Analysis device and analysis method

Abstract

This analysis device for analyzing a work operation is constituted by a computer including an arithmetic unit for executing predetermined arithmetic processing and a storage device connected to the arithmetic unit, and comprises: a shape input unit for receiving an input of shape information of a work object; an object position / posture input unit for receiving an input of position / posture information of the work object; an operation input unit for receiving an input of data of an operation of a work by a working agent; a basic shape extraction unit for extracting a basic shape constituting a shape of the work object by means of the arithmetic unit; a work operation conversion unit for converting an operation of the working agent into a relative operation from a coordinate system of the work object by means of the arithmetic unit; and an operation decomposition unit for decomposing the relative operation into motions belonging to the basic shape by means of the arithmetic unit and outputting a momentum of each decomposed motion component.

Description

Analysis Device and Analysis Method Incorporation by Reference

[0001] This application claims the priority of Japanese Patent Application No. 2025-11256, which was filed on January 27, 2025 (the year of Reiwa 7), and incorporates its content by reference.

[0002] The present invention relates to an analysis device that analyzes the operations of an operator.

[0003] An action analysis method that measures the movements of an object or a work subject, three-dimensionally reconstructs the measured movements, and evaluates the work efficiency in detail is used in various fields such as device design.

[0004] As the background art in this technical field, there are the following prior arts. Patent Document 1 (Japanese Patent Laid-Open No. 2020-177680) discloses a device for analyzing the work performed by an operator, which includes a position information acquisition unit that acquires first position information, which is depth image data output from a depth sensor that photographs the operator, and second position information, which is image data output from a digital camera that photographs the object; an operation identification unit that identifies the position of the operator's hand from the first position information, identifies the position of the object from the second position information, and identifies the content of the operation performed by the operator in the work based on the identified position of the operator's hand and the identified position of the object.

[0005] Currently, data on on-site work is measured without omission, and work analysis engineers individually consider what to focus on in the work, evaluating it in terms of feature quantities and importance, and the man-hours required for evaluation are increasing. In order to apply work analysis to a large number of processes, it is necessary to automatically extract the features to be analyzed.

[0006] A representative example of the invention disclosed in this application is as follows: an analytical device for analyzing work movements, comprising a computer having a computing device that performs predetermined calculation processing and a storage device connected to the computing device, the device comprising: a shape input unit that receives shape information of a work object; an object position and orientation input unit that receives position and orientation information of the work object; an motion input unit that receives data of work movements performed by a worker; a basic shape extraction unit in which the computing device extracts basic shapes that constitute the shape of the work object; a work motion conversion unit in which the computing device converts the movements of the worker into relative movements from the coordinate system of the work object; and a motion decomposition unit in which the computing device decomposes the relative movements into movements belonging to the basic shapes and outputs the momentum of each decomposed motion component.

[0007] According to one aspect of the present invention, work evaluation features can be automatically generated from work actions. Problems, configurations, and effects other than those described above will be clarified by the following description of the embodiments.

[0008] This is a block diagram showing the logical configuration of the analytical device according to an embodiment of the present invention. This is a block diagram of the computing device constituting the analytical device according to an embodiment of the present invention. This is a flowchart of the analysis process according to an embodiment of the present invention. This diagram shows the basic shape extraction process according to an embodiment of the present invention. This diagram shows an example of the output of the motion component output unit according to an embodiment of the present invention. This diagram shows the importance analysis process according to an embodiment of the present invention. This diagram shows an example of the output of the importance analysis process according to an embodiment of the present invention. This diagram shows an example of the deviation alert screen according to an embodiment of the present invention.

[0009] Figure 1 is a block diagram showing the logical configuration of the analysis device 100 in this embodiment.

[0010] The analysis device 100 of this embodiment includes a work subject motion input unit 101, an object position and orientation input unit 102, a work motion conversion unit 103, a shape input unit 104, a basic shape extraction unit 105, a motion decomposition unit 106, a motion component output unit 107, a work information input unit 108, a work importance analysis unit 109, a similar work extraction unit 110, a model motion generation unit 111, a spatial model input unit 112, a body model input unit 113, a comprehensive work score calculation unit 114, a comprehensive work score output unit 115, a deviation determination unit 116, a deviation alert image generation unit 117, and a video output unit 118.

[0011] The worker motion input unit 101 receives a camera image 201 showing the worker's movements, extracts hand skeletal motion information from the input camera image 201, and outputs the extracted hand skeletal motion information to the work motion conversion unit 103. The camera image 201 is a moving image captured by a depth image camera, with color and distance data attached to each pixel. For each frame of the moving image, the image information with color information attached is used to determine the position of each skeletal position in the image using an image processing AI (Artificial Intelligence) model that has learned the correspondence between the color image and the image coordinates of each feature point constituting the hand, i.e., the wrist and each finger joint. Next, the distance information corresponding to the image coordinates is used to convert it into three-dimensional coordinates of each skeletal position in the camera coordinate system. The camera-specific intrinsic parameter matrix for converting from the image coordinate system to the camera coordinate system is obtained by calibration. In this case, the camera image is not limited to one camera; the position of each skeleton in the camera coordinate system for two or more simultaneously captured camera images may be estimated by an image processing AI model, and then converted into three-dimensional coordinates of each skeleton position using the external parameter matrices of the two cameras that have been determined in advance. With this method, since only color information is used for each camera image, distance data does not need to be attached. The work subject motion input unit 101 may receive time-series data of position and posture indicating the work actions of the work subject, rather than the camera image 201. The work subject is a worker or work robot handling an object. The worker's actions are determined from moving images of the work taken with a depth image camera capable of acquiring distance information for each pixel, or from the actions of a body model analyzed from measurement results by a sensing device worn by the worker (for example, a sensor glove that detects the movement of the worker's fingers).

[0012] The object position and orientation input unit 102 receives a camera image 201 of an object related to the work (e.g., a work object), extracts object motion information from the input camera image 201, and outputs the extracted object motion information to the work motion conversion unit 103. Markers are attached to the object related to the work in predetermined locations, and a correspondence relationship between the object related to the work and the markers is defined. The object motion information is obtained by detecting the area of ​​the markers from the camera image 201 and estimating the marker pattern and the position of the markers in the camera coordinate system. If it is not possible to attach markers for reasons such as the object being too small, the object motion information may be estimated using an image processing AI model that has learned the correspondence relationship between the object's orientation and images from various directions generated using a video of the object or a three-dimensional model corresponding to the object. The object position and orientation input unit 102 may also receive time-series data of the position and orientation of the object related to the work (e.g., a work object) instead of a camera image 201. The time-series data of the position and orientation of the object related to the work may be represented by the movement of points in an object model that geometrically represents the object.

[0013] The work motion conversion unit 103 converts the movements of the worker (body model) input to the work subject motion input unit 101 from an absolute coordinate system to relative movements in a relative coordinate system. Specifically, it performs a process of subtracting a three-dimensional vector representing the relative position between the camera coordinate system and the object so that the hand skeletal motion information in the camera coordinate system becomes the relative position with respect to the position of the object motion information at the same time.

[0014] The shape input unit 104 receives an object model representing the shape of an object related to the work. The object model is a geometric model that represents the object related to the work.

[0015] The basic shape extraction unit 105 decomposes the object model input to the shape input unit 104 into basic shapes and extracts one or more basic shapes from the object model representing the shape of the object related to the work.

[0016] The motion decomposition unit 106 decomposes the relative motion converted by the work motion conversion unit 103 into motion components belonging to the basic shape extracted by the basic shape extraction unit 105. If the amount of relative motion is smaller than a predetermined threshold, the motion decomposition unit 106 may determine that the part of the work subject and the work object are moving together, and decompose the motion of the work object into motions belonging to the basic shape.

[0017] The motion component output unit 107 outputs the motion components that have been decomposed by the motion decomposition unit 106. An example of the output of the motion component output unit 107 will be described later with reference to Figure 5.

[0018] The work information input unit 108 receives information about the work to be analyzed. The work information input to the work information input unit 108 includes the importance information of the motion components of each work process included in the work.

[0019] The work importance analysis unit 109 evaluates the work quality in each of the motion components 213 decomposed by the motion decomposition unit 106, and, referring to the work importance information 214, evaluates the importance of each motion component based on the work quality evaluation results.

[0020] The similar work extraction unit 110 uses the values ​​of each motion component analyzed from the worker's immediate preceding action to extract similar work motion information from previously recorded high-quality work records.

[0021] The model motion generation unit 111 generates display data to display high-quality work motions that are closest to the work information whose motion components have been evaluated. This is done by combining a spatial model representing the shape information of the workspace at the time of the work (structures such as walls and desks) input in the spatial model input unit 112, a body model of the work subject input in the body model input unit 113, and an object model representing the shape of objects related to the work input in the shape input unit 104, based on high-quality work motion information recorded in the past and extracted by the similar work extraction unit 110.

[0022] The spatial model input unit 112 receives input of a spatial model representing the shape of the workspace during the operation.

[0023] The body model input unit 113 receives input of a body model representing the worker or work robot that is the subject of the work.

[0024] The overall work score calculation unit 114 uses the decomposed motion components 213 and work importance information 214 to calculate the overall work score from the amount of motion of the motion components belonging to the decomposed basic shape.

[0025] The overall work score output unit 115 generates display data for displaying the calculated overall work score.

[0026] The deviation determination unit 116 uses the decomposed motion components 213 and the work importance information 214 to calculate an evaluation value for each motion component, and determines that there has been a deviation in the work if the evaluation value weighted by importance is less than or equal to a predetermined value.

[0027] The deviation alert image generation unit 117 generates display data for a deviation alert image, which includes the display of the parts of the work that need to be corrected, as shown on the deviation alert screen 800.

[0028] The video output unit 118 outputs the display data generated by the model motion generation unit 111, the overall work score output unit 115, and the deviation alert image generation unit 117 as video.

[0029] Figure 2 is a block diagram of the computer that constitutes the analysis apparatus 100 of this embodiment.

[0030] The analysis device 100 in this embodiment consists of a computer having a processor (CPU) 11, ROM 12, RAM 13, auxiliary storage device 14, input / output device 15, and communication device 16. The analysis device 100 is preferably implemented in a general-purpose computer or a terminal device such as a tablet.

[0031] The processor 11 is an arithmetic unit that executes programs stored in memory. By executing various programs, the processor 11 realizes the functions of each functional unit of the analysis device 100. Note that some of the processing performed by the processor 11 when executing programs may be performed by other arithmetic units (for example, hardware such as ASICs or FPGAs).

[0032] The memory includes a non-volatile memory element, ROM 12, and a volatile memory element, RAM 13. ROM 12 stores immutable programs (e.g., BIOS). RAM 13 is a high-speed, volatile memory element, such as DRAM (Dynamic Random Access Memory), and temporarily stores programs executed by the processor 11 and data used during program execution.

[0033] The auxiliary storage device 14 is, for example, a high-capacity, non-volatile storage device such as a flash memory (SSD) or a magnetic storage device (HDD). The auxiliary storage device 14 also stores data used by the processor 11 when executing a program, and the program that the processor 11 executes. In other words, the program is read from the auxiliary storage device 14, loaded into the RAM 13, and executed by the processor 11 to realize the various functions of the analysis device 100.

[0034] The communication device 16 is a network interface device that controls communication with other devices according to a predetermined protocol.

[0035] The input / output device 15 includes an input device, a camera, and an output device. The input device is an interface such as a touch panel that receives input from the operator. The output device is an interface such as a display device that outputs the program execution results in a format that can be viewed by the user.

[0036] The program executed by the processor 11 is provided to the analyzer 100 via removable media (such as a CD-ROM or flash memory) or a network, and stored in the non-volatile auxiliary storage device 14, which is a non-temporary storage medium. For this reason, the analyzer 100 should have an interface for reading data from the removable media.

[0037] The analysis device 100 may not be a terminal device, but rather a computer system composed of multiple logically or physically configured computers, or it may operate on a virtual computer built on multiple physical computer resources. For example, the multiple programs that realize the functions of the analysis device 100 may each operate on separate physical or logical computers, or multiple programs may be combined and operate on a single physical or logical computer. In this case, a user terminal connected to the analysis device 100 via a network may provide input and output devices.

[0038] Figure 3 is a flowchart of the analysis process performed by the analysis device 100 in this embodiment. The analysis process shown in Figure 3 is performed by the processor 11 of the analysis device 100 executing a program.

[0039] First, a camera image 201, captured by a depth image camera, is input to the work subject motion input unit 101 and the object position and orientation input unit 102.

[0040] The work-oriented motion input unit 101 extracts hand skeletal motion information from the input camera image 201 and outputs the extracted hand skeletal motion information to the work motion conversion unit 103 (202).

[0041] The object position and orientation input unit 102 extracts object motion information from the input camera image 201 and outputs the extracted object motion information to the work motion conversion unit 103 (203).

[0042] Furthermore, object shape information 204 is input to the shape input unit 104. The input object shape information 204 may be an object model representing the shape of the object. The object shape information 204 is output to the work motion conversion unit 103 and the basic shape extraction unit 105.

[0043] The basic shape extraction unit 105 decomposes the input object shape information 204 into basic shapes, and extracts one or more basic shapes from the object model representing the shape of the object related to the work (205). For example, in the modeling of three-dimensional computer graphics, a complex shape can be constructed by the Boolean operation of simple basic shapes called CSG (Constructive Solid Geometry) representation. The Boolean operation considers a three-dimensional shape with volume as a set, and creates a combined shape by combining multiple shapes through set operations such as sum, difference, and product, and creates a divided shape. The input object shape information 204 is converted into a CSG representation expressed by a Boolean operation in a plurality of basic shapes by using a learned AI model. The basic shape of this embodiment is associated with a motion component. The details of the basic shape extraction process executed by the basic shape extraction unit 105 will be described later with reference to FIG. 4.

[0044] In addition, the work operation conversion unit 103 converts the operation of the worker (body model) input to the work subject operation input unit 101 into a relative operation in a relative coordinate system with a point of the work object as the origin from the absolute coordinate system, analyzes the relative operation and the contact amount with respect to the work object, and outputs the mutual relationship information 207 to the operation decomposition unit 106 (206).

[0045] The operation decomposition unit 106 determines whether the relative operation amount is greater than or equal to a threshold value by using the mutual operation information 208 representing the relative operation between the work object and the work subject included in the mutual relationship information 207 (209). If the relative operation amount between the work object and the work subject is greater than or equal to a predetermined threshold value, it is determined that the worker is performing some work on the work object, and the relative operation information is input to the operation decomposition unit 106 (210). On the other hand, if the relative operation amount between the work object and the work subject is smaller than the predetermined threshold value, it is determined that the specific part of the worker and the work object are moving together, and the relative operation information is input to the operation decomposition unit 106 as an object operation (211).

[0046] Then, the operation decomposition unit 106 decomposes the relative operation converted by the work operation conversion unit 103 into motions belonging to the basic shapes decomposed by the basic shape extraction unit 105, and outputs the values 213 of the decomposed motion components to the work importance analysis unit 109 (212). Also, when the amount of the relative operation is smaller than a predetermined threshold, the operation decomposition unit 106 decomposes the operation of the object into motions belonging to the basic shape. For example, when the arc direction belongs to the basic shape of the work object, the operation is decomposed into motions in which the work object moves in the arc direction and the normal direction with respect to the arc surface. Also, when the basic shape of the work object is a cube, the operation is decomposed into motions in which the part is moved within the plane of the cube and in the normal direction with respect to the plane.

[0047] The work importance analysis unit 109 evaluates the work quality in each of the motion components 213 decomposed by the operation decomposition unit 106, and evaluates the importance of each motion component from the evaluation result of the work quality by referring to the work importance information 214 (215). Details of the importance analysis process executed by the work importance analysis unit 109 will be described later with reference to FIG. 6.

[0048] Then, the similar work extraction unit 110 extracts similar works (216).

[0049] The total work score calculation unit 114 calculates the total work score from the operation amounts of the motion components belonging to the decomposed basic shape using the decomposed motion components 213 and the work importance information 214 (217).

[0050] The deviation determination unit 116 calculates the evaluation value (for example, the work success probability in FIG. 7) of each motion component using the decomposed motion components 213 and the work importance information 214, and determines that the work deviates when the importance is greater than or equal to a predetermined value and the evaluation value weighted by the importance is less than or equal to a predetermined value (218).

[0051] Subsequently, the model motion generation unit 111 generates display data to display the high-quality work that is closest to the work information evaluated for its motion components as the motion of the body model. The overall work score output unit 115 generates display data to display the calculated overall work score. The deviation alert image generation unit 117 generates display data to display the deviation judgment result that will be played back on the deviation alert screen 800 (219).

[0052] Figure 4 shows the basic shape extraction process performed by the basic shape extraction unit 105.

[0053] The basic shape extraction unit 105 extracts the basic shape from the object shape information 204. For example, in the example shown in Figure 4, the workpiece 1 is a rectangular parallelepiped 2 with a cylindrical hole A 3A and a cylindrical hole B 3B. That is, the shape of the workpiece 1 is a combination of a rectangular parallelepiped 2 and cylinders 3A and 3B of equal height.

[0054] The basic shape extraction unit 105 extracts one rectangular prism 1 and two cylinders 3A and 3B from the object shape information 204 of the work object 1 using a trained AI model that converts complex shapes into CGS representations constructed by Boolean operations.

[0055] The basic shape extraction unit 105 then refers to the motion components belonging to the basic shapes and identifies the motion components belonging to the rectangular parallelepiped 2 and the cylinders 3A and 3B. The motion components of the basic shapes are a classification of the movements of the worker during work performed on the work object of the basic shape. In the example shown in Figure 4, the motion components of the rectangular parallelepiped 2 are the three orthogonal x, y, and z axes parallel to the sides of the rectangular parallelepiped 2 (perpendicular to the faces of the rectangular parallelepiped 2), and the motion components of the cylinders are the three directions: the θ direction of the arc along the side of the cylinder, the r direction of the diameter of the circular base, and the z direction perpendicular to the circular base.

[0056] For example, in the task of wiping a disc-shaped workpiece, uniform wiping can be achieved by having the main part of the workpiece move at a predetermined speed in the θ direction of the arc. Therefore, the average value of the movement speed in the θ direction is calculated. For example, in a wiping operation that includes forward and reverse movements, the average value of the movement speed in the θ direction will be 0 or close to 0, and it can be determined that the operation is not appropriate.

[0057] Figure 5 shows an example of output from the motion component output unit 107. While Figure 5 shows an example of tabular data output, data may be output in other formats.

[0058] The motion decomposition unit 106 decomposes the relative motion between the work subject and the work object into motion components belonging to the basic shape, and the motion component output unit 107 outputs values ​​indicating the motion in the decomposed motion components.

[0059] In the operation 1, which involves assembling a part onto the workpiece 1 shown in Figure 4, the movement in the y-direction belonging to the rectangular parallelepiped 2 determines the position of the part to be assembled, and the movement in the z-direction places the part onto the workpiece. In operation 1, the motion components belonging to the cylinders 3A and 3B are not extracted.

[0060] Next, in operation 2, where a screw is tightened in hole A, the movement of the screw in the r-direction belonging to cylinder 3A moves the screw to the position of hole A, the movement in the θ-direction rotates the screw in hole A, and the movement in the z-direction inserts the screw into hole A. In operation 2, the motion components belonging to the rectangular parallelepiped 2 and cylinder 3B are not extracted.

[0061] Next, in operation 3, where a screw is tightened in hole B, the movement of the screw in the r-direction belonging to cylinder 3B moves the screw to the position of hole B, the movement in the θ-direction rotates the screw in hole B, and the movement in the z-direction inserts the screw into hole B. In operation 3, the motion components belonging to the rectangular parallelepiped 2 and cylinder 3A are not extracted.

[0062] Figures 6 and 7 show the importance analysis process performed by the work importance analysis unit 109.

[0063] The work importance analysis unit 109 evaluates the work quality in each of the motion components 213 decomposed by the motion decomposition unit 106, and, referring to the work importance information 214, evaluates the importance of each motion component based on the work quality evaluation results.

[0064] For example, as shown in Figure 6, work quality evaluation data is created by dividing the values ​​of each motion component in previously recorded work into a work success group and a work failure group, and the importance of the work is determined by the significance probability of the difference between the average value of the motion component of the work success group and the average value of the motion component of the work failure group. The significance probability is the probability that the result can be judged as having been obtained assuming that there is actually no difference in the average values ​​between the work success group and the work failure group. The smaller the significance probability, the more likely it is that there is a difference between the work success group and the work failure group in the motion component of interest. For example, the importance of the work can be determined by classifying the significance probabilities at a predetermined threshold. For example, in a wiping operation, the work success group and the work failure group can be classified by comparing the number of foreign objects remaining on the surface after wiping with a predetermined threshold. Also, if the information representing the success / failure of the work can be quantitatively evaluated, the correlation coefficient with each motion component is calculated, and the motion component with a large absolute value of the correlation coefficient is judged to be an important component.

[0065] Then, as shown in Figure 7, the overall work score calculation unit 114 calculates the overall work score using the significance probabilities calculated for each motor component by the work importance analysis unit 109. For example, the overall work score is calculated as the average value of the work success probabilities weighted by importance, which is the reciprocal of the significance probability.

[0066] Figure 8 shows an example of a deviation alert screen 800 output by the video output unit 118.

[0067] The deviation alert screen 800 includes a work display area 810, a message display area 820, and an overall work score display area 830.

[0068] In the work display area 810, the movement of the body model 5, which reflects the movements in the work process and is generated by the deviation alert image generation unit 117, is displayed together with the object model of the work object 1. The cylinder 3A, which is the basic shape extracted by the basic shape extraction unit 105 associated with the motion component that detected the deviation and is the part of the work that needs to be corrected, is highlighted and displayed. The message display area 820 displays the message of the deviation judgment result determined by the deviation judgment unit 116. By presenting the deviation judgment result to the worker, an opportunity for correction can be provided. The deviation alert screen 800 can be used in either a method of displaying the result based on the deviation judgment result made on the spot based on real-time hand skeletal movement information obtained during work, or a method of displaying it at any timing using hand skeletal movement information recorded after the completion of work for the purpose of work analysis.

[0069] The overall work score display area 830 displays the overall work score calculated by the overall work score calculation unit 114. By presenting the overall work score to the worker, the work level can be improved.

[0070] According to one aspect of the present invention, since motion components, which are work evaluation features, are automatically generated based on work motion and shape data of the work object, work analysis can be applied to a wide range of tasks in a short period of time, and tasks can be analyzed individually and evaluated with high accuracy. Furthermore, since individual tasks are evaluated in relation to the shape of the objects related to the task, intuitive work support can be provided. In other words, the following effects can be obtained in one embodiment of the present invention.

[0071] Quantitative evaluation of work quality: The worker's actions are broken down into components belonging to the basic shape of the object being worked on (such as rotational components in the arc direction around holes, and components parallel or perpendicular to the plane). The values ​​of each motion component in the work being evaluated are then assessed using statistical values ​​from previously recorded work, and an overall work score is calculated. This allows for feedback to be given to the worker regarding the overall quality of the work.

[0072] Work deviation warning: Compare the movement of the work component being evaluated with the movement of work components recorded in the past. This allows the worker to be shown the possibility of a work deviation.

[0073] Information presentation for work correction: For actions that may deviate from the work process, the basic shape of the work object related to the potentially deviating action component is visually presented along with a reproduction of the worker's action. This provides the worker with an opportunity to correct their work.

[0074] Automation of tasks: While the above describes the work performed by human workers, the present invention can also be applied to tasks performed by robots. When automating tasks with a robot, it is advisable to search for the optimal movements so that the deviation of each movement component in past work records falls within a predetermined value. This enables high work quality to be achieved with the robot.

[0075] Work support: By using the values ​​of each motion component analyzed from the worker's immediate preceding movements, similar work motion information is extracted from previously recorded high-quality work records. This allows for the visual presentation of work examples that are easy for workers to emulate.

[0076] It should be noted that the present invention is not limited to the embodiments described above, but includes various modifications and equivalent configurations within the spirit of the attached claims. For example, the embodiments described above are described in detail to make the present invention easier to understand, and the present invention is not necessarily limited to having all the configurations described above. Furthermore, some of the configurations of one embodiment may be replaced with those of another embodiment. Furthermore, some of the configurations of one embodiment may be added to those of another embodiment. Furthermore, some of the configurations of each embodiment may be added, deleted, or replaced with those of other embodiments.

[0077] Furthermore, each of the aforementioned configurations, functions, processing units, and processing means may be implemented in hardware, for example, by designing them as integrated circuits, or they may be implemented in software by having a processor interpret and execute programs that realize each function.

[0078] Information such as programs, tables, and files that implement each function can be stored in memory, hard disks, SSDs (Solid State Drives), or recording media such as IC cards, SD cards, and DVDs.

[0079] Furthermore, the control lines and information lines shown are those deemed necessary for explanation purposes and do not necessarily represent all control lines and information lines required for implementation. In reality, it can be assumed that almost all components are interconnected.

Claims

1. An analytical device for analyzing work movements, comprising a computer having a calculation device that performs predetermined calculation processing and a storage device connected to the calculation device, the analytical device comprising: a shape input unit that receives shape information of a work object; an object position and orientation input unit that receives position and orientation information of the work object; an motion input unit that receives data of work movements by a worker; a basic shape extraction unit in which the calculation device extracts basic shapes that constitute the shape of the work object; a work motion conversion unit in which the calculation device converts the movements of the worker into relative movements from the coordinate system of the work object; and a motion decomposition unit in which the calculation device decomposes the relative movements into movements belonging to the basic shapes and outputs the momentum of each decomposed motion component.

2. An analytical apparatus according to claim 1, wherein the basic shape extraction unit approximates the work object with a plurality of basic shapes and extracts the basic shape.

3. An analytical apparatus according to claim 1, characterized in that it comprises an importance analysis unit that evaluates the importance of the movement component using work importance information created from work quality evaluation data that evaluates the work quality in each work movement.

4. An analytical apparatus according to claim 3, characterized in that it comprises a score calculation unit that calculates an overall work score for evaluating work movements using the work importance information and the motion components.

5. An analytical apparatus according to claim 3, characterized in that it comprises a deviation determination unit that determines deviations from work using the work importance information and the motion components.

6. The analytical apparatus according to claim 5, wherein the deviation determination unit calculates an evaluation value for each motion component using the work importance information and the motion components, and determines that the work is a deviation when the importance is greater than or equal to a predetermined value and the evaluation value is less than or equal to a predetermined value.

7. An analysis apparatus according to claim 5, characterized in that it comprises a deviation alert image generation unit that generates display data for a deviation alert image, which includes an indication of the part of the work that should be corrected, based on the deviation determination result.

8. An analytical apparatus according to claim 1, wherein the motion decomposition unit decomposes the motion of the work object into motions belonging to the basic shape when the amount of the relative motion is less than a predetermined threshold.

9. An analysis apparatus according to claim 4, wherein the work subject is a robot, and the score calculation unit calculates an overall work score for evaluating the actions of the robot in manipulating a work object using the work importance information and the motion components.

10. The analytical apparatus according to claim 3, wherein the work quality evaluation data is data in which a plurality of motion data are classified into a work success group and a work failure group, and the work importance information is calculated by the statistically significant difference between the work success group and the work failure group in each motion component.

11. An analysis apparatus according to claim 1, wherein the main performer is a person, and the apparatus is equipped with a video output unit that displays the results of the analysis of the movements of the work.

12. An analysis method comprising an analysis device for analyzing work motion, wherein the analysis device is composed of a computer having a computing device for performing predetermined calculation processing and a storage device connected to the computing device, and the analysis method comprises: a shape input procedure in which the computing device receives shape information of a work object; an object position and orientation input procedure in which the computing device receives position and orientation information of the work object; an motion input procedure in which the computing device receives data of work motion by a worker; a basic shape extraction procedure in which the computing device extracts basic shapes that constitute the shape of the work object; a work motion conversion procedure in which the computing device converts the motion of the worker into relative motion from the coordinate system of the work object; and an motion decomposition procedure in which the computing device decomposes the relative motion into motions belonging to the basic shapes and outputs the momentum of each decomposed motion component.

Images