] In the cost adding unit 34, the distance cost from the human centered on the predicted goal of the human is calculated as an additional cost, and for each cell in a predetermined range centered on the predicted goal, the additional cost here is added to the in-map cost. The distance cost here is calculated using the same formula as the distance cost from the human calculated by the social norm cost calculating unit 27, assuming that the human is currently stationary at the predicted goal, and is then multiplied by a preset coefficient.

[0043] In the map determination unit 31, for the cells targeted by the cost addition unit 34, the cost after adding the additional cost calculated by the cost totalization unit 29 is adopted, and for the other cells, the in-map cost is adopted, thereby generating a final cost map in which each cell is associated with a cost value.

[0044] In the first goal candidate selection unit 23, taking into consideration the calculation processing speed and the like, a plurality of cells to be the first goal candidates are extracted from the final cost map determined by the map determination unit 31. Specifically, here, if the cost of a specific cell is lower than the costs of its surrounding eight cells, the specific cell is extracted as a minimum-value cell, and a specific number (for example, up to five) of minimum-value cells are selected as the first goal candidates in order of lowest cost.

[0045] In considering an optimal goal that takes into consideration surrounding people performing the same task, the concession consideration unit 24 assumes that the robot 10 aims for the predicted goal of the moving human, and then extracts a second goal candidate that takes concession into consideration as the destination of the robot 10 when concession is made at the predicted goal, based on the arrival time difference with the human and the cost distribution. In the following explanation of the concession consideration unit 24, the position (coordinates) of the predicted goal of the moving human is referred to as the "concession point."

[0046] This concession consideration unit 24 includes a priority calculation unit 36 ​​which determines the priority in concession between the robot 10 and a human; a cost upper limit calculation unit 37 which calculates the upper limit of the cost acceptable to the high priority side; a concession pattern derivation unit 38 which derives multiple concession patterns; an optimal pattern determination unit 39 which determines an optimal pattern from among the patterns based on the priority and the in-map cost in the basic cost map; and a second goal candidate selection unit 40 which selects the position of the robot 19 after concession in the optimal pattern as a second goal candidate based on the result of the determination by the optimal pattern determination unit 39.

[0047] The priority calculation unit 36 ​​calculates the priority P of a moving human based on the arrival time difference between the human and the robot 10 at the predicted goal estimated by the predicted goal estimation unit 33, using the following formula: H and the robot's priority P R Here, the action of the one that arrives at the predicted goal earlier is set to have priority.

number

[0048] The cost upper limit calculation unit 37 calculates the priority P of the high priority side and the in-map cost C of the cell of the compromise point by the following formula so that the larger the difference in priority is, the smaller the amount of cost increase allowed on the high priority side is. tg Using this, the cost upper limit C L is calculated.

number

[0049] In the yielding pattern derivation unit 38, in a situation where a human H and a robot 10 are heading for a yielding point G as shown in Fig. 3(A), the yielding is assumed to occur in a direction perpendicular to the direction from the yielding point G toward the poster, and a pattern is derived in which one or both of the human H and the robot 10 move so that the distance between them is equal to the personal space PS. Because the calculation is performed in units of cells, for example, in the situation of Fig. 3(A), four patterns shown in Fig. 3(B) to (E) are derived.

[0050] In the optimum pattern determination unit 39, for each yielding pattern, the in-map cost in the cell after yielding on the high priority side is calculated based on the cost upper limit C L and the weighted sum of both costs, C, calculated by the following formula cp The pattern for which is minimal is determined to be the optimal pattern.

number

[0051] In the second goal candidate selection unit 40, the moving position of the robot 10 after the concession in the optimal pattern is added as a second goal candidate taking the concession into consideration. At this time, as the cost of the second goal candidate, the following risk cost is added to the in-map cost of the cell of the moving position of the robot 10 after the concession in the optimal pattern. In other words, even if the robot 10 judges the concession to be appropriate, it is not certain that humans will follow it. Therefore, in order not to force a concession that humans will not follow, when the moving position of the robot 10 after the concession is set as the second goal candidate, the cost of the risk of concession is added to the in-map cost in the basic cost map.

[0052] The risk cost here is C r is calculated by the following formula:

number

[0053] In the above-described goal candidate extraction means 18, a situation cost that quantifies factors based on environmental information to be considered when performing a task is specified for each of the first and second goal candidates. For the first goal candidate, the cost value in the final cost map corresponding to the cell of the point is set as the situation cost, and for the second goal candidate, the cost value after adding the risk cost in the second goal candidate selection unit 40 is set as the situation cost.

[0054] The optimal goal determination means 19 includes a route generation unit 42 that generates route candidates with the first and second goal candidates extracted by the goal candidate extraction means 18 as destinations, a travel cost calculation unit 43 that calculates a travel cost that quantifies the fifth element related to the ease of travel to each route candidate generated by the route generation unit 42, and a final determination unit 44 that determines the optimal goal based on a total cost that is a combination of the situation cost at each goal candidate and the travel cost.

[0055] The path generating unit 42 generates multiple path candidates for the robot 10 using a method disclosed in JP 2020-46773 A and the like already proposed by the present inventors. That is, when it is determined from the detection result of the detection device 12 that interference with a human, object, etc. may occur in the future, a way point (hereinafter referred to as "WP") of a path that avoids interference with the human, object, etc. is set on both the left and right sides of the human, object, etc. Furthermore, on another straight path connecting the WP and the destination with a straight line, when interference with another human, object, etc. may occur in the future, the next WP is set in the same manner on both the left and right sides of the other human or object, and this process is repeated. Finally, multiple path candidates that pass through each WP set in this procedure in order are generated.

[0056] The travel cost calculation unit 43 calculates a travel cost C for each of the route candidates derived by the route generation unit 42 using the following formula similar to the method disclosed in JP 2020-46773 A, etc. m is calculated.

number

[0057] In the final decision unit 44, the goal candidate with the smallest total cost, which is the sum of the situation cost and the movement cost for each goal candidate, is determined as the optimal goal, and the route candidate corresponding to the movement cost at that time is determined as the determined route to the optimal goal. Note that when calculating the total cost, the situation cost and the movement cost may be weighted before being added.

[0058] The motion command means 20 issues a motion control command to the motion unit 11 so that the robot 10 moves along the determined path to the optimal goal determined by the final determination unit 44 .

[0059] In the above-described control device 13, the above-described processes for determining the optimal goal of the robot 10 and deriving the determined path are repeatedly performed at predetermined time intervals (for example, every 0.5 seconds).

[0060] In the case of a task that cannot be completed by traveling to a destination in one go, it is a good idea to assume multiple subgoals, perform the above-mentioned processes for each subgoal, add up the total costs of each subgoal, and find the combination of subgoals with the lowest cost after the additions, as well as the decision route connecting the subgoals in that combination.

[0061] Furthermore, the present invention is not limited to the algorithm described in the above embodiment, and various algorithms can be used as long as they quantify the factors to be considered when executing a task for each goal candidate as a cost, and determine the goal candidate with the lowest cost as the optimal goal.

[0062] Furthermore, the robots that are the subject of the present invention are not limited to the autonomous mobile robot 10 described in each of the above embodiments, but may be mobile bodies that can move autonomously within a specified space, such as automobiles, ships, and aircraft, as well as manipulators such as robot arms that operate within a specified range of space.

[0063] In addition, the configuration of each part of the device in the present invention is not limited to the illustrated configuration example, and various modifications are possible as long as they provide substantially the same effect. [Explanation of symbols]

[0064] 10. Robot 11 Operating unit 12 Detection device 13 Control device (optimal goal search device) 18 Goal candidate extraction method 19 Optimal goal determination method 22 Cost map generator 23 First goal candidate selection section 24. Concession Review Department 36 Priority calculation unit 26 Concession pattern derivation section 39 Optimal pattern determination section 40 Second goal candidate selection section 42 Route Generation Unit 43 Travel cost calculation unit 44 Final Decision Section

Claims

1. A robot comprising: an operating unit that operates to be able to execute a predetermined task involving movement within a preset environment; a detection device that detects position information and speed information of moving objects and stationary objects existing in the vicinity; and a control device that controls the operation of the operating unit based on a detection result of the detection device, the control device comprises: a goal candidate extraction means for extracting a plurality of goal candidates which are candidates for a goal position of the robot when executing the task from a plurality of points in the environment, and calculating a situation cost which quantifies an element based on environmental information to be considered when executing the task for each of the goal candidates; and an optimal goal determination means for planning a path of the robot when each of the goal candidates is set as a destination, and then determining an optimal goal which is the goal position which is optimal for executing the task; The goal candidate extraction means includes a cost map generation unit that generates a cost map in which a cost distribution of each of the points is recorded when each of the points is set as the goal position by using a pre-stored formula, a first goal candidate selection unit that selects a plurality of the points as first goal candidates based on the cost map, and a concession examination unit that determines a second goal candidate taking into consideration a concession between the moving body and the robot, the moving body and the robot moving to perform the same task, based on relative movement information between the moving body and the robot, the yielding examination unit comprises a priority calculation unit which determines a priority in yielding between the robot and the moving body, a yielding pattern derivation unit which derives a plurality of yielding patterns, an optimum pattern determination unit which determines an optimum pattern from among the patterns based on the priority and a map cost in the cost map, and a second goal candidate selection unit which sets the position of the robot after yielding in the optimum pattern as the second goal candidate, The robot is characterized in that the optimal goal determination means determines, from among the goal candidates, the goal candidate having the lowest total cost, taking into account the situation cost and a movement cost related to the effort of moving to each of the goal candidates, as the optimal goal.

2. The robot according to claim 1, characterized in that the second goal candidate selection unit sets the situation cost as a value obtained by adding a risk cost corresponding to the risk that the moving body does not perform a yielding behavior to the map cost at the location of the second goal candidate.

3. The robot according to claim 2, characterized in that the optimal goal determination means comprises a path generation unit that generates path candidates for the robot for each of the goal candidates, a movement cost calculation unit that calculates the movement cost for each path candidate generated by the path generation unit, and a final determination unit that determines the optimal goal based on the total cost for each of the goal candidates.

4. A device having an optimal goal search function that is used for path planning of a robot that operates while executing a predetermined task involving movement within a preset environment, and searches for an optimal goal that is the destination of the robot that is optimal for executing the task from multiple points within the environment based on position information and speed information of moving objects and stationary objects existing in the environment, a goal candidate extraction means for extracting a plurality of goal candidates which are candidates for a goal position of the robot when executing the task from a plurality of points in the environment, and calculating a situation cost which quantifies an element based on environmental information to be considered when executing the task for each of the goal candidates; and an optimal goal determination means for planning a path for the robot when each of the goal candidates is set as a destination, and then determining an optimal goal which is the goal position which is optimal for executing the task, The goal candidate extraction means includes a cost map generation unit that generates a cost map in which a cost distribution of each of the points is recorded when each of the points is set as the goal position by using a pre-stored formula, a first goal candidate selection unit that selects a plurality of the points as first goal candidates based on the cost map, and a concession examination unit that determines a second goal candidate taking into consideration a concession between the moving body and the robot, the moving body and the robot moving to perform the same task, based on relative movement information between the moving body and the robot, the yielding examination unit comprises a priority calculation unit which determines a priority in yielding between the robot and the moving body, a yielding pattern derivation unit which derives a plurality of yielding patterns, an optimum pattern determination unit which determines an optimum pattern from among the patterns based on the priority and a map cost in the cost map, and a second goal candidate selection unit which sets the position of the robot after yielding in the optimum pattern as the second goal candidate, The optimal goal determination means determines, from among the goal candidates, the goal candidate having the lowest total cost, taking into account the situation cost and the travel cost related to the effort of traveling to each of the goal candidates, as the optimal goal.

5. A program for an optimal goal search device that is used for path planning of a robot that operates while executing a predetermined task involving movement within a preset environment, and searches for an optimal goal that is the destination of the robot that is optimal for executing the task from a plurality of points within the environment based on position information and speed information of moving objects and stationary objects existing in the environment, a goal candidate extraction means for extracting a plurality of goal candidates which are candidates for the goal position of the robot when executing the task from a plurality of points in the environment, and calculating a situation cost which quantifies an element based on environmental information to be considered when executing the task for each of the goal candidates; and an optimal goal determination means for planning a path for the robot when each of the goal candidates is set as a destination, and then determining an optimal goal which is the goal position which is optimal for executing the task; The goal candidate extraction means includes a cost map generation unit that generates a cost map in which a cost distribution of each of the points is recorded when each of the points is set as the goal position by using a pre-stored formula, a first goal candidate selection unit that selects a plurality of the points as first goal candidates based on the cost map, and a concession examination unit that determines a second goal candidate taking into consideration a concession between the moving body and the robot, the moving body and the robot moving to perform the same task, based on relative movement information between the moving body and the robot, the yielding examination unit comprises a priority calculation unit which determines a priority in yielding between the robot and the moving body, a yielding pattern derivation unit which derives a plurality of yielding patterns, an optimum pattern determination unit which determines an optimum pattern from among the patterns based on the priority and a map cost in the cost map, and a second goal candidate selection unit which sets the position of the robot after yielding in the optimum pattern as the second goal candidate, A program for an optimal goal search device, characterized in that the optimal goal determination means determines, from among each of the goal candidates, the goal candidate having the lowest total cost, taking into account the situation cost and the travel cost related to the effort of traveling to each of the goal candidates, as the optimal goal.

Images