JP2024072051A - Movement modification system, movement modification method, and working robot - Google Patents

Abstract

To provide an operation modification system, an operation modification method and a working robot, which enable efficient work of a working robot for working using a learning model where an operation of an operator is learned.SOLUTION: An operation modification system includes: a learning part for learning a standard operation model corresponding to a prescribed operation of a sensing object acquired using a sensor, on the basis of sensing information corresponding to the prescribed operation of the sensing object; a model creation part for creating a modification operation model in which execution time of each operation in the standard operation model is set to be shorter than required time of each of the operations at the time of standard operation model creation, by referring to the standard operation model; and a control part for operating the working robot by referring to the modification operation model.SELECTED DRAWING: Figure 1

Description

The present invention relates to a motion modification system, a motion modification method, and a work robot.

In recent years, technology has been developed to allow work robots to perform tasks performed by humans, such as line work in factories. For example, Patent Literature 1 discloses a learning model generation method that generates an optimal learning model for each worker according to the tendencies of the collected data for each worker. The generated learning model can be used to operate a work robot that replaces a worker or works in collaboration with a worker.

JP 2022-135178 A

Due to limitations in the physical abilities of workers, there is a certain limit to how fast they can work. For this reason, if a work robot were to simply use a learning model that has learned the specific movements of a worker, the work robot might not be able to perform the work efficiently.

The present invention aims to provide a behavior modification system, a behavior modification method, and a work robot that enables a work robot to work efficiently using a learning model that has learned the work of a worker.

The behavior modification system according to the present invention comprises a work robot, a sensor, and a management control device capable of communicating with the work robot and the sensor. The management control device comprises a learning unit that learns a standard behavior model corresponding to a predetermined motion of the sensing target based on sensing information corresponding to the predetermined motion of the sensing target acquired using the sensor, a model generation unit that references the standard behavior model and generates a modified behavior model in which the execution time of each motion in the standard behavior model is set shorter than the time required for each motion when the standard behavior model was generated, and a control unit that operates the work robot by referring to the modified behavior model.

The behavior modification system according to the present invention comprises a work robot, a plurality of sensors for sensing a plurality of different sensing targets, and a management control device capable of communicating with the work robot and the plurality of sensors. The management control device comprises a learning unit that learns each of the predetermined behaviors of the plurality of sensing targets and a plurality of standard behavior models corresponding to each of the predetermined behaviors of the plurality of sensing targets based on a plurality of sensing information corresponding to the predetermined behaviors of the plurality of sensing targets obtained using the plurality of sensors, a model generation unit that generates a modified behavior model that integrates at least a portion of the predetermined behaviors of the plurality of sensing targets by referring to the plurality of standard behavior models, and a control unit that operates the work robot by referring to the modified behavior model.

The movement modification method according to the present invention is characterized in that it learns a standard movement model corresponding to a predetermined movement of a sensing target based on sensing information corresponding to the predetermined movement of the sensing target obtained using a sensor, generates a modified movement model by referring to the standard movement model in which the execution time of each movement in the standard movement model is set shorter than the time required for each movement when the standard movement model was generated, and operates the work robot by referring to the modified movement model.

The working robot according to the present invention is characterized by having a drive mechanism for operating the working robot, a learning unit that learns a standard motion model corresponding to a predetermined motion of a sensing target based on sensing information corresponding to the predetermined motion of the sensing target obtained using a sensor, a model generation unit that references the standard motion model and generates a modified motion model in which the execution time of each motion in the standard motion model is set shorter than the time required for each motion when the standard motion model was generated, and a control unit that operates the working robot to control the drive mechanism by reference to the modified motion model.

The present invention provides a method for modifying the behavior of a work robot that performs work using a learning model that has learned the work of a worker, and a work robot that enables the work robot to perform work efficiently.

FIG. 1A is a diagram showing an example of a system configuration in a behavior modification system according to a first embodiment of the present invention, and FIG. 1B is a diagram showing an example of the humanoid robot shown in FIG. 1 is a diagram showing an example of a relationship between a standard behavior model and a behavior modification model in a behavior modification system according to a first embodiment of the present invention. FIG. 1 is a block diagram showing an example of the configuration and functions of an operation modification system according to a first embodiment of the present invention. 2 is a block diagram showing an example of the functions of a management control device in the operation modification system of embodiment 1 according to the present invention. FIG. 4 is an example of a flowchart showing a process of the operation modification system according to the first embodiment of the present invention. 6 is an example of a flowchart showing more detailed processing of the modified action model generation processing indicated in step S103 of FIG. 5 . FIG. 1A is a diagram showing an example of a system configuration in a behavior modification system according to a first modified example of the first embodiment of the present invention, and FIG. 1B is a diagram showing an example of the humanoid robot shown in FIG. FIG. 11 is a block diagram showing an example of functions of a humanoid robot in a behavior modification system according to a first modified example of the first embodiment of the present invention. 13 is an example of a flowchart showing more detailed processing of a modified action model generation process in the action modification system according to the first modification example of the first embodiment of the present invention. FIG. 11 is a diagram showing an example of a relationship between a standard action model and an action modification model in the action modification system according to the first modified example of the first embodiment of the present invention.

The movement modification system, movement modification method, and working robot will be described below with reference to the drawings. However, please note that the technical scope of the present invention is not limited to these embodiments, but extends to the inventions described in the claims and their equivalents.

(Embodiment 1)
FIG. 1 is a diagram for explaining the operation modification system.

Figure 1 (a) is a diagram showing an example of the system configuration of a behavior modification system according to a first embodiment of the present invention. This behavior modification system comprises a first humanoid robot 20a and a second humanoid robot 20b that function as working robots, and a third humanoid robot 20c that functions as a mobile robot. Note that the number of humanoid robots is not limited to this.

Each humanoid robot 20a-c moves to the vicinity of a worker 400 working on the work line 201 in the work area 200 upon receiving instructions from a management control device 60 (see FIG. 3) described later or from each information processing device provided in each humanoid robot 20a-c. The motion modification system senses a predetermined motion of the worker 400 using a third robot sensor 23c (third robot imaging device 24c) provided in the third humanoid robot 20c. The predetermined motions are diverse, and include, for example, assembling parts, moving parts, painting a product, and the movement of the worker himself. In addition, when sensing the worker 400, a known image recognition technology may be used, or the worker 400 and his/her predetermined motion may be recognized by learning using a learning unit 663 (see FIG. 4).

This movement modification system learns a standard movement model corresponding to a specific movement of the worker 400 based on sensing information corresponding to the specific movement of the worker 400 acquired using the third robot sensor 23c (third robot imaging device 24c). The standard movement model is a model that represents the contents of the work items of the worker 400, in other words, a collection of each movement specified in the work items. Then, this movement modification system references the standard movement model to generate a modified movement model in which the execution time of each movement in the standard movement model is set shorter than the time required for each movement when the standard movement model was generated.

This behavior modification system operates the first humanoid robot 20a and the second humanoid robot 20b, which function as work robots, in response to instructions from the management control device 60 or instructions from an information processing device provided in each humanoid robot. At this time, this behavior modification system refers to the modified behavior model.

This behavior modification system can improve the work efficiency of each humanoid robot by operating the first humanoid robot 20a and the second humanoid robot 20b with reference to a modified behavior model in which the execution time of each behavior is set to be short. As an example, if there are six workers on one work line 201 in the workplace 200 and they originally completed 100 products per hour, six humanoid robots are placed on the line 201 and operated at 10 times the speed of the standard behavior model. In other words, this behavior modification system operates each humanoid robot with reference to a behavior modification model set to operate in one-tenth the execution time of the behavior in the standard behavior model. This makes it possible to complete 1,000 products per hour.

In addition, in this behavior modification system, the first humanoid robot 20a and the second humanoid robot 20b that function as working robots may sense their respective robot behaviors using the first robot sensor 23a (first robot image capture device 24a) and the second robot sensor 23b (second robot image capture device 24b) provided therein, respectively.

As an example, the first humanoid robot 20a operates the first body/head drive mechanism 21a (see FIG. 3) in response to an instruction from the management control device 60 or an instruction from the first information processing device 25a so that the sensing area 230a (imaging area 240a) of the first robot sensor 23a (first robot imaging device 24a) targets the first gripping parts 265a, 266a of the first humanoid robot 20a. Similarly, the second humanoid robot 20b operates the second body/head drive mechanism in response to an instruction from the management control device 60 or the first information processing device 25a so that the sensing area 230b (imaging area 240b) of the second robot sensor 23b (second robot imaging device 24b) targets the second gripping parts 265b, 266b of the second humanoid robot 20b. This makes it possible to check whether the robot motion of each humanoid robot is a robot motion that refers to the modified motion model.

Figure 1 (b) is a diagram showing an example of the humanoid robot shown in (a). The humanoid robot 20, which functions as a work robot and a mobile robot, comprises a robot body 21, a robot movement mechanism 22, a robot sensor 23, a robot imaging device 24 included in the robot sensor 23, an information processing device 25, and a robot arm 26.

The humanoid robot 20 can move using a robot movement mechanism 22 provided below the robot body 21, and moves to the vicinity of the work line 201 in the workplace 200 upon receiving instructions from outside the humanoid robot 20, such as from a management control device 60, or by referring to a program stored in the information processing device 25.

The robot main body 21 comprises a robot torso 211 and a robot head 212. The robot torso 211 and the robot head 212 constitute a torso/head drive mechanism, and are capable of changing the sensing area 230 (imaging area 240) of the robot sensor 23 (robot imaging device 24). The configuration of the drive mechanism is not particularly limited, and may be configured such that, for example, the robot head 212 rotates a predetermined angle relative to the robot torso 211, or the robot torso 211 rotates a predetermined angle relative to the robot movement mechanism 22, by a servo motor (not shown).

A robot movement mechanism 22 is provided below the robot body 211, a robot arm 26 is provided on one side of the robot body 211, and a robot sensor 23 is provided in the robot head 212. An information processing device 25 is also provided inside the robot body 21.

The robot movement mechanism 22 may be of any configuration, for example, it may be provided with a rotating body driven by a motor, or may have legs that resemble the shape of a human leg. As an example, when the robot movement mechanism 22 is configured to resemble the shape of a human leg, a servo motor is provided at the location corresponding to the human joint, and the movement mechanism is configured by rotating it by a predetermined angle.

The robot sensor 23 is preferably provided in the robot head 212 and senses the robotic movements of the worker 400, other working robots, or the humanoid robot 20, preferably the robotic movements of the robot arm 26, and more preferably the robotic movements of the grippers 255 and 256. The robot sensor 23 also sequentially acquires information representing at least the distance and angle between the object on which the humanoid robot 20 works and the robot arm 26, which are located in the vicinity of the humanoid robot 20. Examples of the robot sensor 23 include the highest performance camera, a thermal camera, a high-pixel, telephoto, ultra-wide-angle, 360-degree, high-performance camera, radar, solid-state LiDAR, LiDAR, a multi-color laser coaxial displacement meter, vision recognition, or various other sensor groups. These are also examples of the robot imaging device 24. Other examples of the robot sensor 23 include a vibration meter, hardness meter, micro vibration meter, ultrasonic measuring instrument, vibration measuring instrument, infrared measuring instrument, ultraviolet measuring instrument, electromagnetic wave measuring instrument, thermometer, hygrometer, spot AI weather forecast, high-precision multi-channel GPS, low altitude satellite information, or long-tail incident AI data.

Examples of sensing information acquired from the robot sensor 23 include images, distance, vibration, heat, smell, color, sound, ultrasound, radio waves, ultraviolet light, infrared light, humidity, etc., and the image and distance information is preferably acquired by the robot imaging device 24. The robot sensor 23 (robot imaging device 24) performs these sensing operations, for example, every nanosecond. The sensing information is used, for example, for motion capture of the movements of the worker 400, a 3D map of the workplace 200, navigation of the movements and movements of the worker 400 in the workplace 200, analysis of cornering, speed, etc.

The robot arm 26 comprises a right arm 261 and a left arm 262. The right arm 261 comprises a right gripping support part 263 and a right gripping part 265, and the left arm 262 comprises a left gripping support part 264 and a left gripping part 266. The right gripping support part 263 is a mechanism for supporting the right gripping part 265, and the left gripping support part 264 is a mechanism for supporting the left gripping part 266, and may be shaped like a human arm, for example. The gripping parts 265 and 266 are mechanisms for gripping, for example, parts for work, and may be shaped like a human hand, for example.

The robot arm 26 constitutes an arm drive mechanism. The configuration of the drive mechanism is not particularly limited, and for example, if the robot arm 26 is to resemble a human shape, a configuration may be adopted in which servo motors are provided at each joint location, such as a location corresponding to a human shoulder, a location corresponding to an elbow, a location corresponding to a wrist, a location corresponding to a finger joint, and the like, and rotated at a predetermined angle.

The humanoid robot 20 may further be provided with a sensor, for example, on the robot torso 211 (see FIG. 7(b)). In this case, the sensor is located at a different height than the robot sensor 23 located on the robot head 212. The different height allows the sensor to sense the movements of the worker 400 from different angles.

Figure 2 shows an example of the relationship between the standard behavior model and the behavior modification model in this behavior modification system.

As mentioned above, the standard action model is a model that represents a set of actions specified in a work item, and includes multiple actions. As an example, there are a total of 26 actions specified in the work item, and the actions are designated as action A, action B, action C, and action Z.

In contrast, the movement modification model is a model in which the execution time of each movement in the standard movement model is set shorter than the time required for each movement when the standard movement model was generated. For example, assume that the required times for each movement in the standard movement model are T _A seconds for movement A and T _B seconds for movement B. In this case, the movement modification model sets the execution time for movement A to t _{A seconds, which is shorter than T A seconds} , and for movement B to t _B seconds, which is shorter than T _B seconds. The same applies to movements C to Z. This makes it possible to shorten the execution time when a work robot is executed using the movement modification model than when the work robot is executed using the standard movement model.

In addition, in the behavior modification model, it is sufficient that the execution time of at least one behavior is less than the time required for that behavior in the standard behavior model, and it is not necessary that the execution time of all of the behaviors included in the standard behavior model is less than the time required for that behavior in the standard behavior model. In other words, if the execution time of at least one behavior is less than the time required for that behavior in the standard behavior model, then the execution time of each behavior in the standard behavior model will be set shorter than the time required for each behavior when the standard behavior model was generated.

Figure 3 is a block diagram showing an example of the configuration and functions of the operation modification system 100 of this embodiment.

The behavior modification system 100 includes a first humanoid robot 20a, a second humanoid robot 20b, a third humanoid robot 20c, and a management control device 60. The first humanoid robot 20a, the second humanoid robot 20b, and the third humanoid robot 20c are each connected to the communication unit 64 of the management control device 60 via wireless or wired communication, and receive instructions from the management control device 60 and transmit information acquired by each sensor. The humanoid robots 20a-c may also be connected to each other via wireless or wired communication, and send and receive information acquired by each sensor and instructions.

The first humanoid robot 20a, which functions as a work robot, includes a first body/head drive mechanism 21a, a first robot movement mechanism 22a, a first robot sensor 23a, a first robot imaging device 24a included in the first robot sensor 23a, a first information processing device 25a, and a first arm drive mechanism 26a. In this embodiment, the second humanoid robot 20b, which functions as a work robot, and the third humanoid robot 20c, which functions as a mobile robot, are also configured in the same manner as the first humanoid robot 20a.

The first information processing device 25a according to this embodiment includes a CPU (Central Processing Unit) 1212, a RAM (Random Access Memory) 1214, and a graphic controller 1216, which are connected to each other by a host controller 1210. The first information processing device 25a also includes input/output units such as a communication interface 1222, a storage device 1224, a DVD drive, and an IC card drive, which are connected to the host controller 1210 via an input/output controller 1220. The DVD drive may be a DVD-ROM drive, a DVD-RAM drive, or the like. The storage device 1224 may be a hard disk drive, a solid state drive, or the like. The first information processing device 25a also includes input/output units such as a ROM (Read Only Memory) 1230 and a keyboard, which are connected to the input/output controller 1220 via an input/output chip 1240.

The CPU 1212 operates according to the programs stored in the ROM 1230 and the RAM 1214, thereby controlling each unit. The graphics controller 1216 acquires image data generated by the CPU 1212 into a frame buffer or the like provided in the RAM 1214 or into itself, and causes the image data to be displayed on the display device 1218.

The communication interface 1222 communicates with other electronic devices via a network. The storage device 1224 stores programs and data used by the CPU 1212 in the first information processing device 25a. The storage device 1224 may also store sensing information. The DVD drive reads programs or data from a DVD-ROM or the like and provides them to the storage device 1224. The IC card drive reads programs and data from an IC card and/or writes programs and data to an IC card.

The ROM 1230 stores therein a boot program or the like executed by the first information processing device 25a upon activation, and/or a program that depends on the hardware of the first information processing device 25a. The input/output chip 1240 may also connect various input/output units to the input/output controller 1220 via a USB port, a parallel port, a serial port, a keyboard port, a mouse port, etc.

The programs are provided by a computer-readable storage medium such as a DVD-ROM or an IC card. The programs are read from the computer-readable storage medium, installed in the storage device 1224, RAM 1214, or ROM 1230, which are also examples of computer-readable storage media, and executed by the CPU 1212. The information processing described in these programs is read by the first information processing device 25a, and brings about cooperation between the programs and the various types of hardware resources described above. An apparatus or method may be configured by realizing the operation or processing of information in accordance with the use of the first information processing device 25a.

For example, when communication is performed between the first information processing device 25a and an external device, the CPU 1212 may execute a communication program loaded into the RAM 1214 and instruct the communication interface 1222 to perform communication processing based on the processing described in the communication program. Under the control of the CPU 1212, the communication interface 1222 reads transmission data stored in a transmission buffer area provided in the RAM 1214, the storage device 1224, a DVD-ROM, or a recording medium such as an IC card, and transmits the read transmission data to the network, or writes reception data received from the network to a reception buffer area or the like provided on the recording medium.

The CPU 1212 may also cause all or a necessary portion of a file or database stored in an external recording medium such as the storage device 1224, a DVD drive (DVD-ROM), an IC card, etc. to be read into the RAM 1214, and perform various types of processing on the data on the RAM 1214. The CPU 1212 may then write back the processed data to the external recording medium.

Various types of information, such as various types of programs, data, tables, and databases, may be stored on the recording medium and undergo information processing. CPU 1212 may perform various types of processing on data read from RAM 1214, including various types of operations, information processing, conditional decisions, conditional branches, unconditional branches, information search/replacement, etc., as described throughout this disclosure and specified by the instruction sequences of the programs, and write the results back to RAM 1214. CPU 1212 may also search for information in files, databases, etc. in the recording medium.

The above-described program or software module may be stored in a computer-readable storage medium on the first information processing device 25a or in the vicinity of the first information processing device 25a. In addition, a recording medium such as a hard disk or RAM provided in a server system connected to a dedicated communication network or the Internet can be used as a computer-readable storage medium, thereby providing the program to the first information processing device 25a via the network.

The blocks in the flowcharts and diagrams in this embodiment may represent stages of a process in which an operation is performed or "parts" of a device responsible for performing the operation. Particular stages and "parts" may be implemented by dedicated circuitry, programmable circuitry provided with computer-readable instructions stored on a computer-readable storage medium, and/or a processor provided with computer-readable instructions stored on a computer-readable storage medium. The dedicated circuitry may include digital and/or analog hardware circuitry, and may include integrated circuits (ICs) and/or discrete circuits. The programmable circuitry may include reconfigurable hardware circuitry including AND, OR, XOR, NAND, NOR, and other logical operations, flip-flops, registers, and memory elements, such as, for example, field programmable gate arrays (FPGAs), programmable logic arrays (PLAs), and the like.

A computer-readable storage medium may include any tangible device capable of storing instructions that are executed by a suitable device, such that a computer-readable storage medium having instructions stored thereon comprises an article of manufacture that includes instructions that can be executed to create means for performing the operations specified in the flowchart or block diagram. Examples of computer-readable storage media may include electronic storage media, magnetic storage media, optical storage media, electromagnetic storage media, semiconductor storage media, and the like. More specific examples of computer-readable storage media may include floppy disks, diskettes, hard disks, random access memories (RAMs), read-only memories (ROMs), erasable programmable read-only memories (EPROMs or flash memories), electrically erasable programmable read-only memories (EEPROMs), static random access memories (SRAMs), compact disk read-only memories (CD-ROMs), digital versatile disks (DVDs), Blu-ray disks, memory sticks, integrated circuit cards, and the like.

The computer readable instructions may include either assembler instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source or object code written in any combination of one or more programming languages, including object-oriented programming languages such as Smalltalk (registered trademark), JAVA (registered trademark), C++, etc., and conventional procedural programming languages such as the "C" programming language or similar programming languages.

The computer-readable instructions may be provided to a processor of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, or a programmable circuit, either locally or over a local area network (LAN), a wide area network (WAN), such as the Internet, etc., so that the processor of the general-purpose computer, special-purpose computer, or other programmable data processing apparatus, or the programmable circuit, executes the computer-readable instructions to generate means for performing the operations specified in the flowcharts or block diagrams. Examples of processors include computer processors, processing units, microprocessors, digital signal processors, controllers, microcontrollers, etc.

The contents described so far also apply to each information processing device provided in the second humanoid robot 20b and the third humanoid robot 20c.

The management control device 60 is a control device that issues instructions to each humanoid robot 20a-c in order to realize the behavior modification system 100. The management control device 60 also acquires sensing information stored in the storage device of each information processing device.

The management control device 60 is composed of a CPU 60A, a RAM 60B, a ROM 60C, an input/output unit (I/O) 60D, a bus 60E such as a data bus or a control bus that connects these, and a communication unit 64. A storage medium 62 is connected to the I/O 60D.

A communication unit 64 is also connected to the I/O 60D, which transmits and receives sensing information, work manual information, schedule information, etc., between the control system of the humanoid robot 20. The work manual information includes, for example, the name and content of each work item, the order of the work items, and information on the standard work time required for each work item. The schedule information includes, for example, information indicating the work time and start/end times for the entire work, information indicating the work time and start/end times for each work item, and information indicating the worker for each work item.

Figure 4 is a block diagram showing an example of the functions of the management control device 60 in the operation modification system of this embodiment.

The management control device 60 includes a storage medium 62, a communication unit 64, and a processing unit 66.

The storage medium 62 includes, for example, at least one of a semiconductor storage device, a magnetic tape device, a magnetic disk device, or an optical disk device. The storage medium 62 stores driver programs, operating system programs, application programs, data, etc. used for processing in the processing unit 66. For example, the storage medium 62 stores sensing information. The storage medium 62 also stores work manual information and/or schedule information for the worker 400.

The communication unit 64 has a wireless communication interface circuit such as Wi-Fi (registered trademark) and/or a wired communication interface circuit such as Ethernet (registered trademark). The communication unit 64 transmits and receives various information to and from the humanoid robots 20a and 20b through the interface circuits.

The processing unit 66 has one or more processors and their peripheral circuits. The processing unit 66 centrally controls the overall operation of the operation modification system 100, and is, for example, a CPU. The processing unit 66 executes processing by referring to programs (driver programs, operating system programs, application programs, etc.) stored in the storage medium 62. The processing unit 66 can also execute multiple programs (application programs, etc.) in parallel.

The processing unit 66 includes a determination unit 661, a control unit 662, a learning unit 663, and a model generation unit 664. Each of these units is a functional module realized by a program executed by a processor included in the processing unit 66. Alternatively, each of these units may be implemented in the processing unit 66 as firmware.

When there are multiple sensing targets and multiple sensors that sense the targets, the determination unit 661 determines whether the multiple sensors are sensing different targets. This determination may be made using a known image recognition technique or by referring to learning by the learning unit 663.

The control unit 662 operates the first humanoid robot 20a and/or the second humanoid robot 20b, which function as a working robot, by referring to the modified motion model generated by the model generation unit 664. Preferably, the control unit 662 operates the robot movement mechanism, torso/head drive mechanism, and/or arm drive mechanism of each humanoid robot by referring to the modified motion model.

The learning unit 663 learns a standard action model corresponding to the predetermined action of the worker 400 based on sensing information corresponding to the predetermined action of the worker 400 acquired using the third robot sensor 23c (third robot imaging device 24c). This learning is performed, for example, by automatic learning, which is learning in which a learned model is automatically created and judgment/analysis is automatically performed using the learned model. Note that the learning unit 663 may refer to work manual information and/or schedule information when generating the standard action model. This makes it possible to have each humanoid robot 20 perform an appropriate action (task) without reflecting any inappropriate predetermined action of the worker 400.

The model generation unit 664 references the standard behavior model and generates a modified behavior model in which the execution time of each behavior in the standard behavior model is set shorter than the time required for each behavior when the standard behavior model was generated.

(Processing of the behavior modification system according to the first embodiment of the present invention)
FIG. 5 is an example of a flowchart showing the processing of the operation modification system of this embodiment.

First, in response to an instruction from the management control device 60, or an instruction to read a program stored in the storage medium 62 or the storage device of the third information processing device of the third humanoid robot 20c, the third information processing device of the third humanoid robot 20c instructs the third humanoid robot 20c, which functions as a mobile robot, to move to the workshop 200 (step S101). The movement is caused by the operation of the third robot movement mechanism of the humanoid robot 20c. Note that a movement instruction may also be given at this time to the humanoid robots 20a and 20b, which function as work robots.

When moving, an instruction is given so that the sensing area 230c (imaging area 240c) of the third robot sensor 23c (third robot imaging device 24c) of the third humanoid robot 20c targets the worker 400. The placement of the third humanoid robot 20c is performed, for example, by storing a floor plan of the workplace 200 in the storage device and/or storage medium 62 of the third humanoid robot 20c in advance and associating the position of the third humanoid robot 20c with the stored floor plan. Alternatively, the placement of the third humanoid robot 20c may be based on a position optimized through machine learning. The same applies to the placement of the first humanoid robot 20a and the second humanoid robot 20b.

Next, the third robot sensor 23c (third robot imaging device 24c) senses a predetermined movement of the worker 400 on the work line 201 (step S102). In this embodiment, the control unit 662 instructs the third robot sensor 23c (third robot imaging device 24c) to target the worker 400 with the sensing area 230c (imaging area 240c), and the third robot movement mechanism and each drive mechanism of the third humanoid robot 20c are activated.

The sensing information acquired by the third robot sensor 23c (third robot imaging device 24c) is stored in the storage medium 62 via the storage device and/or communication unit 64 of the third humanoid robot 20c. The storage device and storage medium 62 of each humanoid robot function as a storage unit.

The management control device 60 learns a standard action model corresponding to a specific action of the worker 400 based on the sensing information accumulated in the memory unit, in other words, stored, and generates a modified action model by referring to the standard action model, in which the execution time of each action in the standard action model is set shorter than the required time of each action when the standard action model was generated (step S103). As an example, if the required time of one action in the standard action model is 10 seconds, the execution time of that action is set to 5 seconds in the modified action model.

The control unit 662 operates the first humanoid robot 20a and the second humanoid robot 20b, which function as working robots, by referring to the generated modified motion model (step S104). By referring to the generated modified motion model, the first humanoid robot 20a and the second humanoid robot 20b can perform the work (predetermined motion) of the worker 400 faster than the predetermined motion. For example, according to the example shown above, the modified motion model is set to operate in half the execution time of the motion in the standard motion model, so that each humanoid robot 20 can perform the work twice as fast as the predetermined motion of the worker 400.

Figure 6 is an example of a flowchart showing more detailed processing of the modified motion model processing shown in step S103 of Figure 5.

The sensing information is stored in the memory unit (step S201), and the learning unit 663 learns a standard action model corresponding to the specified action of the worker 400 based on the sensing information corresponding to the specified action of the worker 400 acquired using the third robot sensor 23c (third imaging device 24c) (step S202).

The learning unit 663 generates a standard action model based on the learning results (step S203). Note that the learning unit 663 may refer to the work manual information and/or the schedule information of the worker 400 when generating the standard action model.

The model generation unit 664 generates a modified motion model by referring to the standard motion model (step S204). The modified motion model is a model in which the execution time of each motion in the standard motion model is set shorter than the time required for each motion when the standard motion model was generated, so that the robot motions of each humanoid robot 20a, 20b operated by referring to the modified motion model are faster than the robot motions of the same robots operated by referring to the standard motion model.

(Effects of the behavior modification system according to the first embodiment)
According to the movement modification system 100 of this embodiment, the work robot can be operated by referring to a modified movement model that is set to be shorter than the time required for each movement when the standard movement model is generated, thereby making it possible for the work robot to work efficiently.

Furthermore, according to the behavior modification system 100 of this embodiment, when generating a standard behavior model, the work manual information and/or schedule information are referenced, and then the standard behavior model is generated. The worker 400 does not always perform actions faithful to the work, and in some cases may perform unnecessary actions or omit necessary actions. Therefore, by referencing the work manual information and schedule information, it is possible to prevent unnecessary or inappropriate predetermined actions by the worker 400 from being reflected in the standard behavior model.

(First Modification of First Embodiment)
FIG. 7 is a diagram showing an example of an operation modification system according to the first modification of the present embodiment.

Figure 7 (a) is a diagram showing an example of the system configuration of a movement modification system according to Variation 1 of the first embodiment of the present invention. One of the features of this movement modification system is that it senses and learns each of the predetermined movements of multiple sensing targets (workers 400a, 400b). Another feature is that a torso sensor 23'' (torso imaging device 24'') is provided on the humanoid robot 20', which functions as a mobile robot and a working robot. In addition, this movement modification system does not necessarily require a management control device 60, and the movement modification system can be configured using the humanoid robot 20' alone.

Figure 7(b) is a diagram showing an example of the humanoid robot shown in Figure 7(a). The humanoid robot 20', which functions as a mobile and working robot, comprises a robot main body 21', a robot movement mechanism 22', a head sensor 23', a head image capture device 24' included in the head sensor 23', a body sensor 23'', a body image capture device 24'' included in the body sensor 23'', an information processing device 25', and a robot arm 26'.

The robot main body 21' comprises a robot torso 211' and a robot head 212'. The robot torso 211' and the robot head 212' constitute a torso/head drive mechanism 21' (see FIG. 8), and are capable of changing the sensing area 230' (imaging area 240') of the head sensor 23' (head imaging device 24') and the sensing area 230'' (imaging area 240'') of the torso sensor 23'' (torso imaging device 24'').

The head sensor 23' (head imaging device 24') and the torso sensor 23'' (torso imaging device 24'') are positioned at different height positions, so the torso sensor 23'' (torso imaging device 24'') senses the specified movements of each sensing target (workers 400a, 400b) from a different position than the head sensor 23' (head imaging device 24').

Figure 8 is a block diagram showing an example of the functions of a humanoid robot in this movement modification system. In the movement modification system 100', the information processing device 25' includes an information processing unit 66', a communication interface 1222', and a storage device 1224', and the information processing unit 66' includes a determination unit 661', a control unit 662', a learning unit 663', and a model generation unit 664'. That is, in the movement modification system 100', the information processing unit 66' performs the same processing as the processing unit 66 of the management control device 60. The information processing device 25' is configured to be able to communicate with the head sensor 23' (head imaging device 24'), the torso sensor 23'' (head imaging device 24''), the torso/head driving mechanism 21', the robot moving mechanism 22', and the arm driving mechanism 26'.

The humanoid robot 20' of the behavior modification system 100' is equipped with an information processing device 25' and an information processing section 66', so the humanoid robot 20' alone constitutes the behavior modification system.

Referring to FIG. 7(a), for example, the control unit 662' of the humanoid robot 20' instructs the torso sensor 23'' (torso image capture device 24'') to sense the worker 400a, and the head sensor 23' (head image capture device 24'') to sense the worker 400b. In other words, the present movement modification system 100' is provided with a plurality of sensors for sensing a plurality of different sensing targets, and a plurality of pieces of sensing information corresponding to the predetermined movements of a plurality of workers are acquired using the plurality of sensors. Note that the roles of the head sensor 23' (head image capture device 24') and the torso sensor 23'' (torso image capture device 24'') may be reversed. In addition, when each sensor senses, the determination unit 661' may determine whether each sensor is sensing a different target.

Figure 9 is an example of a flowchart showing more detailed processing of the modified behavior model generation process in the behavior modification system related to variant 1 of embodiment 1 of the present invention.

This movement modification system is different from the first embodiment in that in S102, multiple sensors sense the predetermined movements of multiple workers. In addition, this movement modification system includes step S103', which replaces S103.

In step S103', first, each piece of sensing information acquired by the head sensor 23' (head image capture device 24') and the torso sensor 23'' (torso image capture device 24'') is stored in the memory unit (memory device 1224') (step S201'). The learning unit 663' learns each of the predetermined movements of multiple workers (two in this modified example) based on multiple pieces of sensing information (two in this modified example) (step S202').

The learning unit 663' also learns multiple (two in this modified example) standard action models corresponding to the respective predetermined actions of the multiple workers based on the multiple pieces of sensing information (step S203'). For example, if the work of worker 400a is made up of actions A to M, while the work of worker 400b is made up of actions N to Z, the standard action models will be a first standard action model made up of actions A to M and a second standard action model made up of actions N to Z. Note that when generating the standard action models, the learning unit 663' may refer to work manual information and/or schedule information.

The model generation unit 664' generates a modified action model that integrates at least a portion of the specified actions performed by multiple workers (step S204').

Figure 10 is a diagram showing an example of the relationship between a standard behavior model and a behavior modification model in a behavior modification system related to variant 1 of embodiment 1 of the present invention.

For example, in a first standard motion model consisting of motions A to M, motion M is a motion of movement to hand over a certain part. Similarly, in a second standard motion model consisting of motions N to Z, motion N is a motion of movement to receive the part. In this case, motions M and N are unnecessary when performed by a single working robot. Therefore, the model generation unit 664' integrates motions A to L, which are part of the motions in the first standard motion model, and motions O to Z, which are part of the motions in the second standard motion model, to generate a modified motion model consisting of motions A to L and motions O to Z. In this way, the motion modification system 100 can refer to the generated modified motion model and, when operating a humanoid robot 20' that also functions as a working robot, make it possible to have a single humanoid robot 20' perform the predetermined motions that were performed by multiple workers, omitting unnecessary motions as necessary. As a result, it is possible to make the working robot work efficiently.

As an example of this behavior modification system, if one line 201 in the workplace 200 is initially completed by six workers who complete 100 products per hour, then by placing three humanoid robots 20' on the line 201, it becomes possible to complete 100 products per hour. In particular, as mentioned above, if multiple workers are performing different tasks, the humanoid robot 20' can perform those tasks collectively, omitting unnecessary actions as necessary.

The model generation unit 664' may generate a second modified motion model for the modified motion model in which the execution time of each motion is set shorter than the time required for each motion when the modified motion model was generated. In this case, in the example shown above, by operating the humanoid robot 20' with reference to the second modified motion model set to perform the motions in half the execution time of the motions when the modified motion model was generated, it becomes possible to complete 200 products per hour.

(Effects of Modification 1)
According to this behavior modification system, the humanoid robot 20' can constitute a behavior modification system by itself, so that even in places where communication with the management control device 60 is not possible, for example, the working robot can work efficiently by using a learning model that has learned the work of a worker.

In addition, the self-acting robot 20' is equipped with multiple (two in this modified example) sensors (imaging devices), making it possible to conduct work learning for workers even in a space that is too small to sense multiple workers, for example.

Furthermore, according to this movement modification system, multiple standard movement models corresponding to the respective predetermined movements of multiple workers are learned, and a modified movement model that integrates at least some of the predetermined movements of the multiple workers is generated by referring to the standard movement models. This makes it possible to substitute for the work (predetermined movements) of multiple workers with a smaller number of work robots than the number of workers, thereby improving work efficiency. In addition, a second modified model is generated in the modified movement model in which the execution time of each movement is set shorter than the time required for each movement at the time the modified movement model was generated, and the second modified model is referred to in operating the work robot, making it possible to have the work robot work more efficiently.

In this operation modification system, the humanoid robot that functions as both a mobile robot and a work robot does not have to be one, as shown in the example, but may be multiple. In this case, as the number of humanoid robots increases, the number of sensors increases by a multiple of the number of humanoid robots, making it possible to obtain more sensing information at one time, and since the number of work robots also increases, it becomes possible to improve work efficiency, for example, when making each work robot perform the same task.

Although the embodiment of the present invention has been described above, the present invention is not limited to the above-mentioned embodiment of the present invention, and various modifications and applications are possible without departing from the gist of the present invention.

In the present embodiment, the movement modification system 100 has been described as having one mobile robot (humanoid robot) sensing the worker 400. However, the number of mobile robots may be more than this. For example, if there are a large number of mobile robots equipped with sensors and movement mechanisms, it is possible to arrange multiple sensors so as to sense the specific movements of the worker 400 from different positions, heights, and/or directions. This makes it easier to obtain various data necessary for learning the specific movements of the worker 400, and allows sensing to be performed so as to cover each of the specific movements of the worker 400 in its entirety.

In the present embodiment, the standard action model is generated (S103) after sensing (S102) of the worker 400. However, the process from sensing to generating the standard action model does not necessarily have to be performed continuously. For example, when sensing is performed in S102, sensing information is stored (S201), and learning with reference to the sensing information (S202) may be performed a predetermined time (24 hours or one week, etc.) after the sensing. The same applies to the steps between S101 to S105. Also, conversely to the above example, S103 to S105 may be performed while S102 is being performed. In this case, the work robot is operated with reference to the modified action model while the worker 400 is performing a predetermined action, which can further improve work efficiency.

In addition, in this embodiment, the learning of the predetermined actions of the worker has been described as being performed by automatic learning. However, the learning does not necessarily have to be automatic learning, and may be other known machine learning methods, such as deep learning, unsupervised/supervised learning, reinforcement learning, etc.

In addition, in this embodiment, the mobile robot and the working robot are described as being the same humanoid robot. In this case, it is possible to use the mobile robot in combination with the working robot, which can reduce the expenses and costs associated with robot production. However, the mobile robot and the working robot may be different robots.

In addition, in this embodiment, an example of a modified behavior model that is set to execute in half the execution time or one-tenth the execution time of the behavior in the standard behavior model has been described. However, this behavior modification system is not particularly limited as long as the execution time of each behavior in the standard behavior model is set to be shorter than the time required for each behavior when the standard behavior model was generated.

100, 100', 100'' Motion modification system 20, 20a, 20b, 20c, 20' Humanoid robot (mobile robot, working robot)
23, 23a, 23b, 23c, 23', 23'' Robot sensor 230, 230a, 230b, 230c, 230', 230'' Sensing area 24, 24a, 24b, 24c, 24', 24'' Robot imaging device 240, 240a, 240b, 240c, 240', 240'' Imaging area 25, 25a, 25' Information processing device 60 Management control device 62 Storage medium (storage unit)
1224, 1224' storage device (storage unit)

Claims (6)

A working robot,
A sensor;
a management control device capable of communicating with the work robot and the sensor,
The management control device includes:
a learning unit that learns a standard action model corresponding to a predetermined action of the sensing target based on sensing information corresponding to the predetermined action of the sensing target acquired by using the sensor;
a model generation unit that references the standard behavior model and generates a modified behavior model in which an execution time of each behavior in the standard behavior model is set shorter than a required time of each behavior when the standard behavior model was generated;
a control unit that operates the working robot by referring to the modified motion model.
A behavior modification system comprising: The management control device further includes a storage unit that stores work manual information or process chart information of the sensing target,
The learning unit generates the standard operation model by referring to the sensing information and work manual information or work schedule information of the sensing target.
The behavior modification system of claim 1 . A working robot,
A plurality of sensors for sensing a plurality of different sensing targets,
a management control device capable of communicating with the work robot and the plurality of sensors;
The management control device includes:
a learning unit that learns each of the predetermined movements of the plurality of sensing targets and a plurality of standard action models corresponding to each of the predetermined movements of the plurality of sensing targets based on a plurality of sensing information corresponding to the predetermined movements of the plurality of sensing targets acquired using the plurality of sensors;
a model generation unit that generates a modified action model by integrating at least a part of predetermined actions by the plurality of sensing subjects with reference to the plurality of standard action models;
a control unit that operates the working robot by referring to the modified motion model.
A behavior modification system comprising: A plurality of the work robots are provided,
The control unit operates the plurality of work robots.
A behavior modification system according to any one of claims 1 to 3. learning a standard motion model corresponding to a predetermined motion of a sensing target based on sensing information corresponding to the predetermined motion of the sensing target obtained by using a sensor;
generating a modified behavior model by referring to the standard behavior model, the execution time of each behavior in the standard behavior model being set shorter than the required time of each behavior at the time of generating the standard behavior model;
operating the working robot by referring to the modified motion model;
A method for modifying behavior comprising: A work robot,
a drive mechanism for operating the working robot;
a learning unit that learns a standard motion model corresponding to a predetermined motion of a sensing target based on sensing information corresponding to the predetermined motion of the sensing target obtained by using a sensor;
a model generation unit that references the standard behavior model and generates a modified behavior model in which an execution time of each behavior in the standard behavior model is set shorter than a required time of each behavior when the standard behavior model was generated;
a control unit that operates the working robot to control the drive mechanism by referring to the modified motion model;
A working robot comprising:

Images