JP2007319999A - Robot system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To efficiently perform a plurality of tasks in parallel and to execute within a time limit the task requested as an office robot or a housekeeping robot. <P>SOLUTION: The robot device 1 performing the task accepted by an acceptance means 6 by using internal resources 2 is provided with: the acceptance means 6 for accepting the plurality of the tasks; a task storing means 5 storing actions to execute the tasks and service conditions of the internal resources 2 for performing the tasks in parallel as a plurality of task scenarios; and an execution sequence planning means 9 reading the actions and the service conditions for the accepted task out of the task storing means 5 to decide an execution sequence of the task scenario in relation to the service conditions when accepting the task, and acts according to the execution sequence decided by the execution sequence planning means 9. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a robot apparatus having arm / hand driving means, head driving means, leg driving means, image recognition means, speech means, voice recognition means, and the like, and performs a requested task using them.

In recent years, many service robots such as pet robots and humanoid robots have been developed. They are different from conventional industrial robots that perform only certain movements, and humans request what they need to do from time to time and perform the required tasks. The specific action to be performed to perform the requested mission is generally written in a file called a scenario, and the robot acts according to the scenario. A plurality of scenarios are usually recorded. For example, when a human requests a mission by voice, the robot recognizes the mission by voice recognition, extracts a scenario of the corresponding mission from the database, and acts according to the extracted scenario. The more types of scenarios, the more useful robots can perform various tasks.
In addition, in order to make it possible to flexibly handle a variety of responses, and to make the robot more human-like without getting bored, a large number of behavior description modules are arranged in a tree structure, and the desire and satisfaction according to the acquired external information and internal state For example, Patent Document 1 describes that a behavior that is currently being expressed is preferentially continued in order to select a behavior with a high level and appropriately switch the behavior and to provide a consistent response.

JP 2004-283960 A

  In the above prior art, the robot action selection is alternative, and it is difficult to perform a plurality of tasks at the same time and efficiently perform the tasks. In other words, like a human being, you can chat while walking toward the destination, carry documents in the same direction while carrying documents, cook rice while making curry, etc. It is difficult to do casually everyday.

  An object of the present invention is to perform a plurality of tasks efficiently in parallel, as if they were humans, and perform tasks requested as office robots or house robots within a time limit.

  In order to achieve the above object, the present invention provides a plurality of missions in a reception unit that receives a requested mission, and a robot apparatus that acts using the internal resource as a specific operation for the mission received by the reception unit. The mission storage means for storing a plurality of the behaviors for performing the mission and the use conditions of the internal resources when the mission is performed in parallel as mission scenarios, and the mission And the execution order plan means for reading the behavior and the use condition of the received mission from the mission storage means and determining the execution order of the mission scenario in relation to the use condition, Acting according to the execution order determined by the execution order planning means.

  According to the present invention, a robot apparatus that can perform a plurality of tasks in parallel and can efficiently perform work is provided. In addition, since the client can know whether or not the requested mission can be performed and when it is started, the robot apparatus is highly reliable for the client. Further, even if a large number of jobs are requested at once, the specific mission can be more reliably performed, so that the robot apparatus is more reliable for the client.

Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.
FIG. 1 is a diagram showing a configuration of a robot apparatus capable of performing a plurality of tasks in parallel. Reference numeral 1 denotes a robot apparatus body, and 2 denotes an internal resource, which means various driving means, recognition means, and sensors of the robot. In this embodiment, a humanoid robot is assumed, and an arm, hand driving means, head driving means, Leg driving means, speech (display) means, speech recognition means, and the like.
The arm / hand driving means includes an arm / hand for gripping an object or performing a gesture, an actuator for driving them, and an arm / hand control means for performing an intended operation. The head drive means consists of a head that resembles a human head, an actuator that drives the head, and control means. The head is swayed, crawled, or directed at a topical object. To convey the intention of the robot. In addition, it incorporates a camera and an image recognition device to recognize the surrounding environment and identify the client.

The leg driving means is a walking leg or a wheel leg, and moves the robot apparatus to a target position. The utterance (display) means is composed of a speech synthesizer and a speaker, and conveys the intention of the robot apparatus to surrounding people by performing utterances like a human or screen display. The voice recognition means includes a microphone and a voice analysis unit, and recognizes the content of human speech and judges the surrounding situation. Reference numeral 3 denotes an external resource, which is an object handled by an arm / hand drive means that is an internal resource.
In this embodiment, a tray, a hot water bath, a kitchen knife, a cutting board, and the like are assumed. 4 is a client who requests the robot apparatus 1 for a task. In the present embodiment, the client 4 transmits the mission contents to the robot apparatus 1 by voice, and the robot apparatus 1 recognizes the mission contents by voice recognition means. 5 is a mission storage means. It stores mission scenarios that describe the specific behavior of missions and the terms of parallel mission resource usage. For example, the mission scenario is described in a text file, and the mission storage means 5 uses a hard disk.

  Reference numeral 6 denotes a reception unit that constantly monitors the voice recognition unit to determine whether the utterance voice of the mission request has been recognized. Reference numeral 7 denotes a request result response means. The requester 4 is notified of whether or not the requested mission can be performed and the scheduled start time to the client 4 using the utterance means and the head drive means. For example, “Please wait for a few minutes” or “Your request is not accepted because you are now full.” Shake sideways.

  A multi-task scenario execution unit 8 reads a scenario of a task whose execution has been confirmed from the task storage unit 5, and drives various units of the internal resource 2 according to the action details written in the scenario. 9 is a planning means. Plan action execution order to accomplish multiple missions. Reference numeral 10 denotes a control means. The various means in the robot apparatus 1 are controlled.

  Next, examples of mission scenario description contents will be described with reference to FIGS. Each figure describes details of each mission, FIG. 2 guides to the office, FIG. 3 guides to the nearest station, FIG. 4 guides to the toilet, FIG. 5 conveys the weather forecast, FIG. 6 is serving tea, FIG. 7 is carrying documents, FIG. 8 is making curry, FIG. 9 is making salad, FIG. 10 is receiving the package of parcel delivery. The description items of the mission scenario are explained below.

The mission name is a name representing the content of each mission. The standard deadline is the deadline for completing the mission. Used when the client has not specified a deadline. The priority is used when the execution order planning unit 9 determines which mission should be preferentially performed when planning the action execution order, and is fixed according to the type of mission. The normal mission is defined as “ordinary”, the response of important customers is “high”, and the emergency is defined as “best”. The execution order planning means 9 determines the tasks to be preferentially executed as follows.
If a request for a mission with a high priority is received while a mission with a priority of “normal” is being executed, the mission with a priority of “normal” can be interrupted and the mission with a high priority is started. . If a request for a mission with a priority of “normal” is received while performing a mission with a priority of “high”, the mission with a priority of “normal” is postponed. When a mission request with the highest priority is received, all the missions currently being executed are interrupted and the mission with the highest priority is started. In addition to priority, priority is also used as information for determining execution priority. Priorities are determined based on the order of requests. As will be described later, among missions that should be performed within the same time, the earlier the request time, the higher the priority is set, and the higher the priority, the higher the priority. FIG. 19 shows a list of priorities and priorities.

  In the mission scenario, a series of actions are separated by time and each one is named an action. In the action column, a serial number assigned in time series and a comment indicating the contents of each action are described. In the scheduled time column, the scheduled time required to execute each action is described. In the resource name column, the internal resource name of the robot apparatus 1 and the external resource name used for each action are described. In the use range column, describe the use range of each resource in each action. Resources that are not used are described as “not used”. When the entire resource is used, it is described as “whole”. In the case of the leg driving means, the standing position and the moving range are described. For example, in FIG. 2, “client's front fixation” is described as the standing position, and “from the entrance to the office” is described as the movement range. The operation detail column describes the detailed movement of each resource. Actually, information necessary for driving each internal resource, such as a driving angle and an utterance wording, is accurately described in detail.

  The parallel condition column is divided into three items. The other mission simultaneous use column describes whether or not other missions can be used at the same time when the use of other resources becomes necessary while using each resource. As an example that can be used at the same time, FIG. 2 describes that the use of the legs is “possible on the same route”, which indicates that the missions going in the same direction can be performed simultaneously. The interruption condition column describes whether or not the use of the resource can be interrupted and used for other missions when the necessity of use occurs in other missions while using each resource. For example, in FIG. 8, it is described as “possible within 2 minutes. Fire extinguishing when interrupting, ignition when resuming.” During the frying of curry ingredients, it can be interrupted for a short time within 2 minutes. This indicates that the fire should be extinguished when interrupted, and fired when resuming. The omitted column describes whether or not the use of each resource can be omitted when the action is started. For example, the omission of the head driving means in action No. 1 in FIG. 2 indicates that when asking the client for a place to visit, the other party should be looked at, but looking away is also possible.

  Next, the flow of operation of the robot apparatus will be described with reference to FIGS. FIG. 11 is a flowchart showing the flow of operations at the start of a new mission. FIG. 12 is a flowchart showing the flow of operations at the start of action.

  The accepting means 6 starts waiting for a mission request (S1). If there is a mission request (S2), the client's face is stored by the image recognition means built in the head (S3). Thereafter, the execution order planning unit 9 checks the required time, parallel condition, and time limit of the actions after the present of all the missions currently being executed from the corresponding mission scenario of the mission storage unit 5 (S4). The time required for each action of the newly requested mission, the parallel condition, and the time limit are also checked (S5). Then, using these pieces of information, a multi-task execution order described later is planned (S6). As a result, if it is possible to start immediately (S7), it is determined whether or not the client is waiting (S8), and if not, it is necessary to interrupt the task being executed as described later. (S9), and if an interruption is necessary, the action of the mission being executed is interrupted (S10). Then, a new mission is started from action No. 1 (S11). At this time, the action start time is stored. Then prioritize the tasks you are currently performing. The earlier the time when the mission request is received, the higher the priority is given (S12).

  If it is not possible to start immediately in S7, it is determined whether it can be started later (S13). If it can be started later, a new mission start schedule is set (S14). If you can speak (S15), say "Please wait a bit" or "You can start in a few minutes", etc. (S16), if you can talk (S17), give a talk (S18), request Tell people when to start. Since the start time of the action currently being executed is stored in S11, the waiting time from the start time and the scheduled required time to the new task start time can be calculated.

  As a method of setting a new mission start schedule in S14, a new mission start timer may be multiplied to the scheduled end time of the currently executing action, or a new mission will be started upon receiving a completion event of the currently executing action. It may be. If it is determined in S19 that a new mission start timer has arrived or a previous action completion event has been received, a new mission start is scheduled. Then, if it is possible to start immediately as a result of planning a plurality of mission execution orders, the process arrives at S8. In S13, if the mission that is determined to be started later is to be started, the client must be waiting, and the process proceeds to S20. If there is still a client memorized in S3 (S20), if the utterance is possible (S21), then "send me" is uttered (S22), if it is possible to talk (S23), let me talk (S24) .

If it is not possible to start in S13, if the utterance is possible (S25), make the utterance such as "I can't accept it now" or "I can't do it within 20 minutes", and the client can't perform the mission (S26). When the head can be swung (S27), the head is swung sideways (S28).

When there is no mission request in S2 and there is no plan to start a new mission in S19, the operation shifts to the action start operation shown in FIG.
It is determined whether there is any completed action between the time when the previous arrival at S31 and the arrival at this time (S31). If there is a completed action, look into the multiple mission execution plan,
It is determined whether there is a mission to be resumed among the tasks suspended in S10, and if there is a mission that needs to be resumed, it is set that the mission is scheduled to start (S33). If the completed action is the last action of the mission, the mission is completed (S34). Therefore, the priority is updated to the order from which the priority of the mission is removed (S35), and the process returns to S2. If there is a next action instead of the mission completion, the time required for each action after that, the parallel condition, and the time limit are checked (S36).

For all missions currently being carried out, the time required for the action after the present, the parallel condition and the deadline are also examined (S37). A plurality of mission execution orders are planned using the information (S38). As a result, if the next action can be started immediately (S39), it is determined whether or not it is necessary to interrupt the mission being executed as described later (S40). Is interrupted (S41). Then, the action is started (S11). At this time, the action start time is stored.
If it is not possible to start immediately in S39 and it is started later (S43), the fact that the action is scheduled to start is set (S44). If it is impossible to start in S43, if the utterance is possible (S46), the utterance "cannot continue the mission" is uttered (S47). However, at this time, since the requester is not always nearby, a contact telephone number or e-mail address may be acquired in advance, and it may be notified that the mission cannot be continued by telephone or e-mail.

  Next, with reference to FIG. 13, a method for planning the order of performing multiple missions in S6 and S38 will be described. The start time and time limit of the new task in S6, the task scheduled to start the next action in S38, and the start time change task (hereinafter referred to as A) to be described later are obtained (S200). The start time is the start time when designated by the client, or the request time when there is no request. If there is a designation from the client, use the deadline, otherwise use the standard deadline described in the mission scenario. When the completion time is designated by the client, the start time is set to the time obtained by subtracting the sum of the time required for all actions from the completion time.

Thereafter, the actions of A after the present are arranged one by one in time series while making the following judgment (S101). First, it is determined whether the required resource is being used by another task (hereinafter referred to as B) (S102). If it is in use, it is checked whether the resource can be used simultaneously (S103). If it can be used simultaneously, action A is placed in parallel with action B (S104). If they cannot be used simultaneously, it is checked whether the resource A can be omitted (S105). If it can be omitted, the resource is omitted and the action of A is arranged in parallel with B (S106). If it cannot be omitted, it is determined whether the priority is A> B and B can be interrupted (S107). In the case of YES, it is assumed that the action of B is interrupted and the action of A is immediately executed. When this plan is executed, interruption processing is performed in S10 and S41. In the case of NO in S107, the action A is arranged after the action B.

In the case of NO in S102, it is determined whether the required resource is scheduled to be used at the same time for another mission (hereinafter referred to as C) (S114). If it is scheduled to be used, it is checked whether the resource can be used simultaneously (S115). If it can be used simultaneously, action A is placed in parallel with action C (S116). If they cannot be used simultaneously, it is checked whether the resource A can be omitted (S117). If it can be omitted, the resource is omitted and the action of A is arranged in parallel with C (S118). If it cannot be omitted, it is determined whether the priority is A> C (S119). In the case of YES, the action of C is rearranged after the action of A (S120). If rearrangement of C is impossible, a multi-task action execution order plan is performed for C after the end of arrangement of A. That is, the sequence of FIG. 13 is performed with C's mission as A in S200. In the case of NO in S119, the action of the mission with the lower priority is placed after the action of the mission with the higher priority of either A or C (S121).

After performing the above determination and arrangement, it is checked whether the mission can be completed within the time limit (S110). If it is within the time limit, it is determined whether all actions have been arranged (S111), and if there is an action that has not been arranged, the process returns to S101. If the arrangement of all actions has been completed, it is determined that the action of mission A can be started (S112). If it does not fall within the time limit in S110, the task whose start time can be changed is checked (S201). If there is, the start time of the mission is changed (S202), and the execution order plan is performed again for the mission and the A mission. If there is no mission whose start time can be changed in S201, it is determined that it is impossible to start the action of the A mission (S113).

  An example in which the mission scenarios shown in FIGS. 2 to 10 are executed based on the above-described method of planning the multiple mission execution order will be described with reference to timing charts shown in FIGS. In the figure and the following description, the alphabet indicates the mission, and the number indicates the action No. of the mission.

  FIG. 14 is an example of a guidance mission in the office. First, the robot apparatus is requested to guide to the office at the entrance of the building (task A). While moving in the lobby, you are asked to guide you to the nearest station, but because the guidance route to the nearest station is not on the guidance route to the office, you cannot perform tasks in parallel, and the result of the mission action execution order plan It was determined that it could not be performed by the deadline. Therefore, the refusal utterance "I can't accept it right now. Please ask another robot." (Task D) is performed. After that, a person who passes by in the corridor is asked to guide to the toilet (Task B). Reply while moving, and guide to the restroom on the way to the office. Therefore, the guide mission to the toilet ends. Then, on the way from the toilet to the office, A's client asks the world weather forecast for the next week (Task C). He spoke to the target office while speaking the weather forecast, but continued speaking the weather forecast and only told us that he had arrived. Thus, it is possible to perform a plurality of tasks in parallel.

  FIG. 15 is an example of guiding important customers to the office. For important customers, set all office work priorities to “high”. First, a request from an important customer to the office is requested (task A). After that, even if a passer-by asks for guidance to the toilet, the task with low priority is postponed and cannot be performed within the time limit, so the request is ignored. When a demand customer requests a weather forecast (task C), the weather forecast is transmitted because the priorities are equal. After that, when a fire occurs, the priority of evacuation is “highest”, so the guidance to the office is interrupted and evacuated with an important customer (task D). In this way, the possibility of interruption or parallel execution is determined by the high and low priority.

  FIG. 16 is an example of serving tea. A tea is requested (task A). An important customer is asked to carry documents while he or she is drinking tea and carrying it to the reception room and putting a cup of tea in front of the customer (Task B). Since it is a request of an important customer, it has a high priority, but since it has tea, it is strictly prohibited to stop using the hand immediately. However, the distribution of teacups is scheduled to end soon, and it is judged that document transportation can be carried out within the deadline, so he says, "I will accept it soon. Please wait." After that, when the cups are distributed, the document handling mission is started.

  FIG. 17 and FIG. 18 are examples in which curry and salad are completed simultaneously at a specified time. It is requested at 10:00 to make curry and salad to be completed at 12:00. The execution order planning means first obtains a start time obtained by subtracting the required time from the designated completion time. Then, in preparation for an accident in the middle, the start time is advanced five minutes as an allowance time so that the curry does not cool. In curry and salad, curry creation has an earlier creation start time, so each action of curry creation is arranged in time series (FIG. 17A). Then, arrange each action of making salad. However, since the necessary resources for curry creation are used for 10 minutes before the completion time, the salad creation start time is advanced by 10 minutes ((b) of FIG. 17). Wait until the creation start time after planning. When the curry creation time arrives, each action is executed according to the plan (FIG. 18). However, the parcel delivery arrived while the curry ingredients were being fried (task C). The priority for receiving the parcel delivery package is “high”. The estimated time required for receipt of parcels is 2 minutes, and the stir-fry action of curry ingredients can be interrupted if it is less than two minutes. Move to the entrance and pick up your parcel bag and return. Then, the stove is turned on and the frying action is resumed. As a result, curry and salad were completed 3 minutes before the designated time. In this way, it is possible to perform a task with a set completion time in the same way as when a person performs housework.

The block diagram which shows the structure of the robot apparatus which concerns on one embodiment by this invention. The figure which shows the example of description content of mission scenario in one Embodiment. The figure which shows the example of description content of the other mission scenario in one Embodiment. The figure which shows the example of the description content of other mission scenarios in one Embodiment. The figure which shows the example of the description content of other mission scenarios in one Embodiment. The figure which shows the example of the description content of other mission scenarios in one Embodiment. The figure which shows the example of the description content of other mission scenarios in one Embodiment. The figure which shows the example of the description content of other mission scenarios in one Embodiment. The figure which shows the example of the description content of other mission scenarios in one Embodiment. The figure which shows the example of the description content of other mission scenarios in one Embodiment. The flowchart which shows the operation | movement at the time of new mission start in one Embodiment. The flowchart which shows the operation | movement at the time of action start in one Embodiment. 6 is a flowchart illustrating a method for determining an execution order according to an embodiment. The timing chart which shows the example of mission performance in one embodiment. The timing chart which shows the other example of mission execution in one Embodiment. The timing chart which shows the further example of mission execution in one Embodiment. The timing chart which shows the further example of mission execution in one Embodiment. The timing chart which shows the further example of mission execution in one Embodiment. The figure which shows the example of the priority and priority of a mission in one Embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Robot apparatus, 2 ... Internal resource, 3 ... External resource, 4 ... Client, 5 ... Mission storage means, 6 ... Acceptance means, 7 ... Request result reply means, 8 ... Multiple mission scenario execution means, 9 ... Execution order Planning means, 10 ... control means.

Claims (5)

In a robot device that operates using internal resources as a specific operation of a reception unit that receives a requested mission, and the mission received by the reception unit,
The accepting means capable of accepting a plurality of missions;
Mission storage means for storing a plurality of the behaviors for performing the mission and the use conditions of the internal resources when performing the mission in parallel as mission scenarios;
When receiving the mission, execution order planning means for reading the behavior and the use condition of the received mission from the mission storage means, and determining the execution order of the mission scenario in relation to the use condition;
The robot apparatus is characterized in that it acts according to the execution order determined by the execution order planning means. The thing according to claim 1, wherein the mission is a plurality of actions separated by time, and the use condition is defined for each action,
When receiving a plurality of the tasks, the execution order includes performing the actions in parallel by comparing internal resources to be used in the use condition for each action, or executing the actions at different times. A robot apparatus characterized by being determined by a planning means.   The robot apparatus according to claim 1, wherein the use condition describes whether or not use can be omitted for each internal resource.   2. The apparatus according to claim 1, wherein said execution order planning means speaks that said task cannot be performed if it is determined that said task cannot be performed within a specified time limit. Robot device. The robot apparatus according to claim 1, wherein the execution order planning unit speaks that the execution start of the mission is delayed when the execution of the mission cannot be started immediately.

Images