Work operation planning system and method

Abstract

To provide an operation plan of work considering work efficiency and a body load.SOLUTION: The system includes a load analysis device that analyzes a physical load of a worker with a digital human model generated using motion measurement data that measures a motion of the worker and physical function data that specifies a physical function of the worker, and generates physical load data obtained by quantifying the physical load, a data accumulation device that stores learning data including a combination of the motion measurement data, the physical function data, and the physical load data, and a motion design device that generates motion plan data of the user from the physical load condition data of the user based on a learned model that has learned using the learning data so as to output motion plan data that satisfies the physical load condition information and represents a motion during work that is possible with the physical function specified from the physical function information in time series.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a technique for generating an action plan for a task to be performed by a worker. [Background technology]

[0002] When a worker performs a task, stress is placed on the worker's body parts depending on the worker's movements. Conventionally, systems have been proposed that analyze the stress placed on the worker's body parts and provide the worker with a high-stress warning and work improvement information.

[0003] For example, the system in Patent Document 1 has a worker wearing motion sensors and load sensors on their body parts perform work, acquire work motion data and load data based on information detected by each sensor, estimate the posture of the body part from the motion data, estimate the load on the body from the load data, and estimate the physical load on each body part from the estimated posture and load. Furthermore, the system generates work improvement suggestion information based on the physical load on each body part if a target body part is under high load. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent No. 7465842 Summary of the Invention [Problem to be solved by the invention]

[0005] The above-mentioned systems for assessing physical load during work suggest work improvements to reduce physical load, but do not consider improving work efficiency. For example, workers are required to complete a task within a given time limit.

[0006] The present disclosure has been made in consideration of the above circumstances, and its purpose is to propose a system that provides a work motion plan that takes into account both the work efficiency and physical load on the worker. [Means for solving the problem]

[0007] In order to solve the above problem, a work operation planning system according to one aspect of the present disclosure includes: a load analysis device that acquires motion measurement data that records the motion of a worker during work in chronological order and physical function data that identifies the physical functions of the worker, generates a digital human model that reproduces the motion of the worker based on the motion measurement data and the physical function data, analyzes the physical load imposed on the worker in the motion of the worker using the digital human model, and generates physical load data in which the physical load is quantified; a data accumulation device that stores a combination of the motion measurement data, the physical function data, and the physical load data as learning data; The action design device has a trained model trained with the learning data so that when work condition data including work condition information, including physical load tolerance information and work time limits, and physical function information is input, it outputs action plan data that represents, in time series, movements during the work that satisfy the work condition information and are possible with physical functions identified from the physical function information, and generates the action plan data for the user from the user's work condition data using the trained model.

[0008] Further, a work operation planning method according to an aspect of the present disclosure includes: acquiring motion measurement data that records the motion of a worker during work in chronological order and physical function data that identifies the physical functions of the worker, generating a digital human model that reproduces the motion of the worker based on the motion measurement data and the physical function data, analyzing the physical load imposed on the worker in the motion of the worker using the digital human model, and generating physical load data in which the physical load is quantified; storing a combination of the motion measurement data, the physical function data, and the physical load data as learning data in a data accumulation device; and When work condition data including work condition information, including physical load tolerance information and work time limits, and physical function information is input, the operation plan data for the user is generated from the work condition data of the user using a trained model trained with the learning data so as to output operation plan data that satisfies the work condition information and represents in time series movements during the work that are possible with the physical functions identified from the physical function information. [Effects of the Invention]

[0009] According to the present disclosure, it is possible to provide a work motion plan that takes into consideration both the work efficiency and the physical load on the worker. [Brief explanation of the drawings]

[0010] [Figure 1] FIG. 1 is a block diagram showing a schematic configuration of a work operation planning system according to one embodiment of the present disclosure. [Figure 2] Figure 2 is a data flow diagram of the AI ​​model. DETAILED DESCRIPTION OF THE INVENTION

[0011] Next, an embodiment of the present disclosure will be described with reference to the drawings. A work motion planning system according to the present disclosure designs the movements of a worker according to the physical functions of the worker so that the worker can complete a target task within a predetermined time limit while minimizing the physical burden on the worker. The work motion planning system can be used to plan the movements of the human body in various fields, such as the movements of workers during work in industrial fields such as manufacturing, agriculture, and the service industry, and the movements of players in cultural and artistic fields such as sports and musical performances. Below, the work motion planning system will be described in terms of its application to planning the movements of a worker from the start to the completion of a target task in the industrial field.

[0012] Fig. 1 is a block diagram showing a schematic configuration of a work motion planning system 1 according to one embodiment of the present disclosure. The work motion planning system 1 shown in Fig. 1 includes a motion measurement device 2, a load analysis device 3, a data collection device 4, a motion design device 5, and a motion teaching device 6.

[0013] "Motion measurement device 2" The motion measurement device 2 generates motion measurement data 91 that records the motions of a worker while working in chronological order. The motion measurement device 2 generates the motion measurement data 91 using, for example, motion capture technology. Motion capture is a technology that records the motions of a human body in chronological order by recording the positions and movements of parts that characterize the motion of the human body, such as joints. The motion measurement device 2 has a measurement unit 21, a recording unit 22, and a motion information generation unit 23.

[0014] Motion capture technologies are divided into optical, mechanical, and magnetic types depending on the method used to measure the position of the markers. The optical measurement unit 21 includes multiple cameras that capture the movement of a human body with markers attached from different angles. The recording unit 22 records the three-dimensional position and orientation of the markers by calculating the distance to each marker based on the misalignment of the images captured by the multiple cameras. The mechanical measurement unit 21 includes an acceleration sensor, an angular velocity sensor, and a geomagnetic sensor attached to the joints of the human body. The recording unit 22 calculates the three-dimensional movement of the joints based on the values ​​detected by the measurement unit 21 and records the calculated joint movement. The magnetic measurement unit 21 includes a magnetic field generator and a magnetic sensor that detects distortion caused by the movement of a magnetic coil marker attached to the joints of the human body within the magnetic field. The recording unit 22 records the three-dimensional position and orientation of the magnetic coil based on the detected magnetic distortion.

[0015] The motion information generating unit 23 generates motion measurement data 91 that represents the motion of the worker's joints during work in time series, based on the worker's motion quantified by the recording unit 22. The motion measurement data 91 also includes the work time from the start to the completion of the work. The motion measurement data 91 generated by the motion measuring device 2 is output to the load analyzing device 3 together with physical function data 92. The motion measurement data 91 and physical function data 92 are stored in the data collecting device 4.

[0016] The physical function data 92 includes information that can identify the physical functions of the worker, such as the worker's body dimensions, age, sex, dominant hand, and whether or not the worker has any limb disabilities. The physical function data 92 may also include information about the work environment that affects the physical functions, such as differences in the shape and size of the workbench and work jigs, and differences in the shape and weight of the target workpiece. The physical function data 92 may be provided in advance, or may be generated by image processing from a captured image of the worker.

[0017] 《Load analysis device 3》 The load analysis device 3 uses digital human technology to analyze the physical load placed on the worker during work from the motion measurement data 91 and physical function data 92, and generates physical load data that quantifies the physical load. Digital human technology supports the design of products and services by recreating people of various ages and physiques as digital humans on a computer, simulating the movements of muscles and bones using these digital humans, and evaluating the effects of specific motions on the body. The load analysis device 3 is a so-called digital human simulator, and includes a digital human model generation unit 31 and a load analysis unit 32.

[0018] The digital human model generation unit 31 acquires motion measurement data 91 and physical function data 92 and uses these data to generate a digital human model DH that reproduces the motions of the worker during work. The digital human model DH according to this embodiment is a simulation model capable of dynamic analysis that simulates the structure of the human body. One example of such a digital human model DH is a musculoskeletal simulation model that simulates the musculoskeletal structure of the human body and can estimate muscle activity and joint loads from motion measurement data 91 obtained by measuring actual motions based on an inverse dynamics method.

[0019] The digital human model DH has, for example, a large number of body parts, joints, and degrees of freedom, and each body part is connected by a joint. The dimensions of each body part of the digital human model DH are determined based on body function data 92. The movement parameters of the digital human model DH are determined based on motion measurement data 91. By specifying the position and posture of a specific body part based on inverse kinematics, the movement of the digital human model DH is calculated to determine the postures of the intermediate joints leading up to that point. Examples of the movement parameters of the digital human model DH include the movement trajectory of the body part, the movement of the center of gravity, and the position, speed, acceleration, and jerk of each joint. The movement trajectory, the movement of the center of gravity, the joint position, and the like can be specified as three-dimensional coordinates.

[0020] The load analysis unit 32 performs load analysis using the digital human model DH generated by the digital human model generation unit 31. Specifically, it estimates the physical load on the whole body and specific body parts when the digital human model DH reproduces the worker's movements during work. The estimated physical load is expressed as a numerical value. Examples of numerical values ​​that represent the physical load include joint torque, energy consumption, and muscle activity.

[0021] The load analysis unit 32 generates physical load data 93 including the quantified physical load based on the results of the load analysis, and stores the physical load data 93 in the data collection device 4. The physical load data 93 includes, for example, joint torque, energy consumption, movement trajectories of body parts, movement of the center of gravity, position, speed, acceleration, and jerk of each joint. The physical load data 93 is associated with the movement measurement data 91 and physical function data 92 from which it was generated, and stored in the data collection device 4.

[0022] Data Collection Device 4 The data collection device 4 has a learning database 41 that stores a combination of movement measurement data 91, physical function data 92, and physical load data 93 generated by measuring actual work movements as learning data 94. The data collection device 4 may also store movement plan data 95 generated by the action design device 5 described below, and work condition data 96 used by the action design device 5. The data collection device 4 is electrically connected to the movement measurement device 2, load analysis device 3, action design device 5, and movement teaching device 6, and data can be written to and read from the data collection device 4 from these devices.

[0023] "Motion Design Device 5" The action design device 5 generates action plan data 95 using an AI model 50. As shown in Fig. 2, the AI ​​model 50 is configured to output action plan data 95 when task condition data 96 is input. The action design device 5 has a model creation unit 51 and an action plan generation unit 52.

[0024] The model creation unit 51 uses AI (artificial intelligence) to analyze trends in operations in line with work conditions from a large amount of learning data 94, and creates an AI model 50 that outputs operation plan data 95 when work condition data 96 is input. In detail, the model creation unit 51 uses a machine learning algorithm to train a basic AI model using the large amount of learning data 94, and determines optimized parameters. The model creation unit 51 may use a known machine learning algorithm. A trained AI model 50 is created by setting optimized parameters in the basic AI model. In other words, the AI ​​model 50 is a trained model that has been machine-learned using the learning data 94.

[0025] The work condition data 96 includes work content information 961, work condition information 962 including work time limits and allowable physical load ranges, and physical function information 963. The work condition data 96 may be read from pre-stored data or may be input using an input device.

[0026] In this embodiment, the worker performs one type of work, but the worker may be able to perform multiple types of work. In this case, the work content is specified by work content information 961. The information configuration of the work content information 961 is not particularly limited, but examples include a combination of route information consisting of the start point, end point, and via points of a working unit such as a hand based on the position and posture of the worker when standing, and the points on the route where the working unit acts on the target work, and information about the target work such as the dimensions and mass of the target work. If there is only one type of work, the work content information 961 may be omitted.

[0027] The work time limit, which is one piece of work condition information 962, is the upper limit of the work time from the start to the completion of the work. The work time must not exceed the work time limit, but if the work time is significantly shorter than the work time limit, there is a high possibility that the physical strain will increase. The work time limit may be input as a numerical value, or it may be input by selecting from multiple stages such as short, medium, and long. In the case of selective input, a numerical value corresponding to each stage is given in advance, and the work time limit included in the work condition data 96 is replaced with the numerical value corresponding to the selected stage.

[0028] The acceptable physical load range, which is one of the working condition information 962, includes the upper limit of the worker's energy consumption and the upper limit of joint torque. Energy consumption is a function of activity intensity, activity time, and body weight, and is an index of fatigue. Joint torque is the force required to rotate a joint, and is an index of load intensity.

[0029] Like the physical function data 92, the physical function information 963 includes information that can identify the physical functions of the worker using the motion plan generated by the work motion planning system 1, such as the body dimensions, age, gender, dominant hand, and whether or not the worker has any limb disabilities.

[0030] Similar to the motion measurement data 91, the motion plan data 95 is a chronological representation of quantified joint movements of the user during work. The joint movements are represented, for example, by the position and posture of the joint. The joint position can be represented by three-dimensional coordinates, and the joint posture can be represented by the joint rotation angle. The motion plan specified by the motion plan data 95 output from the AI ​​model 50 consists of movements that allow the user to execute the physical functions specified by the physical function information 963. If the user operates in accordance with the motion plan specified by the motion plan data 95 output from the AI ​​model 50, the motion plan will enable the user to complete the work within a work time limit, and the physical load will be within the allowable physical load range, thereby satisfying the work condition information 962 included in the work condition data 96.

[0031] The motion plan generation unit 52 inputs task condition data 96 of the user who will use the motion plan, and generates motion plan data 95 that is optimal for the user, using the trained AI model 50. The motion plan data 95 generated by the motion plan generation unit 52 is stored in the data accumulation device 4.

[0032] <Motion teaching device 6> The action teaching device 6 acquires the action plan data 95 generated by the action design device 5, and teaches the user, who is a learner, the movements of the task in accordance with the action plan data 95. The action teaching device 6 according to this embodiment teaches the user the movements by fusing the real world and the virtual world using XR (cross reality) technology to allow the learner to virtually experience the movements.

[0033] Examples of XR technologies include VR (virtual reality) and MR (mixed reality). VR is a technology that allows users to experience a virtual world as if it were the real world, and users can virtually experience work movements by viewing images created by CG (computer graphics) through VR devices such as head-mounted displays. MR is a technology that blends real space with the virtual world, and by using cameras and sensors built into MR devices such as head-mounted displays, users can virtually experience work movements by viewing images that combine virtual images created by CG with the real space captured by the camera.

[0034] The motion teaching device 6 includes an image generation unit 61 and an output device 62 having a display such as a VR device or an MR device. The image generation unit 61 acquires motion plan data 95 generated by the motion design device 5 and creates a virtual image using CG based on the motion plan data 95. In the case of VR, the image generation unit 61 generates a virtual image of an avatar moving based on the motion plan data 95, outputs the virtual image to the output device 62, and displays the virtual image on the output device 62. A user can learn the motions of a task by tracing the movements of the avatar displayed on the output device 62. In the case of MR, the image generation unit 61 generates a virtual image of an avatar or a task object moving based on the motion plan data 95, and the image generation unit 61 or the output device 62 fuses an image of the real world captured by a camera attached to the output device 62 with the virtual image generated by the image generation unit 61, and displays the fused image on the output device 62. The user can learn the movements of the work by tracing the movements of the avatar displayed on the output device 62 or by virtually moving the work object.

[0035] The components of the work motion planning system 1, namely, the motion measurement device 2 (particularly, the recording unit 22 and the motion information generation unit 23), the load analysis device 3, the motion design device 5, and the motion teaching device 6 (particularly, the image generation unit 61), can be implemented, for example, by a computer. The computer includes, for example, a CPU (Central Processing Unit), memory, an auxiliary storage device, a communication I / F for connecting to a communication network via wired or wireless connection, input devices such as a mouse, keyboard, and touch panel, output devices such as a display and printer, and a media I / F for reading and writing information from and to a portable storage medium. The functions of each component can be realized by loading a predetermined program stored in the auxiliary storage device into memory and executing it on the CPU. The predetermined program may be downloaded from a network via the communication I / F or loaded from a storage medium connected to the media I / F.

[0036] [Summary] The work operation planning system 1 according to the first aspect of the present disclosure includes: a load analysis device 3 that acquires motion measurement data 91 that records the motion of a worker in chronological order and physical function data 92 that identifies the physical functions of the worker, generates a digital human model DH that reproduces the motion of the worker based on the motion measurement data 91 and the physical function data 92, analyzes the physical load imposed on the worker in the motion of the worker using the digital human model DH, and generates physical load data 93 in which the physical load is quantified; a data accumulation device (4) that stores a combination of motion measurement data (91), physical function data (92), and physical load data (93) as learning data (94); The action design device 5 has a trained model 50 that has been trained using training data 94 so as to output, when input, work condition data 96 including work condition information 962 including allowable physical load range information and work time limits, and physical function information 963, action plan data 95 that represents, in time series, movements during work that satisfy the work condition information 962 and are possible with the physical functions identified from the physical function information 963, and generates action plan data 95 for a user from the user's work condition data 96 using the trained model 50.

[0037] Further, a work operation planning method according to a third aspect of the present disclosure includes: Obtaining motion measurement data 91 that records the motion of a worker during work in chronological order and physical function data 92 that identifies the physical functions of the worker, generating a digital human model DH that reproduces the motion of the worker based on the motion measurement data 91 and the physical function data 92, analyzing the physical load imposed on the worker in the motion of the worker using the digital human model DH, and generating physical load data 93 in which the physical load is quantified; Storing a combination of the motion measurement data 91, the physical function data 92, and the physical load data 93 as learning data 94 in the data accumulation device 4; and When work condition data 96 including work condition information 962 including physical load tolerance range information and work time limit information and physical function information 963 is input, action plan data 95 for a user is generated from the user's work condition data 96 using a trained model 50 trained with training data 94 so as to output action plan data 95 that represents, in time series, movements during work that satisfy the work condition information 962 and are possible with the physical functions identified from the physical function information 963.

[0038] The above-described work motion planning system 1 and method can provide motion plan data 95 that quantitatively indicates ideal work movements for the user. The motion plan of the motion plan data 95 is customized for the user, taking into account the user's physical functions based on the user's age, body type, etc., and therefore can be executed within a reasonable range for the user. Furthermore, by executing the motion plan within the range of the user's physical functions, the user can complete the work within the work time limit and with a physical load within the allowable range. In other words, the above-described work motion planning system 1 and method can provide a work motion plan that takes into account both the user's physical load and work efficiency. Furthermore, by presenting such a motion plan, it is possible to objectively evaluate the user's work, which is expected to facilitate agreement between labor and management.

[0039] The work motion planning system 1 according to the second item of the present disclosure is the work motion planning system 1 according to the first item, further comprising a motion teaching device 6 having an image generation unit 61 that acquires a user's motion planning data 95 and generates a virtual image of an avatar or work object moving based on the motion planning data 95, and an output device 62 including a display on which the virtual image is displayed.

[0040] Furthermore, the work motion planning method according to the fourth item of the present disclosure is the work motion planning method according to the third item, further comprising the steps of acquiring user motion planning data 95, generating a virtual image in which an avatar or work object moves based on the motion planning data 95, and displaying and outputting the virtual image on a display.

[0041] According to the above-mentioned work motion planning system 1 and method, the user can learn ideal movements for the user by tracing the movements of the avatar or work object displayed on the display of the output device 62, and can then put the learned movements into practice. Furthermore, the above-mentioned work motion planning system 1 and method uses XR technology, which is familiar to young people, to teach motion planning, and is therefore expected to increase the motivation and work motivation of young people in particular to work in the manufacturing industry.

[0042] The functions performed by the components of the work motion planning system 1 described herein may be implemented in circuitry or processing circuitry, including general-purpose processors, special-purpose processors, integrated circuits, application-specific integrated circuits (ASICs), central processing units (CPUs), conventional circuits, and / or combinations thereof, programmed to perform the described functions. Processors include transistors and other circuits and are considered circuitry or processing circuitry. A processor may also be a programmed processor that executes a program stored in memory. In this specification, a circuit, unit, or means is hardware that is programmed to perform or executes the described functions. The hardware may be any hardware disclosed herein or any hardware known to be programmed to perform or execute the described functions. When the hardware is a processor, which is considered a type of circuitry, the circuit, means, or unit is a combination of hardware and software used to configure the hardware and / or processor.

[0043] The above-described embodiments have been presented for purposes of illustration and description and are not intended to limit the present disclosure to the form disclosed herein. For example, in the foregoing detailed description, various features of the present disclosure are grouped together in a single embodiment for the purpose of streamlining the disclosure, but some of the features may also be combined. Furthermore, the features included in the present disclosure may also be combined into alternative embodiments, configurations, or aspects other than those discussed above. [Explanation of symbols]

[0044] 1: Work motion planning system 2: Motion measurement device 3:Load analysis device 4: Data collection device 5: Motion design device 6: Motion teaching device 50: AI model (trained model) 61: Image generation unit 62: Output device 91: Motion measurement data 92: Physical function data 93: Physical load data 94: Training data 95: Motion planning data 96: Working condition data 962: Working condition information 963 :Physical function information DH: Digital Human Model

Claims

1. a load analysis device that acquires motion measurement data that records the motion of a worker during work in chronological order and physical function data that identifies the physical functions of the worker, generates a digital human model that reproduces the motion of the worker based on the motion measurement data and the physical function data, analyzes the physical load imposed on the worker in the motion of the worker using the digital human model, and generates physical load data in which the physical load is quantified; a data accumulation device that stores a combination of the motion measurement data, the physical function data, and the physical load data as learning data; an action design device that has a trained model trained with the learning data so as to output, when inputting work condition data including work condition information, including physical load tolerance information and work time limits, and physical function information, action plan data that represents, in time series, movements during the work that satisfy the work condition information and are possible with physical functions identified from the physical function information, and that generates the action plan data for the user from the work condition data of the user using the trained model; Work motion planning system.

2. The motion teaching device further includes an image generation unit that acquires the motion plan data of the user and generates a virtual image in which an avatar or a work object moves based on the motion plan data, and an output device that includes a display that displays and outputs the virtual image. The work operation planning system according to claim 1 .

3. acquiring motion measurement data that records the motion of a worker during work in chronological order and physical function data that identifies the physical functions of the worker, generating a digital human model that reproduces the motion of the worker based on the motion measurement data and the physical function data, analyzing the physical load imposed on the worker in the motion of the worker using the digital human model, and generating physical load data in which the physical load is quantified; storing a combination of the motion measurement data, the physical function data, and the physical load data as learning data in a data accumulation device; and generating the motion plan data for a user from the work condition data of the user using a trained model trained with the learning data so as to output, when work condition data including work condition information including allowable physical load range information and work time limits and physical function information, motion plan data representing, in time series, movements during the work that satisfy the work condition information and are possible with physical functions identified from the physical function information; Work motion planning method.

4. acquiring the motion plan data of the user, and generating a virtual image in which an avatar or a work object moves based on the motion plan data; and further comprising displaying and outputting the virtual image on a display. The work motion planning method according to claim 3 .

Images