Estimation device, estimation method, estimation program, and learning model generation device

Abstract

To estimate a state of deterioration of an object including a flexible material having conductivity by using electrical characteristics of the object without using a special detector.SOLUTION: An estimation device (1) includes: a detection unit that detects electrical characteristics between a plurality of detection points (75) predetermined in a flexible material (22) included in an object (2), the flexible material (22) having conductivity and whose electrical characteristics change according to deformation; and an estimation unit (5) that inputs time-series electrical characteristics (4) of an object to be estimated to a learning model (51) and estimates a deterioration state corresponding to the input time-series electrical characteristics, wherein the learning model has been learned to use, as data for learning, time-series electrical characteristics changing according to the deformation of the flexible material and deterioration state information indicating a deterioration state related to the deformation of the flexible material, and receive input of the time-series electrical characteristics and output the deterioration state information.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to an estimation device, an estimation method, an estimation program, and a learning model generation device. [Background technology]

[0002] Conventionally, shape changes occurring in an object have been detected, and the state of a person or object that has deformed the object has been estimated using the detection results. When it comes to detecting shape changes occurring in an object, it is difficult to detect the deformation without inhibiting the deformation of the object. Furthermore, since strain sensors used to detect rigid body deformation such as metal deformation are difficult to use on articles, a special detection device is required to detect the deformation of an object. For example, a technique is known in which the displacement and vibration of an object are measured using a camera, an image of the deformation is acquired, and the amount of deformation is extracted (see, for example, Patent Document 1). Furthermore, a technique related to a flexible tactile sensor that estimates the amount of deformation from the amount of light transmitted through it is also known (see, for example, Patent Document 2). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] International Publication No. 2017 / 029905 [Patent Document 2] Japanese Patent Application Laid-Open No. 2013-101096 Summary of the Invention [Problem to be solved by the invention]

[0004] However, when detecting changes in the shape of an object using a camera and image analysis techniques to detect deformation, such as displacement, the system incorporating the camera and image analysis becomes large and undesirable, leading to larger equipment. On the other hand, objects made of flexible materials that have the flexibility to deform and recover after deformation lose their flexibility, resilience, and restoring force over time. However, evaluation of flexibility relies on the subjective perception of the inspector based on the softness and hardness of the object when it is deformed, and is therefore insufficient to identify the softness of the object. Therefore, for example, when the softness of an object becomes harder than its original softness due to changes over time, it becomes unclear when the softness of the object has decreased. Therefore, there is room for improvement in identifying the state of deterioration of an object.

[0005] The present disclosure aims to provide an estimation device, an estimation method, an estimation program, and a learning model generation device that can estimate the deterioration state of an object by utilizing the electrical characteristics of the object made of a flexible material that is conductive, without using a special detection device. [Means for solving the problem]

[0006] In order to achieve the above object, the first aspect is a detection unit that detects electrical characteristics between a plurality of predetermined detection points on a flexible material of an object that has conductivity and whose electrical characteristics change in response to deformation; an estimation unit that inputs the time-series electrical characteristics detected by the detection unit of the estimation object into a learning model that has been trained to use, as learning data, time-series electrical characteristics that change in response to deformation of the flexible material and degradation state information that indicates a degradation state related to the deformation of the flexible material, and to output the degradation state information, and estimates degradation state information that indicates a degradation state corresponding to the input time-series electrical characteristics; The estimation device includes:

[0007] A second aspect is the estimation device of the first aspect, the electrical property is volume resistivity, the object includes a flexible member, The deterioration state is a state showing a degree of deterioration in which the degree of deterioration increases as at least one physical quantity of the number of deformations, deformation frequency, and elapsed time from the initial state of the object increases.

[0008] A third aspect is the estimation device of the second aspect, The deterioration state indicates a degree of deterioration of at least one of the electrical characteristics when the object is deformed from a pre-deformation shape and the electrical characteristics when the object is restored to the pre-deformation shape in a time-series electrical characteristic.

[0009] A fourth aspect is the estimation device of the third aspect, The degradation state is a state in which the degree of degradation increases as the power of the frequency of the analysis result becomes larger than the power of the frequency at a predetermined time, when the time series electrical characteristics are frequency analyzed.

[0010] A fifth aspect is the estimation device according to any one of the second to fourth aspects, The object includes a urethane material having a structure with at least one of a fibrous and a mesh-like skeleton, or a structure with a plurality of minute air bubbles dispersed therein, at least a portion of which is made electrically conductive.

[0011] A sixth aspect is the estimation device according to any one of the first to fifth aspects, The learning model includes a model generated by learning using a network based on reservoir computing using the flexible material as a reservoir.

[0012] The seventh aspect is The computer detecting electrical characteristics between a plurality of predetermined detection points on a flexible material of an object having conductive flexible material whose electrical characteristics change in response to deformation; Using time-series electrical characteristics that change in response to deformation of the flexible material and degradation state information that indicates a degradation state related to the deformation of the flexible material as learning data, the time-series electrical characteristics of the detected estimation object are input to a learning model that has been trained to input the time-series electrical characteristics and output the degradation state information, and degradation state information that indicates a degradation state corresponding to the input time-series electrical characteristics is estimated. It is an estimation method.

[0013] The eighth aspect is To the computer detecting electrical characteristics between a plurality of predetermined detection points on a flexible material of an object having conductive flexible material whose electrical characteristics change in response to deformation; Using time-series electrical characteristics that change in response to deformation of the flexible material and degradation state information that indicates a degradation state related to the deformation of the flexible material as learning data, the time-series electrical characteristics of the detected estimation object are input to a learning model that has been trained to input the time-series electrical characteristics and output the degradation state information, and degradation state information that indicates a degradation state corresponding to the input time-series electrical characteristics is estimated. This is an estimation program for executing the process.

[0014] The ninth aspect is an acquisition unit that acquires electrical characteristics from a detection unit that detects electrical characteristics between a plurality of predetermined detection points on a flexible material of an object having conductive flexible material whose electrical characteristics change in response to deformation, and acquires deterioration state information that indicates a deterioration state of the flexible material; a learning model generation unit that uses time-series electrical characteristics that change in response to deformation of the flexible material and degradation state information that indicates the degradation state related to the deformation of the flexible material as learning data, inputs the time-series electrical characteristics when the flexible material is deformed, and generates a learning model that outputs degradation state information that indicates the degradation state related to the deformation of the flexible material; A learning model generation device including: [Effects of the Invention]

[0015] According to the present disclosure, it is possible to estimate the deterioration state of an object by utilizing the electrical characteristics of the object having a conductive flexible material without using a special detection device. [Brief explanation of the drawings]

[0016] [Figure 1] FIG. 1 is a diagram illustrating a configuration of an estimation device according to an embodiment. [Figure 2] 10A and 10B are diagrams showing the arrangement of conductive urethane according to an embodiment. [Figure 3] FIG. 1 is a diagram illustrating a conceptual configuration of a learning model generation device according to an embodiment. [Figure 4] FIG. 2 is a diagram illustrating a functional configuration of a learning processing unit according to the embodiment. [Figure 5] FIG. 10 is a diagram illustrating another functional configuration of the learning processing unit according to the embodiment. [Figure 6] FIG. 2 is a diagram illustrating an electrical configuration of the estimation device according to the embodiment. [Figure 7] 10 is a flowchart illustrating a flow of an estimation process according to the embodiment. [Figure 8] 1 is a diagram illustrating an example of a configuration of a degradation state estimating device according to an embodiment; [Figure 9] FIG. 1 is a diagram illustrating a measurement device for measuring a physical quantity of an object according to an embodiment. [Figure 10] FIG. 10 is a conceptual diagram illustrating an example of a result of a durability test of an object according to the embodiment. [Figure 11] FIG. 4 is a diagram illustrating an example of electrical characteristics of an object according to the embodiment. [Figure 12] 1 is a conceptual diagram illustrating the electrical characteristics of an object according to an embodiment in terms of changes in electrical resistance value. [Figure 13] 3 is a conceptual diagram showing frequency characteristics (power spectrum) of electrical characteristics of an object according to the embodiment. FIG. [Figure 14] FIG. 10 is a diagram illustrating an example of a flow of a learning process according to the embodiment. [Figure 15] 10 is a flowchart showing an example of the flow of a process for estimating a deterioration state of an object according to the embodiment. [Figure 16] FIG. 10 is a diagram illustrating a modification of the degradation state estimating unit according to the embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0017] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. Components and processes that perform the same actions and functions are given the same reference numerals throughout the drawings, and duplicated descriptions may be omitted as appropriate. Furthermore, the present disclosure is not limited to the following embodiments, and can be implemented with appropriate modifications within the scope of the purpose of the present disclosure.

[0018] In the present disclosure, a person is a concept including at least one of a human body and an object that can provide a stimulus to a target object using a physical quantity. In the following description, the person will be collectively referred to as a concept including both a human and an object without distinguishing between the human body and the object. In other words, the person will be a collective term for each of the human body and the object, and a combination of the human body and the object.

[0019] First, with reference to Figures 1 to 7, a flexible material to which the technology of the present disclosure is applied and a state estimation process for estimating the state of the imparting side to the flexible material using the flexible material will be described.

[0020] <Flexible material> In this disclosure, the term "flexible material" refers to a material that is at least partially deformable, such as by bending. This term includes soft elastic materials such as rubber, structures with at least one of a fibrous and a mesh-like skeleton, and structures with multiple microscopic air bubbles dispersed therein. Examples of these structures include polymeric materials such as urethane. This disclosure also uses flexible materials that have been imparted with electrical conductivity. The term "flexible material imparted with electrical conductivity" refers to a material that has been imparted with an electrical conductive material, and includes materials in which an electrically conductive material has been imparted to a flexible material to impart electrical conductivity, and materials in which the flexible material is electrically conductive. A suitable flexible material that is imparted with electrical conductivity is a polymeric material such as urethane. In the following description, a member formed by blending and infiltrating (also referred to as impregnation) a conductive material into all or part of a urethane material will be referred to as "conductive urethane" as an example of a flexible material imparted with electrical conductivity. Conductive urethane can be formed by either blending or infiltrating (impregnating) a conductive material, or by combining blending or infiltrating (impregnating) a conductive material. For example, if the conductive urethane formed by infiltration (impregnation) has a higher conductivity than the conductive urethane 22 formed by blending, it is preferable to form the conductive urethane by infiltration (impregnation).

[0021] Conductive urethane has the function of changing its electrical properties in response to a given physical quantity. One example of the physical quantity that causes the electrical property to change is a pressure stimulus value that indicates a pressure stimulus (hereinafter referred to as a pressure stimulus) that deforms a structure, such as bending. Note that pressure stimulus includes pressure applied to a specific location and pressure distribution over a specific range. Other examples of the physical quantity include a moisture content value that indicates a stimulus that changes (alters) the properties of a material by adding moisture, etc. (hereinafter referred to as a material stimulus). Conductive urethane's electrical properties change in response to a given physical quantity. An example of a physical quantity that indicates this electrical property is electrical resistance. Other examples include voltage or current.

[0022] By imparting conductivity to a flexible material with a predetermined volume, conductive urethane exhibits electrical characteristics (i.e., changes in electrical resistance) according to a given physical quantity, and this electrical resistance can be considered the volume resistance of the conductive urethane. Conductive urethane has complexly interconnected electrical pathways, which, for example, expand and contract in response to deformation. Electrical pathways may also be temporarily disconnected or reconnected differently from their previous state. Therefore, conductive urethane exhibits behaviors with different electrical characteristics depending on deformation and alteration according to the magnitude and distribution of stimuli (pressure stimuli and material stimuli) due to a given physical quantity between positions separated by a predetermined distance (e.g., the positions of detection points where electrodes are placed). Therefore, the electrical characteristics change according to the magnitude and distribution of stimuli due to a physical quantity applied to the conductive urethane.

[0023] Furthermore, by using conductive urethane, there is no need to provide detection points such as electrodes at the target locations for deformation and deterioration. Detection points such as electrodes can be provided at least two arbitrary locations on either side of the location where the physical stimulus is applied to the conductive urethane (see, for example, Figure 1).

[0024] Furthermore, to improve the detection accuracy of the electrical properties of the conductive urethane, more than two detection points may be used. Furthermore, the conductive urethane of the present disclosure may be formed as a conductive urethane group consisting of an array of multiple conductive urethane pieces, with the conductive urethane 22 shown in FIG. 1 being one conductive urethane piece. In this case, the electrical properties may be detected for each of the multiple conductive urethane pieces, or the electrical properties of the multiple conductive urethane pieces may be detected in combination. When detecting the electrical properties for each of the multiple conductive urethane pieces, electrical properties such as electrical resistance values ​​can be detected for each arrangement location (e.g., detection sets #1 to #n). As another example, the detection range on the conductive urethane 22 may be divided, and a detection point may be provided for each divided detection range, and the electrical properties may be detected for each detection range.

[0025] <Estimation device> Next, an example of an estimation device that uses conductive urethane to estimate the state of the application side of the conductive urethane will be described.

[0026] 1 shows an example of the configuration of an estimation device 1 capable of executing an estimation process for estimating the state of the application side. The estimation device 1 includes an estimation unit 5 and is connected to an object 2 so that the electrical characteristics of the conductive urethane 22 are input. The estimation device 1 estimates the state of the application side for the conductive urethane 22 included in the object 2. The estimation device 1 can be realized by a computer having a CPU as an execution device that executes the processes described below.

[0027] The deformation and alteration of the conductive urethane described above occurs due to physical quantities applied to the conductive urethane over time. These physical quantities applied over time depend on the state of the side to which they are applied. Therefore, the electrical properties of the conductive urethane, which change over time, correspond to the state of the side to which the physical quantities applied to the conductive urethane are applied. For example, when a pressure stimulus that deforms the conductive urethane or a material stimulus that alters the conductive urethane is applied, the electrical properties of the conductive urethane, which change over time, correspond to the state of the side to which the pressure stimulus is applied, which indicates the position, distribution, and magnitude of the pressure stimulus. Therefore, it is possible to estimate the state of the side to which the pressure stimulus is applied from the electrical properties of the conductive urethane, which change over time.

[0028] The estimation device 1 estimates and outputs the unknown state of the imparting side using a trained learning model 51 through an estimation process described below. This makes it possible to identify the state of the imparting side relative to the object 2 without using special or large equipment or directly measuring the deformation and deterioration of the conductive urethane 22 contained in the object. The learning model 51 is trained using as input the state of the imparting side relative to the object 2 and the electrical characteristics of the object 2 (i.e., electrical characteristics such as the electrical resistance value of the conductive urethane 22 placed on the object 2). The learning of the learning model 51 will be described later.

[0029] The conductive urethane 22 can be placed on a flexible member 21 to form the object 2 (FIG. 2). The object 2, which is made up of the member 21 on which the conductive urethane 22 is placed, includes an electrical characteristic detection unit 76. The conductive urethane 22 only needs to be placed on at least a part of the member 21, and may be placed inside or outside. The conductive urethane only needs to be placed so that the state of the conductive urethane application side can be estimated; for example, it can be placed so that it can come into contact with a person directly or indirectly, or both.

[0030] FIG. 2 shows an example of the arrangement of conductive urethane 22 in the object 2. As shown in the AA cross section of the object 2 as object cross section 2-1, the conductive urethane 22 may be formed to fill the entire interior of the member 21. Also, as shown in object cross section 2-2, the conductive urethane 22 may be formed on one side (front side) of the interior of the member 21, or as shown in object cross section 2-3, the conductive urethane 22 may be formed on the other side (back side) of the interior of the member 21. Furthermore, as shown in object cross section 2-4, the conductive urethane 22 may be formed in part of the interior of the member 21. Also, as shown in object cross section 2-5, the conductive urethane 22 may be separately arranged on the outside of the front side of the member 21, or as shown in object cross section 2-6, the conductive urethane 22 may be arranged on the outside of the other side (back side). When the conductive urethane 22 is arranged outside the member 21, the conductive urethane 22 and the member 21 may simply be laminated together, or the conductive urethane 22 and the member 21 may be integrated by adhesive or the like. Even when the conductive urethane 22 is disposed outside the member 21, the flexibility of the member 21 is not impaired because the conductive urethane 22 is a urethane member having electrical conductivity.

[0031] As shown in FIG. 1, the electrical characteristics (i.e., volume resistance, which is the electrical resistance value) of the conductive urethane 22 are detected by signals from at least two detection points 75 arranged at a distance from each other. The example in FIG. 1 shows detection set #1, which detects the electrical characteristics (time-series electrical resistance values) by signals from two detection points 75 arranged diagonally on the conductive urethane 22. The number and arrangement of the detection points 75 are not limited to those shown in FIG. 1; they may be three or more, or any other positions, as long as they are capable of detecting the electrical characteristics of the conductive urethane 22. The electrical characteristics of the conductive urethane 22 can be measured by connecting an electrical characteristic detection unit 76, which detects the electrical characteristics (e.g., volume resistance, which is the electrical resistance value), to the detection points 75 and using the output thereof.

[0032] In this embodiment, because the conductive urethane 22 is used as the sensor, for example, when a person is present, the sense of discomfort felt by the person is significantly reduced compared to conventional sensors. Therefore, the state of the person receiving the signal is not disturbed during measurement, and measurement and estimation of the state of the receiving side can be performed simultaneously. This is an advantage over conventional sensors that perform measurement and estimation of the state of the receiving side separately, and is particularly beneficial in estimation based on long-term measurement evaluation that tracks time-series changes.

[0033] The estimation unit 5 is a functional unit connected to the object 2 and uses a learning model 51 to estimate the state of the application side based on the electrical characteristics that change in response to at least one of deformation and deterioration of the conductive urethane 22. Specifically, the estimation unit 5 receives time-series input data 4 representing the magnitude of electrical resistance (e.g., electrical resistance value) in the conductive urethane 22. The input data 4 corresponds to state data 3 that represents the state of the application side relative to the object 2, for example, the state related to the behavior of a person who touches the object 2, such as the posture and movement of the person. For example, when a person touches the object 2, the person makes contact in a predetermined state, such as posture, and in response to this state, a stimulus (at least one of a pressure stimulus and a material stimulus) is applied as a physical quantity to the conductive urethane 22 that constitutes the object 2, causing the electrical characteristics of the conductive urethane 22 to change. Therefore, the time-series changing electrical characteristics of the conductive urethane 22 represented by the input data 4 correspond to the state of the application side relative to the object 2, i.e., the conductive urethane 22. Furthermore, the estimation unit 5 outputs, as an estimation result using the trained learning model 51, output data 6 that represents the state of the imparting side corresponding to the electrical characteristics of the conductive urethane 22 that change over time.

[0034] The learning model 51 is a model that has undergone training to derive output data 6 representing the state of the applying side from the electrical resistance (input data 4) of the conductive urethane 22, which changes in response to stimuli (pressure stimuli and material stimuli) given as physical quantities. The learning model 51 is, for example, a model that defines a trained neural network, and is expressed as a collection of information on the weights (strengths) of the connections between the nodes (neurons) that make up the neural network.

[0035] <Learning process> Next, the learning process for generating the learning model 51 will be described. 3 shows the conceptual configuration of a learning model generation device that generates a learning model 51. The learning model generation device includes a learning processing unit 52. The learning model generation device can be configured to include a computer with a CPU (not shown), and is executed as the learning processing unit 52 by a learning data collection process and a learning model generation process executed by the CPU to generate the learning model 51.

[0036] <Learning data collection process> In the learning data collection process, the learning processing unit 52 collects, as learning data, a large amount of input data 4 obtained by measuring the electrical characteristics (e.g., electrical resistance value) of the conductive urethane 22 in time series, using status data 3 indicating the status of the imparting side as labels. Therefore, the learning data includes a large amount of sets of input data 4 indicating the electrical characteristics and status data 3 indicating the status of the imparting side corresponding to the input data 4.

[0037] Specifically, in the learning data collection process, electrical characteristics (e.g., electrical resistance values) that change in response to stimuli (pressure stimuli and material stimuli) corresponding to the state of the object 2 (i.e., the state of the imparting side relative to the conductive urethane 22) when the state of the object 2 is formed are acquired in a time series. Next, state data 3 is assigned as a label to the acquired time series electrical characteristics (input data 4), and the process is repeated until a predetermined number of sets of state data 3 and input data 4 are obtained, or until a predetermined time has elapsed. Sets of these state data 3 indicating the state of the imparting side and the time series electrical characteristics of the conductive urethane 22 (input data 4) acquired for each state of the imparting side become learning data. Note that the state data 3 in the learning data is stored in a memory (not shown) so that it can be treated as output data 6 indicating the state of the imparting side for which the estimated result is correct in the learning process described below.

[0038] The learning data may be associated with time-series information by adding information indicating the measurement time to each of the electrical resistance values ​​(input data 4) of the conductive urethane 22. In this case, for a period determined as the state of the application side, information indicating the measurement time may be added to a set of time-series electrical resistance values ​​of the conductive urethane 22 to associate them with time-series information.

[0039] An example of the learning data described above is shown in the following table. Table 1 is an example of a data set that associates time-series electrical resistance value data (r) with state data (R) indicating the state of the application side as learning data regarding the state of the application side of the conductive urethane 22.

[0040] [Table 1]

[0041] The electrical characteristics (time characteristics based on time-series electrical resistance value data) detected by the conductive urethane 22 can be regarded as a characteristic pattern related to the state of the imparting side relative to the conductive urethane 22. In other words, different stimuli are applied to the conductive urethane 22 in a time series depending on the state of the imparting side relative to the conductive urethane 22. Therefore, it is considered that the time-series electrical characteristics within a predetermined period of time appear as electrical characteristics characteristic of the state of the imparting side. Therefore, a pattern (for example, the distribution shape of the time-series electrical resistance value in the electrical characteristics) shown in the electrical characteristics (time characteristics based on time-series electrical resistance value data) detected by the conductive urethane 22 corresponds to the state of the imparting side and functions effectively in the learning process described below.

[0042] <Learning model generation process> Next, the learning model generation process will be described. The learning model generation device shown in Fig. 3 generates a learning model 51 by the learning model generation process in the learning processing unit 52 using the above-mentioned learning data.

[0043] 4 is a diagram showing the functional configuration of the learning processing unit 52, i.e., the functional configuration of a CPU (not shown) related to the learning model generation process executed by the learning processing unit 52. The CPU (not shown) of the learning processing unit 52 operates as the functional units of the generator 54 and the calculator 56. The generator 54 has a function of generating an output in consideration of the context of the electrical resistance values ​​acquired in time series as input.

[0044] The learning processing unit 52 stores a large number of sets of learning data in a memory (not shown), which are the above-mentioned input data 4 (e.g., electrical resistance value) and output data 6, which is state data 3 indicating the state of the applying side that applied a stimulus to the conductive urethane 22.

[0045] The generator 54 includes an input layer 540, an intermediate layer 542, and an output layer 544, and constitutes a known neural network (NN). Since the neural network itself is a known technology, detailed description will be omitted, but the intermediate layer 542 includes a large number of node groups (neuron groups) having inter-node connections and feedback connections. Data from the input layer 540 is input to the intermediate layer 542, and data resulting from the calculations of the intermediate layer 542 is output to the output layer 544.

[0046] The generator 54 is a neural network that generates generated output data 6A as data representing the state of the application side or data approximating the state of the application side from input data 4 (e.g., electrical resistance values). The generated output data 6A is data that estimates the state of the application side when a stimulus is applied to the conductive urethane 22 from the input data 4. The generator 54 generates generated output data that indicates a state approximating the state of the application side from the input data 4 that is input in a time series. By learning using a large amount of input data 4, the generator 54 is able to generate generated output data 6A that is approximating the state of the application side, such as a person who has applied a stimulus to the object 2, i.e., the conductive urethane 22. In another aspect, by capturing the electrical characteristics of the input data 4 that is input in a time series as patterns and learning the patterns, it is possible to generate generated output data 6A that is approximating the state of the application side, such as a person who has applied a stimulus to the object 2, i.e., the conductive urethane 22.

[0047] The calculator 56 is a calculator that compares the generated output data 6A with the output data 6 of the learning data and calculates the error of the comparison result. The learning processing unit 52 inputs the generated output data 6A and the output data 6 of the learning data to the calculator 56. In response to this, the calculator 56 calculates the error between the generated output data 6A and the output data 6 of the learning data and outputs a signal indicating the calculation result.

[0048] The learning processing unit 52 performs learning of the generator 54 by tuning the weight parameters of the connections between nodes based on the error calculated by the calculator 56. Specifically, the weight parameters of the connections between nodes in the input layer 540 and the intermediate layer 542 in the generator 54, the weight parameters of the connections between nodes in the intermediate layer 542, and the weight parameters of the connections between nodes in the intermediate layer 542 and the output layer 544 are fed back to the generator 54 using a method such as gradient descent or backpropagation. In other words, with the output data 6 of the learning data as the target, the connections between all nodes are optimized so as to minimize the error between the generated output data 6A and the output data 6 of the learning data.

[0049] The generator 54 may use a recurrent neural network that has a function of generating an output taking into account the context of the time-series input, or may use other methods.

[0050] The learning processing unit 52 generates a learning model 51 using the above-mentioned learning data through a learning model generation process. The learning model 51 is expressed as a collection of information on weight parameters (weights or strengths) of connections between nodes as a result of learning, and is stored in a memory (not shown).

[0051] Specifically, the learning processing unit 52 executes the learning model generation process according to the following procedure. In the first learning process, input data 4 (electrical characteristics) is acquired, which is learning data obtained by measuring in time series and has labels containing information indicating the state of the assigning side. In the second learning process, the learning model 51 is generated using the learning data obtained by measuring in time series. That is, a set of information on weight parameters (weights or strengths) of connections between nodes in the learning results learned using a large amount of learning data as described above is obtained. Then, in the third learning process, data expressed as a set of information on weight parameters (weights or strengths) of connections between nodes in the learning results is stored as the learning model 51.

[0052] The estimation device 1 uses the trained generator 54 (i.e., data expressed as a set of information on weight parameters of bonds between nodes as a result of training) as the training model 51. If a sufficiently trained training model 51 is used, it is not impossible to identify the state of the imparting side from the time-series electrical characteristics of the object 2, i.e., the conductive urethane 22 (e.g., characteristics of electrical resistance value that change over time).

[0053] <prc> As described above, the conductive urethane 22 has complexly interconnected electrical pathways, exhibiting behavior in response to changes (deformations) such as expansion / contraction, temporary disconnections, and new connections of the electrical pathways, as well as changes (alterations) in the properties of the material. As a result, the conductive urethane 22 exhibits behavior with different electrical characteristics in response to a given stimulus (e.g., a pressure stimulus). This allows the conductive urethane 22 to be treated as a reservoir that stores data related to the deformation of the conductive urethane 22. In other words, the estimation device 1 can apply the conductive urethane 22 to a network model called physical reservoir computing (PRC) (hereinafter referred to as PRCN). PRC and PRCN are well-known technologies, and detailed description thereof will be omitted. However, PRC and PRCN are suitable for estimating information related to the deformation and alteration of the conductive urethane 22.

[0054] FIG. 5 shows an example of the functional configuration of a learning processing unit 52 employing PRCN. The learning processing unit 52 employing PRCN includes an input reservoir layer 541 and an estimation layer 545. The input reservoir layer 541 corresponds to the conductive urethane 22 included in the object 2. That is, the learning processing unit 52 employing PRCN treats the object 2 including the conductive urethane 22 as a reservoir for storing data related to the deformation and deterioration of the object 2 including the conductive urethane 22 during learning. The conductive urethane 22 exhibits electrical characteristics (electrical resistance values) corresponding to various stimuli, and functions as both an input layer for inputting the electrical resistance values ​​and a reservoir layer for storing data related to the deformation and deterioration of the conductive urethane 22. The conductive urethane 22 outputs different electrical characteristics (input data 4) depending on the stimulus applied depending on the state of the application side, such as a person. Therefore, the estimation layer 545 can estimate the unknown state of the application side from the given electrical resistance value of the conductive urethane 22. Therefore, the learning process in the learning processing unit 52 employing PRCN involves training the estimation layer 545.

[0055] <Configuration of the estimation device> Next, an example of a specific configuration of the above-mentioned estimation device 1 will be further described. Fig. 6 shows an example of the electrical configuration of the estimation device 1. The estimation device 1 shown in Fig. 6 is configured to include a computer as an execution device that executes processes to realize the various functions described above. The above-mentioned estimation device 1 can be realized by causing the computer to execute programs representing the various functions described above.

[0056] The computer functioning as the estimation device 1 includes a computer main unit 100. The computer main unit 100 includes a CPU 102, a RAM 104 such as a volatile memory, a ROM 106, an auxiliary storage device 108 such as a hard disk drive (HDD), and an input / output interface (I / O) 110. The CPU 102, RAM 104, ROM 106, auxiliary storage device 108, and input / output I / O 110 are connected via a bus 112 to enable mutual exchange of data and commands. The input / output I / O 110 is also connected to a communication unit 114 for communicating with external devices, an operation display unit 116 such as a display or keyboard, and a detection unit 118. The detection unit 118 functions to acquire input data 4 (electrical characteristics such as time-series electrical resistance values) from an object 2 including a conductive urethane 22. That is, the detection unit 118 can acquire the input data 4 from an electrical characteristic detection unit 76 that includes the object 2 on which the conductive urethane 22 is disposed and is connected to a detection point 75 on the conductive urethane 22. The detection unit 118 may be connected via the communication unit 114 .

[0057] The auxiliary storage device 108 stores a control program 108P for causing the computer main body 100 to function as the estimation device 1, which is an example of an estimation device of the present disclosure. The CPU 102 reads the control program 108P from the auxiliary storage device 108, loads it into the RAM 104, and executes processing. As a result, the computer main body 100, which has executed the control program 108P, operates as the estimation device 1.

[0058] The auxiliary storage device 108 stores a learning model 108M including the learning model 51, and data 108D including various data. The control program 108P may be provided by a recording medium such as a CD-ROM.

[0059] <Estimation process> Next, the estimation process in the estimation device 1 realized by a computer will be further described. Fig. 7 shows an example of the flow of the estimation process by the control program 108P executed by the computer main body 100. The estimation process shown in Fig. 7 is executed by the CPU 102 when the computer main body 100 is powered on. The CPU 102 reads the control program 108P from the auxiliary storage device 108, loads it into the RAM 104, and executes the process.

[0060] First, CPU 102 reads out learning model 51 from learning model 108M in auxiliary storage device 108 and expands it in RAM 104 to acquire learning model 51 (step S200). Specifically, by expanding a network model (see FIGS. 4 and 5) in which nodes are connected by weight parameters expressed as learning model 51 into RAM 104, learning model 51 in which connections between nodes are realized by weight parameters is constructed.

[0061] Next, CPU 102 acquires unknown input data 4 (electrical characteristics) in time series via detection unit 118, from which the state of the application side due to a stimulus applied to conductive urethane 22 is to be estimated (step S202). Next, CPU 102 estimates output data 6 (the unknown state of the application side) corresponding to the acquired input data 4 using learning model 51 (step S204). CPU 102 then outputs the estimated output data 6 (the state of the application side) via communication unit 114 (step S206), and ends this processing routine.

[0062] In this way, the estimation device 1 can estimate the unknown state of the application side from the electrical resistance value of the conductive urethane 22. Specifically, the estimation device 1 can estimate the state of the application side of a person or the like from input data 4 (electrical characteristics) that changes in response to a stimulus given to the conductive urethane 22 depending on the state of the application side. In other words, it is possible to estimate the state of the application side of a person or the like without using a special or large device or directly measuring the deformation of the flexible member.

[0063] Incidentally, flexible materials, at least a portion of which is deformable, such as by bending, lose softness and hardness (flexibility, resilience, and restoring force) depending on the number of deformations and changes over time. The number of deformations may be the deformation frequency, or a combination of the number of deformations and the deformation frequency. The deformation frequency may be equivalent to the number of deformations or the number of deformations per unit time. Hereinafter, the number of deformations and the number of deformations per unit time will be collectively referred to as the deformation frequency. The time-series electrical characteristics of the conductive urethane 22 described above exhibit characteristics related to the state of the applied side. Therefore, the pattern (e.g., the distribution shape of the time-series electrical resistance values ​​in the electrical characteristics) shown in the electrical characteristics (time characteristics based on the time-series electrical resistance value data) detected by the conductive urethane 22 exhibits characteristics indicating the state of deterioration of flexibility caused by the deformation frequency and changes over time. Therefore, in this embodiment, the degradation of a deformable flexible material is estimated using an estimation device 1 including the conductive urethane 22 described above.

[0064] For simplicity's sake, the following description will focus on an example in which the object 2 includes a non-conductive, deformable, flexible material as the member 21, and a conductive urethane 22 is disposed on the back side of the member 21 (see object cross section 2-5 in FIG. 2). The following description will discuss a case in which softness (flexibility) gradually hardens. However, it goes without saying that the technology of the present disclosure can also be applied to a case in which softness gradually becomes softer. An example of a case in which softness gradually becomes hard is a phenomenon in which air bubbles in a deformable flexible material, such as urethane foam, collapse and do not return to their original shape, causing the material to harden. On the other hand, the technology can also be applied to a case in which the hardness of a deformable flexible material, such as urethane, decreases (a state in which the skeleton of the flexible material gradually softens and the resilience decreases).

[0065] In this embodiment, "deterioration" and "deteriorated state" are concepts that include a state in which the softness of a flexible material has changed over a predetermined period of time, for example, from the time when the flexible material was first formed, with respect to the deformation and recovery from the deformation when a pressure stimulus is applied. Specifically, for example, in terms of the structure and shape of the flexible material, this includes a state in which the restoring force that restores the structure and shape after deformation to the structure and shape before deformation changes (increases or decreases) from the time when the flexible material was first manufactured, and a state in which the pressure that can be applied during deformation changes (increases or decreases) from the pressure at the time when the flexible material was first manufactured.

[0066] <Deterioration state estimation device> Next, an example of a degradation state estimation device that estimates the degradation state of the object 2 that includes the conductive urethane 22 will be described.

[0067] 8 shows an example of the configuration of an estimation device 1A as a degradation state estimation device according to this embodiment. The estimation device 1A shown in FIG. 8 includes a degradation state estimation unit 10 instead of the estimation unit 5 shown in FIG.

[0068] The estimation process in the deterioration state estimation device 1A estimates and outputs the deterioration state of an unknown item (estimation target) using a trained learning model 51. The learning model 51 is trained using training data including the electrical characteristics of the conductive urethane 22 and the deterioration state of the target 2 that includes the conductive urethane 22.

[0069] Next, a learning process for generating the learning model 51 according to this embodiment will be described. Learning is performed using the electrical characteristics (input data 4) of the conductive urethane 22, labeled with deterioration state information (state data 3) indicating the deterioration state of the object 2, as learning data.

[0070] First, the learning data used in the learning process will be described. 9 shows an example of a measuring device 8 that measures a physical quantity of the object 2. The measuring device 8 also functions as a durability test device for the object 2 by repeatedly applying a pressure stimulus (details will be described later).

[0071] In the measuring device 8, a pressure applying unit 83 for applying a pressure stimulus to the object 2 is attached to a fixed unit 82 fixed to a base 81. The pressure applying unit 83 includes a pressure applying main body 83A, an arm 83B that can extend and retract from the pressure applying main body 83A, and a tip end 83C attached to the tip of the arm 83B. The pressure applying main body 83A is fixed to the fixed unit 82, and the arm 83B extends and retracts in response to an input signal, causing the tip end 83C to move in a predetermined direction (the direction of arrow Z and the opposite direction). This allows the pressing member 84 to come into contact with the object 2 placed on the base 81, press the object 2 after contact, or move away from the object 2.

[0072] The object 2 has a conductive urethane 22 placed on a base 81, and a flexible member 21 such as urethane placed on the conductive urethane 22. A pressing member 84 of a predetermined shape is attached to a tip 83C of the pressure-applying unit 83. The object 2 is positioned so that the pressing member 84 attached to the tip 83C of the pressure-applying unit 83 can at least come into contact with the pressing member 84. In this embodiment, as an example of a pressing member 84 of a predetermined shape, a pressing member 84 having a curved tip (for example, a part of a sphere) is used. The pressing member 84 is a member that applies a pressure stimulus to the object 2 with a predetermined pressure. The cross-sectional shape of the pressing member 84 may be any of a rectangle, trapezoid, circle, ellipse, and polygon, or may be any other shape.

[0073] The pressure applying portion 83 operates so that the tip portion 83C presses the pressing member 84 against the object 2 as a result of the arm 83B extending.

[0074] The pressure applying main body 83A is equipped with a force sensor 85 that has the function of detecting force in, for example, six axial directions. The force sensor 85 has the function of detecting the pressing state of the pressing member 84 against the object 2 from the detected force, and the function of detecting the pressure applied to the object 2. This force sensor 85 can detect the physical amount of the pressing state of the pressing member 84 against the object 2 in time series, and can detect the pressure applied to the object 2 in time series. Note that the force sensor 85 can be omitted when testing only the deterioration state.

[0075] The measuring device 8 includes a pressure applying unit 83 and a controller 80 connected to a force sensor 85. The controller 80 includes a CPU (not shown), and controls the pressure applying unit 83 using the CPU (not shown), applies a pressure stimulus to the object 2, and acquires and stores the time-series electrical characteristics resulting from the pressure stimulus applied to the object 2.

[0076] In this embodiment, the controller 80 controls the pressure application unit 83 so that the arm 83B performs a reciprocating motion to extend and retract, thereby applying and releasing a pressure stimulus to the object 2. The controller 80 also acquires the electrical characteristics of the conductive urethane 22 in synchronization with the application and release of the pressure stimulus to the object 2. Therefore, the measurement device 8 can acquire, as one piece of learning data, the electrical characteristics related to the deformation of the object 2 that has been intermittently (e.g., periodically) deformed, in time series.

[0077] Next, the deterioration state of the object 2 will be described. The softness of the object 2 deteriorates due to at least one of the frequency of deformation and changes over time. For example, softness may deteriorate as the frequency of deformation increases, or as the elapsed time, such as the time the object 2 is left standing, increases. Furthermore, softness may also deteriorate as both the frequency of deformation and the elapsed time increase. The deterioration state of the object, which indicates deterioration of the object, is assumed to include, in terms of geometric aspects such as the structure and shape of the object 2, the time from when a pressure stimulus is applied to deform the object until the object returns to its original structure, etc., and the degree of structural difference between the original structure, etc. and the restored structure, etc. Furthermore, in terms of mechanical aspects of the object 2, the degree of difference between the repulsive force of the object 2 when a pressure stimulus is applied to deform the object, and the restoring force of the object 2 when the pressure stimulus is released and the object 2 returns to its original state, etc., is assumed. However, for a flexible object 2, both geometric and mechanical aspects are closely related, and therefore, in this embodiment, the two aspects are not distinguished and are treated as the deterioration state of the object 2.

[0078] Regarding the frequency of deformation, multiple samples (object 2) of the same size were prepared, and each sample was deformed a different number of times within a predetermined time under the same environment, and the deformation state was measured.It was confirmed that the greater the number of deformations, i.e., the greater the frequency of deformation within the predetermined time, the greater the degree of deterioration.In addition, regarding the elapsed time, a sample (object 2) was prepared, and the deformation state was measured at predetermined time intervals under the same environment.It was confirmed that the longer the time passed, i.e., the longer the elapsed time, the greater the degree of deterioration.

[0079] Fig. 10 is a conceptual diagram showing an example of the results of a durability test of the object 2 using the above-described measuring device 8. Fig. 10 shows, as an example, the concept of the electrical characteristics (change characteristics of the electrical resistance value) that change over time when the application and release of a pressure stimulus to the object 2 are repeatedly and continuously performed.

[0080] As shown in FIG. 10, the electrical characteristics of the object 2 fluctuate between an electrical resistance value Ed when a pressure stimulus is applied and an electrical resistance value Eu when the pressure stimulus is released and restored to its original state at the beginning of measurement. The electrical characteristics of the object 2 tend to show a gradual decrease in electrical resistance value from the beginning of measurement as the number of repetitions increases (measurement time t becomes longer). Therefore, in this embodiment, the tendency for the electrical resistance value to decrease is quantified and used as the degradation degree. That is, the degradation degree indicates the degree of the tendency for the electrical resistance value to decrease. The degradation degree can be determined according to the features included in the electrical characteristics.

[0081] The first feature is that the deterioration state of the object 2 is expressed by a gradually decreasing resistance value in the electrical characteristics of the object 2 ( FIG. 10 ). For example, the characteristic relating to the electrical resistance value when a pressure stimulus is applied is represented by a curve Edx. The electrical resistance value on the curve Edx is defined as a base resistance value, and this base resistance value is defined as the deterioration degree. The deterioration degree to which the first feature is applied may use the base resistance value as is, or may be a difference or ratio from the electrical resistance value at the initial measurement stage of the object 2 or at the initial formation stage of the object 2. The deterioration degree of the object 2 may also be associated with a message intuitive to the user. For example, a threshold Eth for the electrical resistance value used to determine the deterioration state may be determined in advance, and a resistance value less than the threshold Eth may be set as an undegraded state, while a resistance value equal to or greater than the threshold Eth may be set as a deteriorated state. The threshold Eth may be determined in advance through experiments or other means in accordance with the geometric or mechanical characteristics of the object 2, such as the initial shape and softness of the object 2 at the initial measurement stage. In the above, the state is defined using the threshold value Eth as a boundary, but the number of thresholds is not limited to one, and multiple thresholds may be defined to define the degradation state in stages.The degree of degradation may be determined by using the amount of change indicated by the current value relative to the initial base resistance value or electrical resistance value as a degradation index for determining degradation.

[0082] The second feature is that the deterioration state of the object 2 is reflected in the electrical characteristics for one cycle of application and release of pressure stimulus or in the electrical characteristics for a predetermined cycle. Specifically, the deterioration state of the object 2 is reflected in a change in the electrical resistance value in the electrical characteristics for application and release of pressure stimulus.

[0083] FIG. 11 shows an example of the electrical characteristics of the object 2 based on the above-described test results. In FIG. 11, electrical characteristics 41 show an example of the electrical characteristics of the object 2 when the application and release of pressure stimuli was repeated a predetermined number of times in the initial stage of measurement. Electrical characteristics 42 show the electrical characteristics of the object 2 at an elapsed time after the application and release of pressure stimuli was continuously repeated from the initial stage of measurement, and a predetermined number of times (increased frequency of deformation) of pressure stimuli was applied. In FIG. 11, P1 and P2 indicate the time periods when two pressure stimuli were applied in the same cycle, at the initial stage and the elapsed stage. As shown in FIG. 11, the electrical characteristics of the object 2 differ between the initial stage and the elapsed stage. Although FIG. 11 shows the elapsed time when the frequency of the predetermined deformation was increased, the elapsed time may also be a time during which the object 2 was left alone without applying or releasing pressure stimuli. Furthermore, the elapsed time may include continuous repetition of the application and release of pressure stimulation, or may include intermittent repetition, and may be the time when the frequency of deformation is continuously or intermittently increased and a predetermined time has elapsed after the increase.

[0084] FIG. 12 is a conceptual diagram illustrating the electrical characteristics of the object 2 shown in FIG. 11 from the perspective of changes in electrical resistance value. FIG. 12 shows the concept of electrical characteristics related to the ratio between the measured electrical resistance value and the average electrical resistance value as a time characteristic. Note that FIG. 12 shows electrical characteristics measured using a flat-plate-shaped pressing member 84 with a substantially flat contact portion. The time characteristic at the beginning of the measurement is shown by a solid line as characteristic 43, and the time characteristic over time is shown by a dotted line as characteristic 44. Also, FIG. 12 shows the pressure characteristic of the pressure stimulus applied to the object as characteristic 45. Note that in FIG. 12, P1a and P2a indicate the time when the pressure stimulus was applied to the object 2 at its maximum value (immediately before the pressure stimulus was released).

[0085] As shown in Figure 12, in terms of changes in electrical resistance, the electrical characteristics of the object 2 differ between the time when a pressure stimulus is applied and the time when it is released. Specifically, with regard to the change trend (e.g., rate of change) in the electrical resistance when the pressure stimulus is applied, there is a slight difference between the trend Wa at the beginning of measurement and the trend Wb over time. In contrast, with regard to the change trend (e.g., rate of change) in the electrical resistance when the pressure stimulus is released, there is a tendency for the difference between the trend Wc at the beginning of measurement and the trend Wd over time to become larger. Furthermore, the difference at the time when the pressure stimulus is released tends to be larger compared to the difference at the beginning of measurement when the pressure stimulus was applied (the distance between the solid line and the dotted line is large in Figure 12).

[0086] Therefore, in the second feature, the change in the electrical resistance value in the electrical characteristics when a pressure stimulus is applied and released is applied as the deterioration state of the object 2. Specifically, for example, the change tendency of the electrical resistance value (e.g., the rate of change) or the difference between the minimum values ​​of the change in the electrical resistance value is used as the deterioration degree indicating the deterioration trend. As described above, the deterioration degree to which the second feature is applied may be associated with a message that is intuitive to the user, as in the deterioration degree to which the first feature is applied, or a threshold value for determining the deterioration state may be set in advance, and a non-degraded state or a deteriorated state may be set relative to the threshold value. Furthermore, the deterioration degree may use the amount of change indicated by the current value relative to the initial value as a deterioration index for determining deterioration.

[0087] The deterioration level to which the second feature is applied can be less affected by fluctuations in the absolute value of the electrical characteristic than the deterioration level to which the first feature is applied, making it possible to identify the deterioration state of the object 2.

[0088] The third feature is that the deterioration state of the object 2 is reflected in a change in frequency of the electrical characteristics of the object 2, that is, a change in frequency of the time characteristics of the time-series electrical resistance value detected in the object 2.

[0089] Fig. 13 is a conceptual diagram showing an example of the frequency characteristics (power spectrum) of the electrical characteristics of the object 2. In Fig. 13, characteristic 46 shows an example of the conceptual power spectrum of the electrical characteristics 41 (e.g., Fig. 12) of the object 2 when the application and release of pressure stimuli is repeated a predetermined number of times in the initial stage of measurement. Also, characteristic 42 shows the power spectrum of the electrical characteristics 42 (e.g., Fig. 12) of the object 2 at an elapsed time after the application and release of pressure stimuli is continuously repeated from the initial stage of measurement and a predetermined number of times (increase in frequency of deformation).

[0090] As shown in FIG. 13 , the flexibility of the object 2 deteriorates as at least one of the physical quantities of the deformation frequency and the elapsed time increases (e.g., as the number of deformations increases or the elapsed time increases), and a phenomenon in which the flexibility changes from its previous state appears in the high-frequency components. That is, the high-frequency components change as the deformation frequency increases, or as the elapsed time, such as the time the object 2 is left unused, increases. The high-frequency components may also change as both the deformation frequency and the elapsed time increase. In the example shown in FIG. 13 , the high-frequency components gradually decrease as the time elapses with an increase in the deformation frequency, indicating that the flexibility gradually becomes softer and deteriorates. Note that while FIG. 13 illustrates the elapsed time associated with a predetermined increase in the frequency of deformation, the elapsed time may also be the time the object is left unused, or may be a continuous or intermittent repetition of the deformation, or may be a continuous or intermittent increase in the frequency of deformation and a predetermined time after the increase.

[0091] Specifically, in the frequency characteristics (power spectrum) of the electrical characteristics, consider the power difference Ps between a first frequency f1 and a second frequency f2 that is higher than the first frequency f1. The power difference Ps2 after a certain period of time (characteristic 47) tends to be larger than the power difference Ps1 at the beginning of measurement (characteristic 46).

[0092] Therefore, in the third feature, a change in frequency in the electrical characteristics of the object 2 is applied as the deterioration state of the object 2. Specifically, for example, the change tendency of the frequency in the electrical resistance value (for example, the amount of change or the rate of change) is set as the deterioration degree indicating the deterioration tendency. As described above, the deterioration degree to which the third feature is applied may be associated with a message that is intuitive to the user, or a threshold value for determining the deterioration state may be determined in advance, and an undegraded state or a deteriorated state may be set relative to the threshold value. Furthermore, the deterioration degree may use the amount of change indicated by the current value relative to the initial value as a deterioration index for determining deterioration.

[0093] Compared to the deterioration level to which the first or second feature is applied, the deterioration level to which the third feature is applied can be less affected by fluctuations in the absolute value of the electrical characteristics and by relative fluctuations in the electrical characteristics, making it possible to identify the deterioration state of the object 2 with higher accuracy.

[0094] The above-described deterioration level makes it possible to set deterioration state information (state data 3) indicating the deterioration state of the object 2. Therefore, by deriving the above-described features indicated in the electrical characteristics of the object 2, it becomes possible to associate the deterioration state of the object 2 with the time-series electrical characteristics related to the deformation of the object 2 as another piece of learning data.

[0095] Therefore, the learning data is a set of information indicating the time-series electrical characteristics of the object 2 (i.e., the conductive urethane 22) and the deterioration state of the object 2 relative to the electrical characteristics. An example of the learning data is shown in the table below. Table 2 is an example of a data set that associates time-series electrical resistance data (r) with condition data (R) indicating the deterioration state as learning data regarding the deterioration state of the object 2 including the conductive urethane 22.

[0096] [Table 2]

[0097] Table 2 details the status data (R) in Table 1, corresponding to the degradation state. The status data (R) may correspond to any of the features appearing in the electrical characteristics described above. For example, the degradation level (E) to which the first feature is applied may be applied. Examples of degradation levels to which the first feature is applied include the base resistance value, the difference or ratio from the initial electrical resistance value, a message indicating the degradation level, and a degradation index derived from these degradation levels. For the second feature, the degradation level may be determined by a change in the electrical resistance value in the electrical characteristics, such as a change trend (W) in the rate of change of the electrical resistance value, or the like. For the degradation level to which the third feature is applied, the degradation level may be determined by a change in the frequency in the electrical characteristics, such as a change amount or rate of change in the electrical resistance value, or the like. Therefore, the time-series electrical characteristics of the object 2 (conductive urethane 22) characteristically reflect the degradation state of the object 2, and therefore function effectively in the learning process.

[0098] Next, a learning process for generating a learning model 51 relating to the deterioration state of the object 2 according to this embodiment will be described.

[0099] 14 shows an example of the flow of the learning process executed in the learning processing unit 52. The learning process is performed by the CPU (not shown) in the learning processing unit 52 described above.

[0100] In step S110, the electrical characteristics (input data 4) of the object 2 are acquired. In the next step S111, the electrical characteristics (input data 4) of the object 2 are first analyzed to derive a degradation level representing the above-mentioned characteristics, thereby acquiring status data 3 indicating the degradation state of the object 2. In this step S111, the input data 4, which is the electrical characteristics of the object 2, is associated with the status data 3 indicating the degradation state of the object 2 as a result of the analysis, and a set of input data 4 (electrical resistance) labeled with the status data 3 (degradation level) is acquired as training data. Next, in step S112, a training model 51 is generated using the acquired training data. That is, a set of information on weight parameters (weights or strengths) of connections between nodes in the training results learned using a large amount of training data as described above is acquired. Then, in step S114, data expressed as a set of information on weight parameters (weights or strengths) of connections between nodes in the training results is stored as the training model 51.

[0101] Next, a process for estimating the deterioration state of the object 2 using the learning model 51 according to this embodiment will be described.

[0102] As described above, the deterioration state estimation device 1A uses a trained learning model 51 generated by the method exemplified above, and if a sufficiently trained learning model 51 is used, it is not impossible to identify the deterioration of the object 2 from the electrical characteristics of the object 2. The degradation state estimation device 1 for an object is an example of an estimation unit and an estimation device of the present disclosure.

[0103] 15 shows an example of the flow of a process for estimating the deterioration state of the object 2 by the control program 108P executed by the computer main body 100. The estimation process shown in FIG. 15 is executed by the CPU 102 when the computer main body 100 is powered on.

[0104] First, in step S201, the learning model 51 that has been trained for the above-mentioned degradation state estimation is read from the learning model 108M in the auxiliary storage device 108 and expanded into RAM 104, thereby obtaining the learning model 51 and constructing the learning model 51.

[0105] Next, in step S203, the time-series input data 4 (electrical characteristics) of the unknown object 2 (estimation object) is acquired via the detection unit 118.

[0106] Next, in step S205, output data 6 (deterioration state of the estimation object) corresponding to the input data 4 (electrical characteristics) acquired in step S203 is estimated using the learning model 51 acquired in step S201. Then, in the next step S207, the output data 6 (deterioration state of the estimation object) of the estimation result is output via the communication unit 114, and this processing routine ends.

[0107] In step S207, it is also possible to output support information such as a notification of the replacement time. For example, the above-mentioned intuitive message to the user may be output as the deterioration level to the communication unit 114 or the operation display unit 116. Specifically, a message indicating either an undegraded state when the level is below a threshold for determining the deterioration state, or a degraded state where replacement is recommended when the level is equal to or above the threshold, may be output as information. By outputting support information in this manner, the user can replace the object 2 at an appropriate time when it reaches a predetermined deterioration state.

[0108] The estimation process shown in FIG. 15 described above is an example of a process executed by the estimation method of the present disclosure.

[0109] As described above, according to the present disclosure, it is possible to estimate the deterioration state of an estimation object whose deterioration state is unknown from the electrical characteristics of the estimation object. That is, it is possible to estimate the deterioration state of the estimation object by utilizing the electrical characteristics of the object 2 when it is deformed, without using a special detection device.

[0110] Furthermore, by outputting the estimation result of the deterioration state of the estimation target, the deterioration state can be estimated including the flexibility of the conductive urethane 22 that functions as a sensor included in the target 2. This makes it possible to replace the target 2 at an appropriate time when it reaches a predetermined deterioration state, even if its function as a sensor deteriorates in the early stages of its formation.

[0111] <Modification> The above-mentioned degradation state estimating unit 10 can be configured by function in accordance with the features included in the above-mentioned electrical characteristics.

[0112] 16 shows a modified example of the configuration of the degradation state estimation unit 10. The degradation state estimation unit 10 includes a degradation state analysis unit, a comparison determination unit, and a threshold value storage unit.

[0113] The degradation state analysis unit is a functional unit that analyzes the degradation state of the object 2 containing the conductive urethane 22 using time-series electrical characteristics (input data 4) that change in response to deformation of the conductive urethane 22. The degradation state analysis unit derives a degradation level indicated by a feature included in the electrical characteristics, which is any one of the first to third features described above. The comparison / determination unit is connected to a threshold storage unit that stores a threshold for determining the degradation state, compares the degradation state of the analysis result with the threshold for determining the degradation state, and outputs the comparison result as a degradation level. The threshold storage unit may store the threshold for determining the degradation state in the ROM 106 or the auxiliary storage device 108, for example, to associate the threshold with a message that is intuitive to the user. Furthermore, the comparison / determination unit may, for example, use the threshold for determining the degradation state to determine that the derived degradation level is not deteriorated if it is less than the threshold, or that it is deteriorated if it is equal to or greater than the threshold. In this way, by configuring the degradation state estimation unit 10 by function corresponding to the feature included in the electrical characteristics, the unit can be easily applied to hardware configurations such as independent device configurations and electrical circuit configurations.

[0114] In the present disclosure, a case where conductive urethane is used as an example of a flexible member has been described, but the flexible member need only have flexibility, and is of course not limited to the conductive urethane described above.

[0115] Furthermore, the technical scope of the present disclosure is not limited to the scope described in the above embodiments. Various modifications or improvements can be made to the above embodiments without departing from the gist of the present disclosure, and such modifications or improvements are also included in the technical scope of the present disclosure.

[0116] Furthermore, in the above embodiment, the estimation process and learning process are described as being realized by a software configuration using a flowchart, but this is not limited to this, and for example, each process may be realized by a hardware configuration.

[0117] Furthermore, a part of the estimation device, for example, a neural network such as a learning model, may be configured as a hardware circuit. [Explanation of symbols]

[0118] 1 Estimation device 2. Object 3. State Data 4. Input Data 5 Estimation part 6 Output data 6A Generated Output Data 21 Materials 22 Conductive urethane 51 Learning Model 52 Learning processing unit 54 Generator 56 Arithmetic unit 75 detection points 76 Electrical characteristics detection unit< / prc>

Claims

1. a detection unit that detects electrical characteristics between a plurality of predetermined detection points on a flexible material of an object that has conductivity and whose electrical characteristics change in response to deformation; an estimation unit that uses, as learning data, time-series electrical characteristics that change in response to deformation of the flexible material due to a pressure stimulus applied to the object and degradation state information that indicates a degradation state related to the deformation of the flexible material, and inputs the time-series electrical characteristics detected by the detection unit of the estimation object into a learning model that has been trained to input the time-series electrical characteristics and output the degradation state information, and estimates and outputs the degradation state information corresponding to the input time-series electrical characteristics; An estimation device comprising:

2. the electrical property is volume resistivity, the object includes a flexible member, The deterioration state information indicates a deterioration degree in which the degree of deterioration increases as at least one physical quantity of the number of deformations, the deformation frequency, and the elapsed time from the initial state of the object increases. The estimation device according to claim 1 .

3. The degradation state information indicates the degradation state regarding the deformation of the flexible material, and the degradation state indicates the degree of degradation of at least one of the electrical characteristics of the object when it is deformed from a shape before deformation and the electrical characteristics when it is restored to the shape before deformation in a time series of electrical characteristics. The estimation device according to claim 2 .

4. The deterioration state information indicates a deterioration state related to the deformation of the flexible material, which indicates a degree of deterioration in which, when a frequency analysis of time-series electrical characteristics is performed, the degree of deterioration increases as the power of the frequency of the analysis result becomes larger than the power of the frequency at a predetermined time. The estimation device according to claim 2 .

5. The object includes a material having a structure having at least one of a fibrous and a mesh-like skeleton, or a structure having a plurality of minute air bubbles dispersed therein, and at least a part of the urethane material is made electrically conductive. The estimation device according to any one of claims 2 to 4.

6. The learning model includes a model generated by learning using a network based on reservoir computing using the flexible material as a reservoir. The estimation device according to any one of claims 1 to 5.

7. The computer detecting electrical characteristics between a plurality of predetermined detection points on a flexible material of an object having conductive flexible material whose electrical characteristics change in response to deformation; The time-series electrical characteristics that change in response to deformation of the flexible material due to pressure stimuli applied to the object are associated with degradation state information that indicates a degradation state related to the deformation of the flexible material, and the time-series electrical characteristics of the detected estimation object are input to a learning model that has been trained to input the time-series electrical characteristics and output the degradation state information, and the degradation state information corresponding to the input time-series electrical characteristics is estimated and output. Estimation method.

8. To the computer detecting electrical characteristics between a plurality of predetermined detection points on a flexible material of an object having conductive flexible material whose electrical characteristics change in response to deformation; The time-series electrical characteristics that change in response to deformation of the flexible material due to pressure stimuli applied to the object are associated with degradation state information that indicates a degradation state related to the deformation of the flexible material, and the time-series electrical characteristics of the detected estimation object are input to a learning model that has been trained to input the time-series electrical characteristics and output the degradation state information, and the degradation state information corresponding to the input time-series electrical characteristics is estimated and output. A guessing program to execute the process.

9. an acquisition unit that acquires electrical characteristics from a detection unit that detects electrical characteristics between a plurality of predetermined detection points on a flexible material of an object having conductive flexible material whose electrical characteristics change in response to deformation, and acquires deterioration state information that indicates a deterioration state of the flexible material; a learning model generation unit that generates a learning model by associating time-series electrical characteristics that change in response to deformation of the flexible material due to pressure stimuli applied to the object with degradation state information that indicates a degradation state related to the deformation of the flexible material as learning data, inputting the time-series electrical characteristics when the flexible material is deformed due to pressure stimuli applied to the object, and outputting the degradation state information that indicates a degradation state related to the deformation of the flexible material; A learning model generation device comprising:

Images