Estimation device and estimation system

Abstract

The system appropriately detects the strength of the worker's fingers while suppressing interference with the worker's work caused by the sensor. [Solution] The estimation device comprises an acquisition unit which acquires a first detection result from a first sensor that is provided away from the palm side of the worker's hand and detects a first physical quantity related to the flexion of the worker's fingers, and a second detection result from a second sensor that is attached to the worker's forearm and detects a second physical quantity related to the movement of the forearm muscles, and an estimation unit which uses the first detection result and the second detection result to estimate the finger force generated in the fingers.

Description

Technical Field

[0001] The present disclosure relates to an estimation device and an estimation system.

Background Art

[0002] Patent Document 1 discloses a wearable sensor having a sensor that measures operations related to fingers, such as finger movements.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In the technology of Patent Document 1, at least a part of the sensors included in the wearable sensor is attached to the palm side portion of the operator's hand. Due to such sensors attached to the palm side portion, there is a possibility that the work by the operator may be hindered.

Means for Solving the Problems

[0005] The present disclosure can be realized in the following forms.

[0006] (1) According to one aspect of the present disclosure, an estimation device is provided. The estimation device is provided away from the palm side portion of the operator's hand, and includes an acquisition unit that acquires a first detection result by a first sensor that detects a first physical quantity related to the bending of the operator's fingers, and a second detection result by a second sensor that is attached to the forearm of the operator and detects a second physical quantity related to the movement of the muscles of the forearm, and an estimation unit that estimates the finger force generated in the fingers using the first detection result and the second detection result. This configuration does not require sensors to be attached to the palm side of the worker's hand, thus suppressing interference with the worker's work due to sensors, while allowing for appropriate estimation of finger strength using the first and second detection results. (2) In the above embodiment, the acquisition unit may further acquire a third detection result from a third sensor that is provided away from the palm side of the worker's hand and detects a third physical quantity relating to the flexion of the worker's wrist, and the estimation unit may further estimate the finger strength using the third detection result. In this embodiment, the finger strength can be estimated more effectively by using the third detection result in addition to the first and second detection results. (3) In the above configuration, the first sensor and the third sensor may be attached to the back of the worker's hand. This configuration allows the first sensor and the third sensor to be concentrated on the worker's hand while preventing the sensors from hindering the worker's work. (4) In the above embodiment, the estimation unit may estimate the finger strength using a machine learning model that has been trained to output a prediction result of the finger strength based on the first detection result, the second detection result, and the third detection result. According to this embodiment, the finger strength can be estimated by integrating the first detection result, the second detection result, and the third detection result using a simple method. This disclosure can be implemented in forms other than the estimation device described above, such as an estimation system, an estimation method, a program for implementing the estimation method, a non-temporary recording medium on which the program is recorded, or a program product. The program product may be provided, for example, as a recording medium on which the program is recorded, or as a program product that can be distributed via a network. [Brief explanation of the drawing]

[0007] [Figure 1] This is an explanatory diagram showing the schematic configuration of the estimation system. [Figure 2] This is a conceptual diagram illustrating the process of estimating finger strength. [Figure 3]This is a flowchart showing the processing steps, including the estimation process. [Modes for carrying out the invention]

[0008] A. First Embodiment: Figure 1 is an explanatory diagram showing the schematic configuration of the estimation system 10 in the first embodiment. The estimation system 10 is used to estimate the hand strength of worker WK performing the work. Details of hand strength will be described later.

[0009] The estimation system 10 is used in the workplace where worker WK performs work. In this embodiment, the workplace is a factory FC for manufacturing vehicles. The work in this embodiment is a variety of tasks for manufacturing vehicles, including, for example, tasks related to vehicle assembly, tasks related to the installation of parts into vehicles, and tasks related to vehicle inspection. Such tasks include, for example, taking out, fitting, aligning, temporarily placing, tightening, removing, joining, attaching, temporarily attaching, and positioning parts. Such tasks may involve worker WK grasping parts with their hands or pinching parts with their fingers FN. Such tasks may also involve worker WK using various tools. These tools may be used, for example, while being grasped by worker WK's hands or pinched by their fingers FN.

[0010] The above-mentioned finger force represents the force generated in the fingers FN of worker WK in conjunction with the work performed by worker WK. More specifically, the finger force is generated when worker WK grips or holds tools or parts. In this embodiment, the estimation system 10 estimates the magnitude of the finger force by performing the estimation process described later. The finger force may be estimated for each of worker WK's hands, or for each of worker WK's fingers FN. In this embodiment, the finger force is estimated for each of worker WK's hands.

[0011] The estimation system 10 comprises a first sensor 50, a second sensor 60, and an estimation device 100. Furthermore, in this embodiment, the estimation system 10 includes a third sensor 70.

[0012] In this embodiment, the first sensor 50, the second sensor 60, and the third sensor 70 are each configured as wearable sensors. More specifically, the first sensor 50 and the third sensor 70 are mounted on a glove 80 that can be worn on the hand of worker WK, and are integrated into the glove 80. The second sensor 60 is mounted on a band 90 that can be attached to the forearm FA of worker WK, and is attached to the forearm FA of worker WK via the band 90. In this embodiment, the glove 80 is shaped to leave the fingertips exposed, but in other embodiments, it may be shaped to cover the fingertips. Also, the glove 80 may be worn over, for example, any other work glove worn on the hand of worker WK.

[0013] The first sensor 50 is located away from the palmar portion PM of the worker WK's hand. The palmar portion PM includes not only the palm itself but also the palmar portion of the fingers FN. In this embodiment, the first sensor 50 is located on a part of the worker WK's body that is different from the palmar portion PM. More specifically, the first sensor 50 is located on the back portion BK of the glove 80. The back portion BK is the part of the glove 80 that corresponds to the back portion BH of the worker WK's hand and is the part opposite the front portion FR of the glove 80. The back portion BH includes not only the back of the hand itself, i.e., the part opposite the palm, but also the back portion of the fingers FN. The front portion FR is the part of the glove 80 that corresponds to the palmar portion PM of the worker WK. With this configuration, when the glove 80 is worn on the worker WK's hand, the first sensor 50 is attached to the back portion BH of the worker WK.

[0014] The first sensor 50 detects a first physical quantity. The first physical quantity is a first physical quantity relating to the flexion of worker WK's finger FN. The detection result by the first sensor 50 is also called the first detection result. The first detection result is associated with information indicating the timing at which the first detection result was detected. The first sensor 50 transmits the detected first detection result to the estimation device 100.

[0015] In this embodiment, the first sensor 50 mechanically detects a first physical quantity. More specifically, in this embodiment, the first sensor 50 is configured as a sensor group including a plurality of bending sensors 51. In this embodiment, two bending sensors 51A and 51B are arranged for one finger FN. Each bending sensor 51 is arranged along the skeleton of each finger FN. Bending sensor 51A is located in the back of the hand portion BH between bending sensor 51B and the fingertip. Bending sensor 51B is located in the back of the hand portion BH between bending sensor 51A and the wrist WR. In this embodiment, the bending sensor 51 is configured as a resistance-changing type bending sensor 51, and is configured such that the electrical resistance of the bending sensor 51 changes according to the degree of bending of the bending sensor 51. With this configuration, the first sensor 50 mechanically detects the degree of bending of each finger FN as a first physical quantity. In other embodiments, the bending sensor 51 may be configured as, for example, a capacitive type bending sensor.

[0016] The second sensor 60 is attached to the forearm FA of worker WK. The second sensor 60 detects a second physical quantity. The second physical quantity is a physical quantity related to the movement of the muscles in the forearm FA. The detection result by the second sensor 60 is also called the second detection result. The second detection result is associated with information indicating the timing at which the second detection result was detected. The second sensor 60 transmits the detected second detection result to the estimation device 100.

[0017] In this embodiment, the second sensor 60 mechanically detects a second physical quantity. More specifically, in this embodiment, the second sensor 60 is configured as a surface pressure sensor for detecting the movement of the muscles of the forearm FA. The second sensor 60 as a surface pressure sensor may be configured as, for example, a resistance-type surface pressure sensor or a capacitive-type surface pressure sensor. The second sensor 60 is in the form of a sheet with enough flexibility to be deformable along the shape of the forearm FA, and is attached so as to be in close contact with at least a part of the forearm FA, thereby detecting the degree of muscle movement in each part of the forearm FA associated with work as a surface pressure distribution. More specifically, as the worker WK works, the muscles of the forearm FA contract or relax, and the degree of muscle prominence and muscle stiffness of the forearm FA change, which changes the degree to which the second sensor 60 is pressed by the muscles of the forearm FA. The second sensor 60 detects this change in the degree of pressure applied by the muscles of the forearm FA. With this configuration, the second sensor 60 mechanically detects the surface pressure distribution, which represents the degree of muscle movement in the forearm FA, as a second physical quantity. The method of mechanically detecting or analyzing muscle activity, such as muscle contraction and relaxation, and the resulting changes in muscle prominence and stiffness, is also called Force Myography (FMG). In other words, in this embodiment, the second sensor 60 is configured as a sensor capable of performing FMG.

[0018] As described above, the detected second physical quantity reflects not only finger strength but also flexion of the finger force (FN) and wrist strength (WR). In other words, even if the finger strength is the same, if the degree of flexion of the finger force (FN) and wrist strength (WR) differs, different second physical quantities can usually be detected.

[0019] The third sensor 70 is provided away from the palmar part PM. In the present embodiment, the third sensor 70 is provided at a position different from the palmar part PM among the body parts of the operator WK. More specifically, the third sensor 70 is arranged on the back part BK of the glove 80. With such a configuration, when the glove 80 is worn on the hand of the operator WK, the third sensor 70 is worn on the dorsal part BH. In the present embodiment, the third sensor 70 is arranged near the wrist WR of the operator WK among the dorsal parts BH. More specifically, the third sensor 70 is arranged between the first sensor 50 and the wrist WR.

[0020] The third sensor 70 detects a third physical quantity. The third physical quantity is a physical quantity related to the flexion of the wrist WR of the operator WK. The detection result by the third sensor 70 is also referred to as the third detection result. Information indicating the timing when the third detection result is detected is associated with the third detection result. The third sensor 70 transmits the detected third detection result to the estimation device 100. In other embodiments, for example, a transmission unit for aggregating the first detection result and the third detection result and transmitting them to the estimation device 100 may be mounted on the glove 80, and the first detection result and the third detection result may be transmitted to the estimation device 100 via the transmission unit.

[0021] In the present embodiment, the third sensor 70 mechanically detects the third physical quantity. More specifically, the third sensor 70 is configured as an inertial measurement unit (IMU) including a three-axis acceleration sensor, a three-axis gyro sensor, and a three-axis geomagnetic sensor. The third sensor 70 mechanically detects the acceleration and angular velocity generated in the wrist WR of the operator WK as the third physical quantity. It is possible to obtain the position and angle of the wrist WR of the operator WK by using the integration of the detected acceleration and angular velocity. Furthermore, it is possible to obtain the degree of flexion of the wrist WR based on the position and angle of the wrist WR. <{

[0022] In this embodiment, the first sensor 50 and the third sensor 70 are attached to the back of the hand portion BH, and the second sensor 60 is attached to the forearm portion FA, so that none of the sensors, such as the first sensor 50, the second sensor 60, or the third sensor 70, are attached to the palm portion PM.

[0023] The estimation device 100 is composed of a computer comprising a processor 101, a memory 102 including ROM and RAM, an input / output interface 103, and an internal bus 104. The processor 101, memory 102, and input / output interface 103 are connected via the internal bus 104 to enable bidirectional communication. A communication device 105 and a display device 106 are connected to the input / output interface 103. The communication device 105 can communicate directly or indirectly with the first sensor 50, the second sensor 60, and the third sensor 70 by wired or wireless communication. The display device 106 is composed of, for example, a liquid crystal display and displays various information such as information related to the estimation results by the estimation system 10. Various information such as the program PG1 and the prediction model 210 is stored in the memory 102. The processor 101 realizes various functions, including those of an acquisition unit 110, an estimation unit 120, and a processing unit 190, by executing the program PG1.

[0024] Figure 2 is a conceptual diagram illustrating the flow of finger force estimation in this embodiment. As shown in Figure 2, the acquisition unit 110 acquires the first detection result DR1 and the second detection result DR2. In this embodiment, the acquisition unit 110 also acquires the third detection result DR3.

[0025] The estimation unit 120 performs estimation processing. Estimation processing is the process of estimating the hand strength of worker WK using the first detection result DR1 and the second detection result DR2 acquired by the acquisition unit 110. In the estimation processing in this embodiment, the estimation unit 120 further estimates the hand strength of worker WK using the third detection result DR3 acquired by the acquisition unit 110. The estimation unit 120 records the hand strength estimated by the estimation processing as estimation result ER in the memory 102. The estimation unit 120 also outputs the estimation result ER. More specifically, the estimation unit 120 displays the estimation result ER on the display device 106 or causes the processing unit 190 to perform subsequent processing described later.

[0026] In this embodiment, the estimation unit 120 estimates finger strength using the prediction model 210. The prediction model 210 is a machine learning model trained to output a predicted finger strength result PR based on a first detection result DR1, a second detection result DR2, and a third detection result DR3. In this embodiment, the prediction model 210 is trained to output a predicted finger strength result PR as input, with the first detection result DR1, the second detection result DR2, and the third detection result DR3 included.

[0027] In this embodiment, the prediction model 210 is trained by supervised learning using a training dataset. The training dataset includes multiple training data and multiple labels. In the training dataset, each training data is associated with a label. The training data corresponds to explanatory variables, and the labels correspond to the target variable. In this embodiment, the training data used includes information representing the first detection result DR1, the second detection result DR2, and the third detection result DR3. The labels used are data representing the magnitude of finger strength. Such a training dataset is prepared, for example, by measuring the first, second, and third physical quantities while the worker WK is fitted with each sensor, and measuring the grip strength corresponding to finger strength using a general-purpose dynamometer. Various machine learning models such as random forests, support vector machines (SVMs), and neural networks can be used as the prediction model 210. Furthermore, in other embodiments, the learning method for the prediction model 210 is not limited to supervised learning. For example, the prediction model 210 may be trained by unsupervised learning or reinforcement learning.

[0028] The processing unit 190 uses the estimated result ER from the estimation system 10 to perform subsequent processing. The subsequent processing is a process for utilizing the estimated result ER, and includes, for example, an analysis process for analyzing the estimated result ER. In the analysis process, the processing unit 190 analyzes, in real time or retrospectively, the appropriateness of the worker's WK state and the appropriateness of the work style performed by the worker WK, for example, by comparing the finger strength as the estimated result ER with a predetermined standard finger strength according to the work. Such analysis processing may be used, for example, for quality assurance of products produced at the workplace or for safety evaluation of work at the workplace. The processing unit 190 may, for example, display the processing results of the subsequent processing on the display device 106. The content of the subsequent processing is not limited to the above.

[0029] Figure 3 is a flowchart showing the processing procedure including the estimation process in this embodiment. The processing procedure shown in Figure 3 is executed by the processor 101 of the estimation device 100, for example, at predetermined time intervals.

[0030] In step S100 of Figure 3, the acquisition unit 110 acquires each detection result from each sensor. More specifically, in step S100, the acquisition unit 110 acquires the first detection result DR1, the second detection result DR2, and the third detection result DR3. In step S105, the estimation unit 120 performs estimation processing. More specifically, in step S105 in this embodiment, the estimation unit 120 inputs each detection result acquired in step S100 into the prediction model 210 and estimates the finger strength by outputting the finger strength prediction result PR from the prediction model 210. Also in step S105, the estimation unit 120 records the estimated finger strength as the estimation result ER in the memory 102. In step S110, the estimation unit 120 outputs the estimation result ER.

[0031] According to the estimation device 100 in this embodiment described above, the hand strength of worker WK is estimated using the first detection result DR1 from the first sensor 50 and the second detection result DR2 from the second sensor 60. The first sensor 50 is provided away from the palmar portion PM and detects a first physical quantity related to the flexion of the fingers FN. The second sensor 60 is attached to the forearm portion FA and detects a second physical quantity related to the movement of the muscles of the forearm portion FA. In this way, it is not necessary to attach a sensor to the palmar portion PM, so it is possible to suppress interference with the gripping of tools and parts by the worker WK's hand, or with the gripping of tools and parts by the fingers FN, due to a sensor on the palmar portion PM. As a result, it is possible to suppress interference with the worker WK's work due to a sensor. Furthermore, unlike this embodiment, it is difficult to estimate finger strength using only the second detection result DR2, which reflects both finger strength and the degree of finger FN flexion, or to appropriately estimate finger strength using only the first detection result DR1, which simply reflects the degree of finger FN flexion. In contrast, in this embodiment, finger strength can be appropriately estimated by using both the first detection result DR1 and the second detection result DR2.

[0032] Furthermore, according to this embodiment, compared to, for example, a case where a pressure sensor or load sensor for directly detecting finger force is provided on the palm side PM, direct contact between the sensor and parts or tools due to work can be suppressed, and damage to the sensor can be suppressed. Also, compared to, for example, a case where a protective structure for the purpose of suppressing damage to such sensors is provided on the front part FR of the glove 80, an increase in the thickness of the front part FR can be suppressed, and work can be prevented from being hindered due to the thickness of the front part FR can be suppressed.

[0033] Furthermore, in this embodiment, the finger strength is estimated using the third detection result DR3 obtained by the third sensor 70. The third sensor 70 is located away from the palmar portion PM and detects a third physical quantity related to the flexion of the worker WK's wrist WR. By using the third detection result DR3 in addition to the first detection result DR1 and the second detection result DR2, the finger strength can be estimated more effectively while suppressing interference with the worker WK's work due to the sensor. More specifically, for example, the finger strength can be estimated with good accuracy even when the estimation process is performed in a situation where the degree of flexion of the worker WK's wrist WR may differ depending on the estimation timing at which the finger strength is estimated.

[0034] Furthermore, in this embodiment, the first sensor 50 and the third sensor 70 are mounted on the back of the hand portion BH. This allows the first sensor 50 and the third sensor 70 to be concentrated on the worker WK's hand while suppressing interference with the worker WK's work due to the sensors. Also, as in this embodiment, the first sensor 50 and the third sensor 70 can be compactly integrated into a hand-mounted device such as a glove 80.

[0035] Furthermore, in this embodiment, the first sensor 50 and the third sensor 70 mechanically detect the first and third physical quantities, respectively. This suppresses interference with the detection of each physical quantity due to disturbances such as foreign objects, compared to, for example, a case where the first sensor 50 and the third sensor 70 are configured to detect each physical quantity optically, and allows for more robust detection of each physical quantity. As a result, finger force can be estimated with greater robustness.

[0036] Furthermore, in this embodiment, the estimation unit 120 estimates finger strength using a pre-trained prediction model 210 that outputs a finger strength prediction result PR based on the first detection result DR1, the second detection result DR2, and the third detection result DR3. Therefore, finger strength can be estimated by comprehensively utilizing the first detection result DR1, the second detection result DR2, and the third detection result DR3 using a simple method.

[0037] B. Other embodiments: (B1) In the above embodiment, the estimation unit 120 uses the third detection result DR3 in the estimation process, but it is not necessary to use the third detection result DR3. That is, the estimation unit 120 only needs to estimate the finger strength using at least the first detection result DR1 and the second detection result DR2 in the estimation process. In this case, the estimation unit 120 may estimate the finger strength using, for example, a machine learning model that has been trained to predict the finger strength based on the first detection result DR1 and the second detection result DR2. Even in this configuration, the finger strength can be estimated accurately if the estimation process is performed in a situation where, for example, the degree of flexion of the worker WK's wrist WR does not change according to the estimation timing, or in a situation where the change in the degree of flexion according to the estimation timing is relatively small. Furthermore, in this configuration in which the third detection result DR3 is not used in the estimation process, the acquisition unit 110 does not need to acquire the third detection result DR3. Also, in this configuration, the estimation system 10 does not need to be equipped with a third sensor 70.

[0038] (B2) In the above embodiment, the first sensor 50 is configured as a sensor group including a plurality of bending sensors 51, and the third sensor 70 is configured as an IMU, but it is not limited to this. For example, the first sensor 50 may be configured as an IMU. Also, the third sensor 70 may be configured as, for example, one or more bending sensors. Also, in the above embodiment, the first sensor 50 and the third sensor 70 detect the first and third physical quantities mechanically, respectively, but it is not limited to this. For example, the first sensor 50 and the third sensor 70 may be configured as optical sensors that detect each physical quantity optically. Optical sensors include, for example, cameras and Lidar (Light Detection And Ranging) devices. Also, in this case, the functions of the first sensor 50 and the third sensor 70 may be realized by, for example, a single optical sensor.

[0039] (B3) In the above embodiment, the first sensor 50 and the third sensor 70 are attached to the back of the hand portion BH. However, the first sensor 50 and the third sensor 70 do not have to be attached to the back of the hand portion BH as long as they are provided away from the palm portion PM. For example, the first sensor 50 and the third sensor 70 may be attached to the side of the worker WK's hand. The side of the hand includes the outer surface of the hand and the side of the fingers FN. The first sensor 50 and the third sensor 70 may also be attached to a part of the worker WK's arm other than the hand. Furthermore, the first sensor 50 and the third sensor 70, configured as optical sensors, may be attached to a body part other than the worker WK's arm, or may be provided away from the worker WK.

[0040] (B4) In the above embodiment, a machine learning model is used in the estimation process, but a machine learning model is not required. For example, the finger strength may be estimated in the estimation process by using a pre-prepared rule-based system. Such a rule-based system may be configured to calculate an index value corresponding to the component derived from finger strength by subtracting a component derived from the flexion of the finger FN from an action value (e.g., a value representing muscle stiffness) that represents the degree of movement of the forearm FA muscles, and to output a predicted value of finger strength based on the calculated index value. In this case, when calculating the index value, a component derived from the flexion of the wrist WR may be further subtracted from the action value. In this case, the action value is calculated based on the second detection result DR2. The component derived from the flexion of the finger FN is calculated based on the first detection result DR1. The component derived from the flexion of the wrist WR is calculated based on the third detection result DR3.

[0041] (B5) In the above embodiment, the prediction model 210 is a machine learning model that has been trained to output a prediction result PR, taking the first detection result DR1, the second detection result DR2, and the third detection result DR3 as inputs. In contrast, the prediction model 210 only needs to be configured to output a prediction result PR based on the first detection result DR1, the second detection result DR2, and the third detection result DR3, and does not need to take the first detection result DR1, the second detection result DR2, and the third detection result DR3 as inputs. For example, the prediction model 210 may be configured to take a predicted value based on the first detection result DR1 as input instead of the first detection result DR1. Also, the prediction model 210 may be configured to take a predicted value based on the second detection result DR2 as input instead of the second detection result DR2. Also, the prediction model 210 may be configured to take a predicted value based on the third detection result DR3 as input instead of the third detection result DR3. The predicted values ​​input to the prediction model 210 may be output using, for example, one or more machine learning models different from the prediction model 210, or they may be output using a rule-based system. Such machine learning models are trained to output predicted values, for example, by taking one or two of the first detection result DR1, the second detection result DR2, and the third detection result DR3 as input. The estimation unit 120 may also be configured to change the prediction model 210 used, for example, depending on one or two of the first detection result DR1, the second detection result DR2, and the third detection result DR3. For example, the estimation unit 120 may switch between using two prediction models 210 that output predicted result PR, taking the second detection result DR2 and the third detection result DR3 as input, depending on whether the degree of flexion of the finger FN as the first detection result DR1 is above a predetermined level or below a predetermined level. Even in this configuration, the prediction model 210 can output a prediction result PR based on the first detection result DR1, the second detection result DR2, and the third detection result DR3.

[0042] (B6) In the above embodiment, a surface pressure sensor is used as the second sensor 60, but it is not limited to this. For example, various sensors that realize FMG may be used as the second sensor 60. For example, various piezoelectric sensors and various capacitive sensors may be used as such second sensors 60. Furthermore, the shape and material of such second sensors 60 may be arbitrary. For example, a functional rubber material may be used for the second sensor 60, or a functional fiber material that can realize smart textile (E-Textile) technology may be used. In addition, similar to the second sensor 60, a functional rubber material may be used for the first sensor 50 and the third sensor 70, or E-Textile technology may be used. Furthermore, the second sensor 60 is not limited to a sensor that realizes FMG, but for example, an EMG sensor that can detect the movement of muscles in the forearm FA using electromyography (EMG) may be used. The EMG sensor has electrodes for detecting electrical signals generated in the muscle in conjunction with muscle movement, and detects a second physical quantity using the electrodes.

[0043] This disclosure is not limited to the embodiments described above, and can be implemented in various configurations without departing from its spirit. For example, the technical features in the embodiments corresponding to the technical features in each form described in the summary of the invention can be replaced or combined as appropriate in order to solve some or all of the above-described problems, or to achieve some or all of the above-described effects. Furthermore, if a technical feature is not described as essential in this specification, it can be deleted as appropriate. [Explanation of Symbols]

[0044] 10...Estimation system, 50...First sensor, 51, 51A, 51B...Bend sensor, 60...Second sensor, 70...Third sensor, 80...Globe, 90...Band, 100...Estimation device, 101...Processor, 102...Memory, 103...Input / Output interface, 104...Internal bus, 105...Communication device, 106...Display device, 110...Acquisition unit, 120...Estimation unit, 190...Processing unit, 210...Prediction model

Claims

1. An acquisition unit that acquires a first detection result from a first sensor, which is located away from the palm side of the worker's hand and detects a first physical quantity related to the flexion of the worker's fingers, and a second detection result from a second sensor, which is attached to the worker's forearm and detects a second physical quantity related to the movement of the muscles of the forearm. An estimation device comprising: an estimation unit that estimates the finger force generated in the fingers using the first detection result and the second detection result.

2. An estimation device according to claim 1, The acquisition unit further acquires a third detection result from a third sensor, which is provided separately from the palmar portion and detects a third physical quantity related to the flexion of the worker's wrist. The estimation unit further estimates the finger strength using the third detection result, and is an estimation device.

3. The estimation device according to claim 2, The first sensor and the third sensor are an estimation device attached to the back of the worker's hand.

4. The estimation device according to claim 3, The estimation unit is an estimation device that estimates the finger strength using a machine learning model that has been trained to output a prediction result of the finger strength based on the first detection result, the second detection result, and the third detection result.

5. A first sensor is provided at a position away from the palm side of the worker's hand and detects a physical quantity related to the flexion of the worker's fingers, A second sensor is attached to the forearm of the worker and detects physical quantities related to the movement of the muscles in the forearm, An acquisition unit that acquires a first detection result from the first sensor and a second detection result from the second sensor, An estimation system comprising: an estimation unit that estimates the force generated in the fingers using the first detection result and the second detection result.

Images