Robot control device, robot control method, and program
The robot control device enhances the sense of agency by aligning user control inputs with predicted actions, addressing the imbalance in existing methods by integrating user input and predicted action similarity to ensure accurate robot performance.
Patent Information
- Application Number
- JP2022024177
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2022-02-18
- Publication Date
- 2025-07-25
- Estimated Expiration
- 2042-02-18
AI Technical Summary
Existing robot control methods fail to balance the sense of agency in movement by either ignoring user control inputs or failing to accurately reflect them, leading to a diminished sense of agency in both master-slave and automatic action systems.
A robot control device that acquires first control input information from the user, calculates second control input information based on predicted target actions, and determines the similarity between the two to adjust control inputs, ensuring the robot performs the intended action when the user has an intention to do so.
Improves the sense of agency during robot control by accurately reflecting user intentions and ensuring the robot performs the intended actions, even when user inputs may be inaccurate or untrained.
Smart Images
Figure 0007713196000006 
Figure 0007713196000007 
Figure 0007713196000008
Abstract
Description
Technical Field
[0001] The present invention relates to a robot control device, a robot control method, and a program.
Background Art
[0002] Users have been able to control various robots. For example, Non-Patent Document 1 discloses a technique of inputting data indicating a user's inertia information and surface electromyogram as control information into a robot and operating the robot to perform a user's target motion.
[0003] Also, it has become possible to infer a user's target motion from the user's biological information. For example, Non-Patent Document 2 discloses a technique of estimating a target motion required by a user from the user's brain wave information.
Prior Art Documents
Non-Patent Documents
[0004]
Non-Patent Document 1
Non-Patent Document 2
Summary of the Invention
Problems to be Solved by the Invention
[0005] When controlling a robot by a human, the following two viewpoints are considered to be requirements for the sense of agency in movement.
[0006] Requirement 1: A person's behavior (control input) is reflected in the movement.
[0007] Requirement 2: A person's target movement is performed through the robot.
[0008] Regarding Requirement 1, like the technology described in Non-Patent Document 1, it is necessary to have a master-slave type relationship in which the robot operates obediently to the user's control input. However, in the case of a user who is not used to controlling a robot, there may be a certain deviation between the control input required for the target movement and the actual control input. In such a master-slave type relationship, it is difficult to perform the target movement and Requirement 2 is not satisfied. Therefore, there is a problem of low sense of agency in movement. Also, there is a method of correcting the control input through force feedback assist or the like when the user makes a control input, but the user's spontaneous behavior may become passive and the sense of agency may be lowered.
[0009] On the one hand, the technique described in Non-Patent Document 2 estimates the user's target action, and since the robot can automatically perform it, Requirement 2 is satisfied. However, since it becomes an action system that ignores the user's control input, it does not satisfy Requirement 1, and there is still a problem of low sense of agency.
[0010] The disclosed technique aims to improve the sense of agency during robot control.
Means for Solving the Problem
[0011] The disclosed technique includes a first control input acquisition unit that acquires first control input information input by a user, a second control input calculation unit that calculates second control input information for realizing the target action based on information indicating the target action of the user, and a similarity between the first control input information and the second control input information. calculate the similarity, and the similarity Based on a target operation execution intention determination unit that determines whether or not the user has an intention to perform the target operation, and when it is determined that the user has an intention to perform the target operation, a control information calculation unit that calculates control information for controlling the robot based on the first control input information, the second control input information, and the similarity. It is a robot control device.
Effect of the Invention
[0012] According to the disclosed technique, the sense of agency during robot control can be improved.
Brief Description of the Drawings
[0013]
Figure 1
Figure 2
Figure 3
Mode for Carrying Out the Invention
[0014] Hereinafter, embodiments of the present invention (hereinafter referred to as "the present embodiments") will be described with reference to the drawings. The embodiments described below are merely examples, and the embodiments to which the present invention is applied are not limited to the following embodiments.
[0015] (Overview of Robot Control Device) The robot control device according to the present embodiment acquires first control input information based on data such as the user's surface electromyogram. Then, the robot control device acquires the predicted result of the target operation required by the user, calculates second control input information for realizing the predicted result, and calculates the similarity between the first control input information and the second control input information. The robot control device controls the robot based on the calculated control information according to the calculated similarity.
[0016] Hereinafter, control for operating a robot arm composed of two motors on a two-dimensional plane by two-axis control input will be exemplified. Note that the content of the control is an example, and other controls may be used.
[0017] (Example of Functional Configuration of Robot Control Device) FIG. 1 is a diagram showing an example of the functional configuration of a robot control device. The robot control device 10 includes a target operation acquisition unit 11, a first control input acquisition unit 12, a control operation control coordinate calculation unit 13, a second control input calculation unit 14, a target operation execution intention determination unit 15, a control information calculation unit 16, and a robot control unit 17.
[0018] The target operation acquisition unit 11 acquires information indicating the target operation predicted by an external device or the like. The external device predicts the target operation required by the user through, for example, analysis of electroencephalograms by applying the technology shown in Non-Patent Document 2. Note that the target operation is the operation required of the robot as the purpose for which the user controls the robot. When the user is not accustomed to the control operation, the control may not necessarily be performed as intended.
[0019] The content of the predicted operation is, for example, a reaching operation in which the robot arm approaches an object. Note that the content of the predicted operation is an example, and other operations by robots having other degrees of freedom may also be used.
[0020] The target motion acquisition unit 11 obtains, from the two-dimensional position of the object and the two-dimensional position of the robot arm, the target rotation angles R = (R 1 , R 2 ) of two motors required for the reaching motion to the object as the target motion.
[0021] The first control input acquisition unit 12 acquires control input information (hereinafter referred to as first control input information) from an external device such as a biological signal measurement device that measures biological signals such as data indicating surface electromyogram, or a joystick-type two-axis signal measurement device. That is, the first control input information is information indicating the control input entered by the user. Specifically, the first control input information is the control input information c t = (c 1,t , c 2,t ) acquired at time t.
[0022] The control motion control coordinate calculation unit 13 calculates the control coordinates of the control motion by taking the target rotation angle R = (R 1 , R 2 ) acquired by the target motion acquisition unit 11 as an input.
[0023] Specifically, the control motion control coordinate calculation unit 13 calculates the control coordinates of the control motion to be input for performing the target motion in the control space with the robot arm at the time point (initial state) when the target rotation angle is input from the target motion acquisition unit 11 as the origin. In this control space, the coordinates are determined based on the time-series control input. For example, the control motion control coordinate calculation unit 13 calculates the coordinates (x t , y t ) in the control space at time t as follows.
[0024]
Equation
[0025] Also, according to the characteristics of the robot, the environment, etc., the control coordinates c θ = (x θ , yθ ) and the relationship between the rotation angle θ = (θ 1 , θ 2 ) of each motor and the like is preset based on the following f.
[0026] (θ 1 , θ 2 ) = f(c θ ) Therefore, the control operation control coordinate calculation unit 13 calculates the control coordinates c R = (x R , y R ) of the control operation that realizes the target rotation angle R according to the following formula.
[0027] c R = (x R , y R ) = f -1 (R 1 , R 2 ) The control coordinates c R of the control operation that realizes the target rotation angle R are absolute coordinates in the control space based on the initial state.
[0028] The second control input calculation unit 14 calculates control input information (hereinafter referred to as second control input information) for realizing c R from the control coordinates at each time using the control coordinates c R of the control operation that realizes the target rotation angle R calculated by the control operation control coordinate calculation unit 13 as an input.
[0029] For example, the second control input information for realizing c R from the control coordinates at time t is calculated as follows.
[0030]
Equation
[0031] The target action execution intention determination unit 15 calculates the similarity between the first control input information and the second control input information, and determines the intention to execute the target action. For example, in the case of a user who is not used to controlling a robot, there may be a certain degree of deviation between the first control input information and the second control input. Therefore, the target action execution intention determination unit 15 determines whether there is an intention to actually execute the target action predicted by the external device based on the similarity between the first control input information and the second control input information.
[0032] For example, the target action execution intention determination unit 15 calculates the cosine similarity as follows.
[0033]
Equation
[0034] Then, when the calculated cosine similarity is greater than the threshold value (for example, 0.5), the target action execution intention determination unit 15 determines that there is an intention to execute the target action, and in other cases, determines that there is no intention to execute the target action. Here, the threshold value may be determined in advance for each user, or may be varied according to the usage time, usage frequency, etc. of the robot.
[0035] The control information calculation unit 16 calculates control information for controlling the robot based on the determination result by the target action execution intention determination unit 15, the first control input information, and the second control input information.
[0036] Specifically, when the target action execution intention determination unit 15 determines that there is an intention to execute the target action, the control information calculation unit 16 calculates the control information At as shown in Equation (1).
[0037]
Equation
[0038] Also, when the target operation execution intention determination unit 15 determines that there is no intention to execute the target operation, the control information calculation unit 16 calculates the control information A t as follows.
[0039] A t = c t
[0040] The robot control unit 17 drives and controls two motors based on the control information A t The robot control unit 17 drives and controls two motors based on the control information A.
[0041] The robot control device 10 repeatedly executes the processing of each part described above from t = 1 until a predetermined condition as shown in Equation (2) is satisfied. Note that d is a predetermined constant.
[0042]
Equation
[0043] (Operation of the Robot Control Device) Next, the operation of the robot control device 10 will be described with reference to the drawings.
[0044] Figure 3 is a diagram showing an example of the flow of robot control processing. The robot control device 10 starts robot control processing according to a user's operation or the like.
[0045] The target motion acquisition unit 11 acquires information indicating the predicted target motion by receiving it from an external device or the like (step S101). Next, the first control input acquisition unit 12 acquires first control input information by receiving it from an external device such as a biological signal measurement device or a two-axis signal measurement device (step S102).
[0046] Subsequently, the control motion control coordinate calculation unit 13 calculates the control coordinates of the control motion based on the information indicating the target motion acquired by the target motion acquisition unit 11 (step S103).
[0047] Next, the second control input calculation unit 14 calculates second control input information based on the control coordinates of the control motion (step S104).
[0048] Then, the target motion execution intention determination unit 15 calculates the similarity between the first control input and the second control input (step S105). The target motion execution intention determination unit 15 sets t = 1 and j = 1 (step S106) and determines whether the similarity is greater than the threshold Th (step S107).
[0049] If the target motion execution intention determination unit 15 determines that the similarity is greater than the threshold Th (step S107: YES), it determines that there is an intention to execute the target motion. In this case, the control information calculation unit 16 calculates the control information according to formula (1) (step S108). The robot control device 10 increments t (step S109) and proceeds to step S114.
[0050] Also, if the target motion execution intention determination unit 15 determines that the similarity is not greater than the threshold Th (step S107: NO), it determines that there is no intention to execute the target motion. In this case, the control information calculation unit 16 sets A t =C tCalculate control information thereby (step S110). The robot control device 10 increments j (step S111) and determines whether j is greater than T (step S112). Then, when the robot control device 10 determines that j is greater than T (step S112: YES), it erases the information indicating the target operation (step S113) and proceeds to step S114. Also, when the robot control device 10 determines that j is not greater than T (step S112: NO), it skips the process of step S113 and proceeds to the process of step S114.
[0051] Note that T is a threshold value set in advance, and is a value set as the upper limit of the number of times when the target operation is erased when it is continuously determined that there is no intention to perform the target operation. The threshold value T may be set when reasons such as low prediction accuracy of the target operation or an abnormality occurring in the prediction process of the target operation are considered. Note that the robot control device 10 may skip the processes of step S112 and step S113 when the threshold value T is not set.
[0052] Next, the robot control unit 17 controls the robot based on the calculated control information (step S114). Then, the robot control device 10 determines whether the formula (2) is satisfied (step S115). When the robot control device 10 determines that the formula (2) is not satisfied (step S115: NO), it returns to the process of step S107.
[0053] Also, when the robot control device 10 determines that the formula (2) is satisfied (step S115: YES), it ends the robot control process.
[0054] (Hardware configuration example according to this embodiment) The robot control device 10 can be realized, for example, by causing a computer to execute a program describing the processing contents described in this embodiment. Note that this "computer" may be a physical machine or a virtual machine on the cloud. When using a virtual machine, the "hardware" described here is virtual hardware.
[0055] The above program can be recorded on a computer-readable recording medium (such as a portable memory), saved, and distributed. It is also possible to provide the above program through a network such as the Internet or email.
[0056] FIG. 3 is a diagram showing an example of the hardware configuration of the above computer. The computer in FIG. 3 includes a drive device 1000, an auxiliary storage device 1002, a memory device 1003, a CPU 1004, an interface device 1005, a display device 1006, an input device 1007, an output device 1008, etc., which are mutually connected by a bus B.
[0057] A program for realizing the processing on the computer is provided, for example, by a recording medium 1001 such as a CD-ROM or a memory card. When the recording medium 1001 storing the program is set in the drive device 1000, the program is installed from the recording medium 1001 via the drive device 1000 into the auxiliary storage device 1002. However, the installation of the program does not necessarily have to be performed from the recording medium 1001, and it may be downloaded from another computer via a network. The auxiliary storage device 1002 stores the installed program and also stores necessary files, data, etc.
[0058] When there is an instruction to start a program, the memory device 1003 reads and stores the program from the auxiliary storage device 1002. The CPU 1004 realizes the functions related to the device according to the program stored in the memory device 1003. The interface device 1005 is used as an interface for connecting to a network. The display device 1006 displays a GUI (Graphical User Interface) or the like according to a program. The input device 1007 is composed of a keyboard, a mouse, buttons, or a touch panel, etc., and is used to input various operation instructions. The output device 1008 outputs the calculation result. Note that the above computer may be provided with a GPU (Graphics Processing Unit) or a TPU (Tensor processing unit) instead of the CPU 1004, or may be provided with a GPU or a TPU in addition to the CPU 1004. In that case, the GPU or the TPU may execute the processing that requires special calculations, and the CPU 1004 may execute the other processing, and the processing may be shared and executed in this way.
[0059] (Effect of the present embodiment) According to the robot control device 10 according to the present embodiment, based on the similarity between the first control input information acquired based on the measurement result of the user's biological signal or the like and the second control input information calculated based on the predicted target motion, it is determined whether there is an intention to perform the target motion, and the robot is controlled based on the control information calculated according to the determination result. Thereby, even when the first control input information is such that the user's intended motion cannot be performed, the robot can be controlled to perform the target motion based on the similar second control input information. Therefore, the sense of the moving body during robot control can be improved.
[0060] (Summary of the embodiment) This specification describes at least the robot control device, the robot control method, and the program described in each of the following items. (Item 1) A robot control device for controlling a robot, A first control input acquisition unit that acquires first control input information input by a user; A second control input calculation unit that calculates second control input information for realizing the target operation based on information indicating the target operation of the user; A control information calculation unit that calculates control information for controlling the robot based on the similarity between the first control input information and the second control input information, comprising: A robot control device. (Item 2) Further comprising a target operation execution intention determination unit that determines whether the user has an intention to perform the target operation based on the similarity between the first control input information and the second control input information; When it is determined that the user has an intention to perform the target operation, the control information calculation unit calculates control information obtained by combining the first control input information and the second control input information with weighting based on the similarity; The robot control device according to Item 1. (Item 3) When the similarity is greater than a threshold value, the target operation execution intention determination unit determines that the user has an intention to perform the target operation; The robot control device according to Item 2. (Item 4) When it is determined that the user has no intention to perform the target operation, the control information calculation unit uses the first control input information as the control information; The robot control device according to Item 2 or Item 3. (Item 5) The target operation execution intention determination unit and the control information calculation unit repeatedly execute the process until a preset condition is satisfied; The robot control device according to any one of Items 2 to 4. (Item 6) A robot control method executed by a robot control device for controlling a robot, comprising: A step of acquiring first control input information input by a user; A step of calculating second control input information for realizing the target operation based on information indicating the target operation of the user; A step of calculating control information for controlling the robot based on the degree of similarity between the first control input information and the second control input information. Robot control method. (Item 7) A program for causing a computer to function as each part in the robot control device according to any one of Items 1 to 5.
[0061] Although the present embodiment has been described above, the present invention is not limited to such a specific embodiment, and various modifications and changes are possible within the scope of the gist of the present invention described in the claims.
Explanation of Signs
[0062] 10 Robot control device 11 Target operation acquisition unit 12 First control input acquisition unit 13 Control operation control coordinate calculation unit 14 Second control input calculation unit 15 Target operation execution intention determination unit 16 Control information calculation unit 17 Robot control unit 1000 Drive device 1001 Recording medium 1002 Auxiliary storage device 1003 Memory device 1004 CPU 1005 Interface device 1006 Display device 1007 Input device 1008 Output device
Claims
1. A first control input acquisition unit that acquires first control input information input by a user; A second control input calculation unit that calculates second control input information for realizing the target operation based on information indicating the target operation of the user; A target operation execution intention determination unit that calculates the similarity between the first control input information and the second control input information, and determines whether the user has an intention to perform the target operation based on the similarity; When it is determined that the user has an intention to perform the target operation, a control information calculation unit that calculates control information for controlling the robot based on the first control input information, the second control input information, and the similarity. A robot control device comprising the above.
2. The control information calculation unit When it is determined that the user has an intention to perform the target operation, the first control input information and the second control input information are weighted based on the similarity and combined to calculate control information for controlling the robot. The robot control device according to Claim 1.
3. The target operation execution intention determination unit determines that the user has an intention to perform the target operation when the similarity is greater than a threshold value. The robot control device according to Claim 1 or 2.
4. The target operation execution intention determination unit and the control information calculation unit repeatedly execute the process until a predefined condition is satisfied. The robot control device according to any one of Claims 1 to 3.
5. A robot control method executed by a robot control device for controlling a robot, comprising: A step of acquiring first control input information input by a user; A step of calculating second control input information for realizing the target operation based on information indicating the target operation of the user; A step of calculating the similarity between the first control input information and the second control input information, and determining whether the user has an intention to perform the target operation based on the similarity; When it is determined that the user has an intention to perform the target operation, a step of calculating control information for controlling the robot based on the first control input information, the second control input information, and the similarity. A robot control method comprising the above.
6. In the step of calculating control information for controlling the robot, the robot control device When it is determined that the user intends to perform the target operation, the first control input information and the second control input information are combined with weighting based on the similarity, and control information for controlling the robot is calculated. The robot control method according to claim 5.
7. A program for causing a computer to function as each part in the robot control device according to any one of claims 1 to 4.
Citation Information
Patent Citations
Robot motion control device
JP2017196678A