JP2023000311A - Prediction device, prediction method and prediction program - Google Patents

Abstract

To provide a prediction device, a prediction method and a prediction program that enable a walking trainee to actively works for training.SOLUTION: A prediction device comprises an acquisition part, a generation part and a prediction part. The acquisition part acquires first information including information associated with walking actions of a walking trainee and second information including information associated with a body and symptoms of the walking trainee. The generation part inputs the first information and second information acquired by the acquisition part to a model having learnt information on other walking trainees in walking training done in the past and correlations between details and results of the walking training done in the past, and generates training details for the walking trainee based upon a result output from the model. The prediction part predicts a future walking function of the walking trainee based upon the training details generated by the generation part.SELECTED DRAWING: Figure 1

Description

The present disclosure relates to a prediction device, a prediction method, and a prediction program that perform prediction processing related to walking training.

Various studies have been conducted on the movement of the human body. For example, techniques have been developed for improving a person's movement function and efficiently performing rehabilitation based on information obtained from measured walking patterns and sensors attached to the body.

As an example, there is known a technique of visualizing the walking state of a subject based on a gait pattern and walking parameters and allowing the subject to imagine the walking habits of the subject (see Patent Document 1 below). Also, there is known a technique that enables more active walking training to be performed based on the remaining walking ability of a walking disabled person (see Patent Document 2 below).

Japanese Patent Application Laid-Open No. 2003-204953 Japanese Patent No. 6175050

According to the conventional technology, since the walking state can be presented to the pedestrian, effective walking training can be performed.

However, the prior art presents the subject's walking state or assists walking, and it is difficult for the walking trainer to imagine the results of the training. In other words, the walking trainee cannot imagine how the walking function will be improved by how much training is continued for how long. For this reason, there is a possibility that the walking trainee may lose motivation or feel uneasy about the training.

Therefore, the present disclosure proposes a prediction device, a prediction method, and a prediction program that allow a walking trainee to enthusiastically engage in training.

In order to solve the above problems, the prediction device of one embodiment according to the present disclosure includes first information including information on the walking motion of the walking trainee, and second information on the body and symptoms of the walking trainee. An acquisition unit that acquires information and the first information and second information acquired by the acquisition unit, information of other walking trainees in walking training that has been performed in the past, and the information that has been performed in the past A generation unit that inputs the correlation between the content and results of walking training that has been learned into a learned model, and based on the results output from the model, generates training content for the walking trainee, and generated by the generation unit and a prediction unit that predicts the future walking function of the walking trainee based on the training content.

It is a block diagram which shows typically the flow of the process which the prediction apparatus which concerns on embodiment performs. It is a sequence diagram which shows the processing procedure of the prediction system which concerns on embodiment. It is a figure which shows the structural example of the user terminal and prediction apparatus which concern on embodiment. It is a figure which shows an example of the user information storage part which concerns on embodiment. It is a figure which shows an example of the learning data storage part which concerns on embodiment. It is a figure which shows an example of the model memory|storage part which concerns on embodiment. It is a figure which shows an example of the prediction result memory|storage part which concerns on embodiment. 4 is a flowchart (1) showing the procedure of information processing according to the embodiment; 4 is a flowchart (2) showing the procedure of information processing according to the embodiment; It is a hardware block diagram which shows an example of the computer which implement|achieves the function of a prediction apparatus.

Embodiments of the present disclosure will be described in detail below with reference to the drawings. In addition, in the following embodiment, the overlapping description is abbreviate|omitted by attaching|subjecting the same code|symbol to the same site|part.

(1. Embodiment)
[1-1. An example of prediction processing according to the embodiment]
FIG. 1 is a block diagram schematically showing the flow of prediction processing according to the embodiment. A prediction device 100 illustrated in FIG. 1 is an information processing device that executes prediction processing according to the embodiment. The prediction device 100 is used, for example, by facilities that perform walking training such as rehabilitation, and training instructors such as physical therapists.

In gait training carried out in rehabilitation, etc., it is more effective if the person undergoing gait training (hereinafter referred to as "walking trainee") is conscious of the future walking function such as the recovery period and recovery content of training. On the other hand, when the walking trainee cannot imagine the future walking function, the walking trainee loses motivation for training, and the training result may decrease.

Therefore, as described below, the prediction device 100 predicts the content of how much the walking function will be recovered by the training of the walking trainee, and presents it to the walking trainee. Thereby, the prediction device 100 assists the walking trainee to train enthusiastically. Hereinafter, an overview of the processing executed by the prediction device 100 will be described along the flow with reference to FIG. 1 .

First, the prediction device 100 acquires a learning data set (step S10). Specifically, the prediction device 100 acquires various data in which information of a walking trainee in walking training performed in the past and the contents and results of the walking training are set. The information of the walking trainee is, for example, the physical information of the walking trainee (sex, age, height, weight, inseam, sole size, etc.). In addition, the information of the walking trainee is information on the symptoms that triggered the walking training (specific cases such as fractures and sprains, elapsed time from onset, symptom classification, symptom location, presence or absence of paralysis, in daily life content such as how much operation is possible). In addition, the information of the walking trainee includes walking data obtained before and after walking training and during training (for example, foot pressure during walking obtained by various sensors during walking, acceleration, stride length, etc.).

The content of walking training is the training content and training time actually decided by the training instructor (for example, walking training using parallel bars for 30 minutes, stretching, functional recovery of joints and muscles using equipment, etc.). . In addition, the results of walking training include the degree of recovery of walking function (becoming able to walk using a cane, being able to walk independently, being able to go up and down stairs on one's own, etc.) and recovery. It is the period (for example, how long it takes to reach the above-mentioned degree of recovery from the onset of symptoms).

The prediction device 100 acquires the learning data set from, for example, data input to an electronic medical record, a log recorded by a training instructor, or the like. The learning data set may be a combination of the above-mentioned various types of information, information of the walking trainee in the past walking training (so-called input in machine learning), and one correct data (machine learning so-called labels) may be combined. Specifically, the learning data set includes the information of the walking trainee (for example, (gender, age, height, weight, case, etc.) vector obtained by digitizing (normalizing) each input) and Correct data (information that "walking training using assistance" was performed for "1 hour" on "1 day", and as a result, "walking using a cane" became possible "2 months later") and consists of a combination of

Note that the training content may be used not only as a label but also as an input. That is, the learning data set uses the information (input) of the walking trainee as "what kind of training" was performed "how many hours" in "one day", and as a result, "two months later" using a cane The correct data (label) may be the fact that the patient is able to walk on the ground. Various known techniques may be used for generating and acquiring such learning data sets and quantifying (normalizing) each item. In this specification, the individual information of the walking trainee in the walking training performed in the past is referred to as learning data, and the set of the learning data and the correct data that is the contents and results of the walking training is referred to as the learning data set. .

The prediction device 100 performs learning processing based on the plurality of learning data sets obtained in step S10 (step S11). Specifically, the prediction device 100 learns the correlation between the information of the walking trainee in past walking training and the content and result of the walking training.

That is, the prediction device 100 uses the information of the walking trainee in the past walking training as an input (also referred to as a feature, an explanatory variable, etc.), and the contents and results of the walking training as labels (correct data, teacher data, (also referred to as an objective variable) is generated (step S12). Although illustration in FIG. 1 is omitted, the prediction device 100 may generate a plurality of models. For example, the prediction device 100 includes a model that has learned the correlation between "walking trainee information" and "training content", "walking trainee information and training content" and "recovery content", and "walking trainee information and training" and a model that learned the correlation with the "recovery period" may be generated separately. Then, the prediction device 100 may perform each process using an appropriate model among the plurality of models in the generation process and the prediction process described later.

Then, the prediction device 100 that has generated the model performs prediction processing using the model when information is acquired from the person to be predicted. In the example of FIG. 1, the prediction device 100 performs prediction processing on the user 20 who is a walking trainee to be predicted.

When performing the prediction process according to the embodiment, the prediction device 100 acquires data to be input to the model (step S15). First, the prediction device 100 acquires information regarding the user 20 . Specifically, the prediction device 100 acquires physical information such as gender and age of the user 20 via the user terminal 10 such as a smartphone or tablet terminal. The prediction device 100 may acquire information transmitted from the user terminal 10, or may acquire data of information obtained through an interview conducted by the training instructor. In this specification, "user" may mean "user terminal used by the user". For example, "a user transmits physical information" actually means "a user terminal used by a user transmits physical information".

Also, the prediction device 100 acquires symptom information of the user 20 via an electronic medical record input by a doctor or the like who examined the user 20 . Alternatively, the prediction device 100 may acquire data of symptom information determined by a physical therapist or the like.

The prediction device 100 also acquires walking data obtained by measuring the walking state of the user 20 with a sensor. For example, the prediction device 100 acquires the acceleration, pressure, stride length, and the like in walking of the user 20 via sensors installed in a training facility or the like where rehabilitation is performed.

Then, the prediction device 100 inputs various information obtained in step S15 to the model generated in step S12 (step S16).

Subsequently, the prediction device 100 generates training content for the user 20 based on information output from the model. Also, the prediction device 100 predicts the recovery content and recovery period of the user 20 based on the generated training content (step S17).

For example, based on the information output from the model, the prediction device 100 generates training content such as "what kind of training" and "how many hours" to be performed in "one day". Specifically, the prediction device 100 selects the content and time with a high score among the training content output from the model to which the information of the user 20 is input, that is, the content determined to be appropriate for the user 20. and generate training content. For example, when the model generated by the prediction device 100 outputs the appropriateness of the training as a score from 0 to 1 (it is judged that the training content is more appropriate as it approaches 1), the prediction device 100 receives the input Output a score based on the information. As an example, the prediction device 100 determines that if the score of “walking training using a walking aid” for “1 hour” is 0.5 and the score of “stretching” for “2 hours” is 0.3, more A high-score “one-hour” “walking training using a walking aid” is generated as the training content for the user 20 . When generating training content by combining a plurality of menus, prediction device 100 may generate training content by selecting menus in descending order of training content with a higher score.

Subsequently, the prediction device 100 predicts the future walking function of the user 20, that is, the recovery period and recovery content, based on the generated training content. For example, the prediction device 100 predicts the recovery period and recovery details of the user 20 using a model that has learned the correlation between the input user information, the training content to be performed, and the recovery period and recovery details. That is, the prediction device 100 uses a model that outputs information such as what kind of training a past trainee who has similar physical information and symptom information to the user 20 will perform and what result will occur. By doing so, the future walking function of the user 20 is predicted.

Then, the prediction device 100 transmits the generated training content and the predicted result to the user terminal 10 and controls to display them on the screen of the user terminal 10 . As a result, the user 20 can grasp the content of training to be performed by the user 20, and can predict the recovery period.

Note that the user 20 may provide feedback on training content and predictions generated by the prediction device 100 . For example, the user 20 provides feedback such as requesting a change in training content that the user does not feel comfortable with, presenting desired training content, or changing the content so as to shorten the predicted recovery period. Upon receiving such feedback, the training instructor may change the training content based on the expert's knowledge, or may input further information to the prediction device 100 to generate new training content.

For example, the prediction device 100 compares the training content desired by the user 20 with the training content output from the model that has learned the correlation between the "walking trainee information" and the "training content". Then, if the score corresponding to the training content desired by the user 20 and the training content output from the model and the score are within a predetermined range (for example, within 0.2), the prediction device 100 predicts the training content to the user. 20 performs processing such as replacing with desired training contents. On the other hand, if the score corresponding to the training content desired by the user 20 and the training content output from the model and the score exceed a predetermined range, the prediction device 100 rejects the training content desired by the user 20. good too. For example, if the score corresponding to the training content desired by the user 20 is 0.2 and the score corresponding to the training content output from the model is 0.3, the prediction device 100 predicts that the difference between the two scores is 0.2. 2 or less and within the predetermined range, and the training content desired by the user 20 is adopted. On the other hand, if the score corresponding to the training content desired by the user 20 is minus 0.3 and the score corresponding to the training content output from the model is 0.3, the prediction device 100 determines that the difference between the two scores is a predetermined value. Reject the training content desired by the user 20 as out of range. Such score adjustment is a matter of design and may be arbitrarily set by a training instructor such as a physical therapist or a doctor.

The outline of the processing executed by the prediction device 100 has been described above with reference to FIG. Here, the prediction process according to the embodiment may be executed by the prediction system 1 including the prediction device 100. FIG. Therefore, the configuration and processing flow of the prediction system 1 will be described with reference to FIG. 2 and subsequent drawings.

FIG. 2 is a sequence diagram showing the processing procedure of the prediction system 1 according to the embodiment. As shown in FIG. 2 , prediction system 1 includes user terminal 10 , prediction device 100 , and sensor 30 .

The sensor 30 is a measuring device installed in a facility or the like where walking training is performed. For example, the sensor 30 is a pressure mat spread on the floor, a pressure sensor provided in shoes or insoles for walking training, a motion analysis camera, or the like. It is assumed that the sensor 30 has a communication function for transmitting measured information (walking data) to the user terminal 10 and the prediction device 100 .

As shown in FIG. 2, the sensor 30 measures walking data of the user 20 during exercise (step S20). The sensor 30 then transmits the measured walking data to the prediction device 100 (step S21).

The prediction device 100 receives the walking data measured by the sensor 30 (step S22). In addition, the user 20 measures his/her height, weight, and other physical information as well as the measurements by the sensor 30, and inputs the measurements to the user terminal 10 (step S23). The user 20 also inputs symptom information such as a specific case judged by a doctor or a training instructor to the user terminal 10 together with the physical information. The user 20 inputs information on an application installed in the user terminal 10, for example.

The user terminal 10 transmits the physical information and symptom information input by the user 20 to the prediction device 100 (step S24). The prediction device 100 receives physical information and symptom information from the user terminal 10 (step S25). Then, the prediction device 100 converts the information obtained in steps S22 and S25 into data, inputs the data into the model, generates training content, and predicts the future walking function of the user 20 (step S26).

The prediction device 100 transmits the generated training content and prediction content to the user terminal 10 (step S27). The user terminal 10 receives the training content and prediction content (step S28), and displays the results on the screen (step S29). This allows the user 20 to grasp his/her own training content and recovery prediction.

As described above with reference to FIGS. 1 and 2, the prediction device 100 according to the embodiment includes the first information including information (walking data) on the walking motion of the walking trainee, the walking trainee's body and and obtaining second information including information about the symptoms. Then, the prediction device 100 uses the first information and the second information as the correlation between the information of other walking trainees in walking training performed in the past and the content and results of walking training performed in the past. is input to the learned model, and based on the results output from the model, training content for the walking trainee is generated. In addition, the prediction device 100 predicts the future walking function of the walking trainee based on the generated training content.

That is, the prediction device 100 uses a model generated based on information on past training to generate training content to be performed for the user 20, and predicts the period until recovery and the function to be recovered. As a result, the prediction device 100 allows the user 20 to grasp the content of the training before the training, thereby motivating the user 20 to perform the training. In addition, the prediction device 100 receives feedback from the user 20 on the presented training content, thereby assisting the user 20 in a more proactive attitude toward training and improving training results.

[1-2. Configuration of User Terminal and Prediction Apparatus According to Embodiment]
Next, configurations of the user terminal 10 and the prediction device 100 that execute information processing according to the embodiment will be described. FIG. 3 is a diagram showing a configuration example of the user terminal 10 and the prediction device 100 according to the embodiment.

First, the user terminal 10 will be described. As shown in FIG. 3 , the user terminal 10 has a communication section 11 , a reception section 12 , a transmission section 13 , a display control section 14 and a user information storage section 15 . Note that the user terminal 10 may have an input unit (for example, a touch panel, etc.) that receives various operations from the user 20 or the like, and a display unit (for example, a liquid crystal display, etc.) for displaying various information.

The communication unit 11 is realized by, for example, a NIC (Network Interface Card) or the like. The communication unit 11 is connected to a network N (for example, the Internet) by wire or wirelessly, and transmits and receives information to and from the sensor 30, the prediction device 100, and the like via the network N.

Control units such as the reception unit 12, the transmission unit 13, and the display control unit 14 are stored inside the user terminal 10 by, for example, a CPU (Central Processing Unit), an MPU (Micro Processing Unit), a GPU (Graphics Processing Unit), or the like. The program is executed using a RAM (random access memory) or the like as a work area. Also, the control unit is a controller, and is implemented by an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array).

The reception unit 12 receives various types of information. For example, the reception unit 12 receives input of physical information (sex, age, height, weight, inseam, sole size, etc.) input from the user 20 . The reception unit 12 also receives the user's symptom information (a specific case, elapsed time from the onset, symptom category, symptom location, presence or absence of paralysis, details of activities of daily living, etc.). The reception unit 12 appropriately stores the received information in the user information storage unit 15 .

The transmission unit 13 transmits various information. For example, the transmission unit 13 transmits various information received by the reception unit 12 to the prediction device 100 . In addition, when the reception unit 12 receives feedback on training content from the user 20 , the transmission unit 13 transmits the information to the prediction device 100 . Also, when information measured by the sensor 30 is accepted by the user terminal 10 , the transmission unit 13 transmits the information to the prediction device 100 .

The display control unit 14 controls to display various information on the screen of the user terminal 10 . For example, the display control unit 14 displays on the screen a user interface for the user 20 to input physical information and symptom information. In addition, the display control unit 14 displays training content and prediction results generated by the prediction device 100 on the screen.

The user information storage unit 15 is implemented by, for example, a semiconductor memory device such as a RAM (Random Access Memory) or a flash memory, or a storage device such as a hard disk or an optical disk.

FIG. 4 shows an example of information stored in the user information storage unit 15. As shown in FIG. FIG. 4 is a diagram showing an example of the user information storage unit 15 according to the embodiment. In the example shown in FIG. 4, the user information storage unit 15 has items such as "user ID", "physical information", and "symptom information". Further, the "physical information" has items such as "sex", "age", "height", "weight", "inseam", and "sole size". The "symptom information" has items such as "concrete case", "elapsed time from onset", "symptom category", "symptom site", "presence or absence of paralysis", and "life activity".

In the examples shown in FIGS. 4 to 7, the information stored in the user information storage unit 15 and the storage unit 120, which will be described later, may be conceptually indicated as "A001". Each information to be explained is stored in the user information storage unit 15 and the storage unit 120 . In addition, the information shown in FIG. 4 is an example, and the user information storage unit 15 stores, in addition to the information shown in FIG. You may Further, the user information storage unit 15 may store walking data of the user 20 measured by the sensor 30 .

Next, the prediction device 100 will be described. As shown in FIG. 3 , prediction device 100 includes communication unit 110 , storage unit 120 , and control unit 130 . Note that the prediction device 100 includes an input unit (for example, a keyboard, a mouse, etc.) that receives various operations from an administrator or the like who manages the prediction device 100, and a display unit (for example, a liquid crystal display, etc.) for displaying various information. may have.

The communication unit 110 is implemented by, for example, a NIC. The communication unit 110 is connected to the network N by wire or wirelessly, and transmits and receives information to and from the user terminal 10, the sensor 30, and the like via the network N.

The storage unit 120 is realized by, for example, a semiconductor memory device such as a RAM or flash memory, or a storage device such as a hard disk or an optical disk. Storage unit 120 has learning data storage unit 121 , model storage unit 122 , and prediction result storage unit 123 . Each storage unit will be described below in order.

FIG. 5 shows an example of information stored in the learning data storage unit 121. As shown in FIG. FIG. 5 is a diagram showing an example of the learning data storage unit 121 according to the embodiment. In the example shown in FIG. 5, the learning data storage unit 121 has items such as "data ID", "input", and "label". "Input" has items such as "physical information", "symptom information", and "gait data". The "label" has items such as "training content", "recovery content", and "recovery period".

“Data ID” is identification information for identifying a learning data set. An “input” is a feature in a learning data set and various parameters that can be explanatory variables. "Physical information" is data corresponding to the physical information shown in FIG. 4, and is physical information such as sex, height and weight of each walking trainee. "Symptom information" is data corresponding to the symptom information shown in FIG. 4, and is information indicating symptoms such as specific cases of each walking trainee.

The “label” is correct data in the learning data set and various parameters that can be objective variables. "Training content" indicates the specific content of the training implemented for the walking trainee. "Recovery content" indicates the specific content of the walking function acquired (recovered) by the walking trainee after the training of the walking trainee. "Recovery period" indicates the training period that has elapsed until the walking trainee reaches the recovery content.

Note that the data set stored in the learning data storage unit 121 is an example, and the combination of inputs and labels may be arbitrarily rearranged according to the model to be generated. For example, the learning data set may be data whose inputs are "physical information" and "symptom information" and whose label is only "training content".

Next, FIG. 6 shows an example of information stored in the model storage unit 122. As shown in FIG. FIG. 6 is a diagram showing an example of the model storage unit 122 according to the embodiment. In the example shown in FIG. 6, the model storage unit 122 has items such as "model ID", "learning data", and "date and time of generation".

"Model ID" is identification information for identifying the generated model. “Learning data” indicates a learning data set used for model generation. "Created date and time" indicates the date and time when the model was created. Even after the model is generated, the model may be updated as appropriate by increasing the learning data set or changing the learning method.

Next, FIG. 7 shows an example of information stored in the prediction result storage unit 123. As shown in FIG. FIG. 7 is a diagram illustrating an example of the prediction result storage unit 123 according to the embodiment. In the example shown in FIG. 7, the prediction result storage unit 123 has items such as "user ID", "training content", "recovery period", and "predicted date and time".

"User ID" is identification information for identifying a walking trainee. "Training content" is information generated based on the information output from the model, and indicates the content of training to be performed for the walking trainee. The "recovery period" is information predicted by the prediction device 100 based on the generated training content, and indicates a period in which the walking trainee is expected to recover to a predetermined state. “Predicted date and time” indicates the date and time when the prediction process was performed by the prediction device 100 .

Returning to FIG. 3, the description is continued. The control unit 130 is realized, for example, by executing a program stored inside the prediction device 100 using a RAM or the like as a work area by a CPU, MPU, GPU, or the like. Also, the control unit 130 is a controller, and may be implemented by an integrated circuit such as an ASIC or FPGA, for example.

As shown in FIG. 3, the control unit 130 includes an acquisition unit 131, a learning unit 132, a generation unit 133, and a prediction unit 134, and implements or executes the information processing functions and actions described below. . Note that the internal configuration of the control unit 130 is not limited to the configuration shown in FIG. 3, and may be another configuration as long as it performs information processing described later.

Acquisition unit 131 acquires various types of information. For example, the acquisition unit 131 acquires first information including information about the walking motion of the user 20 and second information including information about the body and symptoms of the user 20 .

Specifically, the acquisition unit 131 acquires at least one of acceleration, pressure, and stride length in walking of the user 20 detected by the sensor 30 as the first information.

In addition, the acquisition unit 131 acquires at least one of the sex, age, height, weight, inseam, and sole size of the user 20 as second information as physical information. The acquiring unit 131 also acquires at least one of the specific case of the user 20, the elapsed time from the onset, the symptom classification, the symptom location, the presence or absence of paralysis, and the contents of the daily living activities as symptom information.

Moreover, the acquisition part 131 may acquire the training content or training period which the user 20 desires as 3rd information. The acquisition unit 131 may acquire training content or a training period desired by the user 20 before inputting the first information and the second information into the model, or may provide the user 20 with the generated training content or recovery period. After presentation, the training content or training period desired by the user 20 may be obtained as feedback.

The acquisition unit 131 appropriately stores the acquired information in the storage unit 120 . Further, the acquisition unit 131 may acquire information used for processing by the learning unit 132, the generation unit 133, and the prediction unit 134 from the storage unit 120 as appropriate.

The learning unit 132 learns the correlation between the information of the walking trainee in walking training performed in the past and the content and result of the walking training performed in the past. Then, the learning unit 132 generates a model capable of outputting training details, recovery details, and recovery period to be performed for the user 20 when information of the user 20 who intends to perform walking training is input. The learning unit 132 stores the generated model in the model storage unit 122 . As described above, the learning unit 132 may use various known techniques for model generation, training data set generation, normalization, and the like.

The generation unit 133 inputs the first information and the second information acquired by the acquisition unit 131 to the model generated by the learning unit 132, and based on the results output from the model, trains the user 20. to generate

Moreover, the generation unit 133 may generate the training content for the user 20 based on the training content or the training period desired by the user 20 . For example, the generation unit 133 generates a plurality of training contents suitable for the user 20 based on the first information and the second information. Then, when the training content desired by the user 20 is included in the generated training content, the generation unit 133 adds a predetermined score, and then determines the training content to be actually presented to the user 20 . Thereby, the generation unit 133 can generate training content that meets the user's 20 desires.

Further, when the generating unit 133 acquires the desired training content or training period from the user 20 after presenting the generated training content to the user 20, the score corresponding to the training content desired by the user 20 and the score from the model It may be determined whether or not to replace the desired training content by comparing with the output training content.

The prediction unit 134 predicts the future walking function of the user 20 based on the training content generated by the generation unit 133 .

For example, the prediction unit 134 predicts, as the future walking function of the user 20, the details of the recovery of the walking function of the user 20, the recovery prediction period, or the period until a different training content is newly implemented. The period until a different training content is newly implemented means, for example, the period from the end of "walking training using a cane" to the next step of walking training. That is, the prediction unit 134 may not only indicate the period until the walking function reaches the predicted recovery content, but may also predict the period for each stage of walking training. That is, the prediction unit 134 can prevent the user 20 from losing motivation for training by showing the recovery period up to the latest stage step by step.

The prediction unit 134 transmits the predicted result to the user terminal 10 and displays it on the screen of the user terminal 10 . After that, when receiving feedback from the user 20, the prediction unit 134 may display a new prediction result on the screen. For example, when the prediction unit 134 receives feedback from the user 20 such as "I want to train in a shorter period", the prediction unit 134 uses a model that has learned the correlation between the recovery period and the training content to shorten the recovery period. Generate training content such as Then, the prediction unit 134 causes the user terminal 10 to display a new training period predicted in accordance with the generated training content. In this way, the prediction unit 134 can predict a training period that meets the user's 20 request by performing prediction processing while interacting with the user 20 .

[1-3. Information processing procedure according to the embodiment]
Next, the procedure of information processing according to the embodiment will be described with reference to FIGS. 8 and 9. FIG. First, with reference to FIG. 8, the flow of learning processing executed by the prediction device 100 according to the embodiment will be described. FIG. 8 is a flowchart (1) showing the procedure of information processing according to the embodiment.

As shown in FIG. 8, the prediction device 100 acquires a learning data set (step S101). For example, the prediction device 100 automatically acquires the information of the electronic medical record, or through the input work by the training instructor, the information of the walking trainee in the walking training performed in the past, and the information performed in the past A training data set is obtained in which the contents and results of the walking training are combined.

Subsequently, the prediction device 100 generates a model based on the acquired learning data set (step S102). Specifically, the prediction device 100 generates a model showing the correlation between the information of the walking trainee in the walking training performed in the past and the details and results of the walking training performed in the past. Then, the prediction device 100 stores the generated model in the storage unit 120 (step S103).

Next, the flow of prediction processing executed by the prediction device 100 according to the embodiment will be described with reference to FIG. 9 . FIG. 9 is a flowchart (2) showing the procedure of information processing according to the embodiment.

As shown in FIG. 9, the prediction device 100 acquires user information (physical information and symptom information) of the user 20 who is the target of prediction processing (step S201). The prediction device 100 also acquires walking data of the user 20 (step S202). Note that the prediction device 100 may acquire walking data prior to step S201.

Subsequently, the prediction device 100 determines whether or not the training instructor has heard a request from the walking trainer such as the user 20 or the training instructor (step S203). The walking trainee's request is, for example, the training content and training period desired by the user 20, or the training content and training period recommended to the user 20 by the training instructor.

When the walking trainee desires (step S203; Yes), the prediction device 100 acquires the content desired by the walking trainee (step S204). On the other hand, when the walking trainee does not have a desire (step S203; No), the prediction device 100 skips the process of step S204.

Subsequently, the prediction device 100 inputs the acquired various information into the model (step S205). The prediction device 100 then transmits the output information (the generated training content and prediction result) to the user terminal 10 (step S206).

After that, the prediction device 100 determines whether or not feedback has been received from the user 20, the training instructor, or the like (step S207). The feedback is, for example, the training content and training period desired by the user 20, or a request to change the presented training content.

When feedback is obtained (step S207; Yes), the prediction device 100 executes generation processing or prediction processing based on the feedback information, and further transmits updated information to the user terminal 10 (step S208). On the other hand, when no feedback is obtained (step S207; No), the prediction device 100 skips the process of step S208 and ends the prediction process.

(2. Modified example of embodiment)
[2-1. Device configuration]
In the above-described embodiment, the user terminal 10 according to the present disclosure is a terminal handled by the user 20, which is a smart phone or a tablet terminal. However, the configuration of the user terminal 10 is not limited to this example as long as it can execute the processing shown in the embodiment.

For example, the user terminal 10 may be composed of multiple smartphones or tablet terminals. Specifically, the user terminal 10 may be composed of a plurality of terminals including a smartphone used by the user 20 and a tablet terminal provided by a training instructor or a doctor.

The sensor 30 includes a pressure sensor, a camera for measuring the walking state of the user 20, a plurality of measuring devices including a thermometer and a hygrometer, and a terminal device having a function of transmitting and receiving measured data. may be configured.

[2-2. learning data]
In the above-described embodiment, the learning data set is a combination of information of other walking trainees in walking training performed in the past, and the contents and results of the training. However, the learning data set is not limited to this. For example, the learning data set may incorporate the knowledge of experts such as doctors and physical therapists.

For example, the correct data is not only the content and results of walking training performed in the past, but also the content of walking training determined by the training instructor based on the information of other walking trainees in walking training performed in the past and It may be a prediction result. That is, the prediction device 100 uses the information of other walking trainees in walking training performed in the past as learning data, and the training instructor determines based on the contents and results of walking training performed in the past or the learning data. The training content for the user 20 may be generated using a model that has learned a data set having the walking training content and the prediction result as correct data.

The content of walking training is often determined based on expert knowledge. For this reason, the learning data set may use not only the actual past training, but also the training contents judged by experts based on the current state of the walking trainee and the prediction of the recovery period. As a result, the prediction device 100 can generate a model for predicting training contents and results in a more realistic manner, taking into account the expert's judgment. In addition, the prediction device 100 can learn a large amount of learning data sets by acquiring not only learning data sets based on actual cases, which tend to be insufficient, but also data based on expert knowledge as learning data sets. A highly accurate model can be generated.

[2-3. Generation and prediction processing based on the wishes of the walking trainee]
The prediction device 100 may acquire the training content and period desired by the walking trainee through interviews with experts, for example.

For example, the prediction device 100 extracts the training content or training period desired by the user 20 from the conversation or message exchange between the user 20 and the training instructor, and acquires the extracted training content or training period as third information. You may Specifically, the prediction device 100 performs speech recognition of a conversation between the user 20 and the training instructor, or acquires text data of a message, performs morphological analysis of the data, and performs training content desired by the user 20. Get duration.

For example, the prediction device 100 recognizes the content uttered by the user 20, such as "I want to complete the training within one hour", "I want to finish the training as soon as possible", or "I want to do the training without an assistant". is acquired by voice recognition. Then, the prediction device 100 may generate training content and a training period based on the acquired information. As a result, the prediction device 100 can generate training content or the like that meets the desires of the user 20 without the user 20 performing an input operation on the user terminal 10 .

In this way, such processing can be realized through interactive processing between the user terminal 10 and the prediction device 100 . That is, the prediction device 100 acquires the third information via the user terminal 10 displaying the predicted future walking function of the user 20 . Furthermore, the prediction device 100 generates new training content for the user 20 based on the third information, and information about the future walking function of the user 20 newly predicted based on the newly generated training content. is displayed on the user terminal 10 . That is, when the training content is displayed on the user terminal 10, the prediction device 100 may acquire the third information by recognizing the voice of the content uttered by the user 20 viewing the display. The third information may be obtained via user 20 input at 10 . When the prediction device 100 obtains the third information, the prediction device 100 displays on the user terminal 10 training content and prediction results reflecting the third information. In this way, according to the prediction device 100, the content of training can be determined in such a manner that the user 20 and the prediction device 100 interact with each other.

(3. Other embodiments)
The processes according to the above-described embodiments may be implemented in various different forms other than the above-described embodiments.

For example, among the processes described in the above embodiments, all or part of the processes described as being automatically performed can be manually performed, or the processes described as being performed manually can be performed manually. All or part of this can also be done automatically by known methods. In addition, information including processing procedures, specific names, various data and parameters shown in the above documents and drawings can be arbitrarily changed unless otherwise specified. For example, the various information shown in each drawing is not limited to the illustrated information.

Also, each component of each device illustrated is functionally conceptual, and does not necessarily need to be physically configured as illustrated. In other words, the specific form of distribution and integration of each device is not limited to the one shown in the figure, and all or part of them can be functionally or physically distributed and integrated in arbitrary units according to various loads and usage conditions. Can be integrated and configured.

Also, the above-described embodiments and modifications can be appropriately combined within a range that does not contradict the processing content.

Also, the effects described in this specification are only examples and are not limited, and other effects may be provided.

(4. Effect of Prediction Device According to Present Disclosure)
As described above, the prediction device (the prediction device 100 in the embodiment) according to the present disclosure includes an acquisition unit (the acquisition unit 131 in the embodiment), a generation unit (the generation unit 133 in the embodiment), a prediction unit (the The form includes a prediction unit 134). An acquisition part acquires the 1st information containing the information regarding walking motion of a walking trainee, and the 2nd information containing the information regarding a walking trainee's body and symptom. The generation unit combines the first information and the second information acquired by the acquisition unit with information of other walking trainees in walking training that has been performed in the past, and the content and results of walking training that has been performed in the past. are input to the learned model, and based on the results output from the model, training content for the walking trainee is generated. The prediction unit predicts the future walking function of the walking trainee based on the training content generated by the generation unit.

In this way, the prediction device according to the present disclosure predicts the future walking function before training is performed, so that it is possible to support the walking trainee to actively participate in training. Therefore, it is possible to improve training results. can. That is, according to the prediction device, there is an effect that the walking trainee can enthusiastically work on the training.

Moreover, a prediction part predicts the recovery content of a walking trainee's walking function, a recovery prediction period, or the period until newly implementing different training contents as a walking function in the future of a walking trainee.

In this way, the prediction device predicts the recovery content, the recovery prediction period, and the period until the next stage of training, so that the walking trainee can grasp the training period and motivate the training. .

In addition, the generation unit uses information of other walking trainees in walking training performed in the past as learning data, and the content and result of walking training performed in the past or the walking determined by the training instructor based on the learning data The training content for the walking trainee is generated using a model that has learned the data set with the training content and the prediction results as correct data.

In this way, the prediction device can generate a more accurate model by learning using a learning data set incorporating knowledge determined by an expert such as a training instructor.

Moreover, an acquisition part acquires the training content or training period which a walking trainee desires as 3rd information. The generation unit generates training content for the walking trainee based on the third information acquired by the acquisition unit.

In this way, the prediction device can construct an environment in which the walking trainee can actively work on treatment and rehabilitation by generating training content that incorporates the desire of the walking trainee.

(5. Hardware configuration)
Information devices such as the user terminal 10 and the prediction device 100 according to the above-described embodiments are implemented by a computer 1000 configured as shown in FIG. 10, for example. Hereinafter, the prediction device 100 according to the embodiment will be described as an example. FIG. 10 is a hardware configuration diagram showing an example of a computer 1000 that implements the functions of the prediction device 100. As shown in FIG. The computer 1000 has a CPU 1100 , a RAM 1200 , a ROM (Read Only Memory) 1300 , a HDD (Hard Disk Drive) 1400 , a communication interface 1500 and an input/output interface 1600 . Each part of computer 1000 is connected by bus 1050 .

The CPU 1100 operates based on programs stored in the ROM 1300 or HDD 1400 and controls each section. For example, the CPU 1100 loads programs stored in the ROM 1300 or HDD 1400 into the RAM 1200 and executes processes corresponding to various programs.

The ROM 1300 stores a boot program such as a BIOS (Basic Input Output System) executed by the CPU 1100 when the computer 1000 is started, a program depending on the hardware of the computer 1000, and the like.

The HDD 1400 is a computer-readable recording medium that non-temporarily records programs executed by the CPU 1100 and data used by the programs. Specifically, HDD 1400 is a recording medium that records an information processing program according to the present disclosure, which is an example of program data 1450 .

Communication interface 1500 is an interface for connecting computer 1000 to an external network 1550 (for example, the Internet). For example, CPU 1100 receives data from another device via communication interface 1500, and transmits data generated by CPU 1100 to another device.

Input/output interface 1600 is an interface for connecting input/output device 1650 and computer 1000 . For example, the CPU 1100 receives data from input devices such as a keyboard and mouse via the input/output interface 1600 . The CPU 1100 also transmits data to an output device such as a display, speaker, or printer via the input/output interface 1600 . Also, the input/output interface 1600 may function as a media interface for reading a program or the like recorded on a predetermined recording medium. Media include, for example, optical recording media such as DVDs (Digital Versatile Discs) and PDs (Phase change rewritable discs), magneto-optical recording media such as MOs (Magneto-Optical disks), tape media, magnetic recording media, semiconductor memories, and the like. is.

For example, when the computer 1000 functions as the prediction device 100 according to the embodiment, the CPU 1100 of the computer 1000 implements the functions of the control unit 130 and the like by executing an information processing program loaded on the RAM 1200 . The HDD 1400 also stores an information processing program according to the present disclosure and data in the storage unit 120 . Although CPU 1100 reads and executes program data 1450 from HDD 1400 , as another example, these programs may be obtained from another device via external network 1550 .

As described above, the embodiments of the present application have been described in detail with reference to the drawings. It is possible to implement the invention in other forms.

1 prediction system 10 user terminal 30 sensor 100 prediction device 110 communication unit 120 storage unit 121 learning data storage unit 122 model storage unit 123 prediction result storage unit 130 control unit 131 acquisition unit 132 learning unit 133 generation unit 134 prediction unit

Claims (6)

An acquisition unit that acquires first information including information on the walking motion of the walking trainee and second information including information on the body and symptoms of the walking trainee,
Correlation between the first information and the second information acquired by the acquisition unit, information of other walking trainees in walking training performed in the past, and the content and result of the walking training performed in the past A generation unit that inputs sex into a learned model and generates training content for the walking trainee based on the results output from the model;
A prediction unit that predicts the future walking function of the walking trainee based on the training content generated by the generation unit;
A prediction device comprising: The prediction unit
As the walking function in the future of the walking trainee, predict at least one of the recovery content of the walking trainee's walking function, the recovery prediction period, and the period until a different training content is newly implemented,
The prediction device according to claim 1, characterized by: The generating unit
Using the information of other walking trainees in the walking training performed in the past as learning data, the content and results of the walking training performed in the past, or the walking training content determined by the training instructor based on the learning data and generating training content for the walking trainee using the model that has learned a data set with the prediction result as correct data,
The prediction device according to claim 1 or 2, characterized by: The acquisition unit
As the third information, the training content or training period desired by the walking trainee is acquired,
The generating unit
generating training content for the walking trainee based on the third information acquired by the acquisition unit;
The prediction device according to any one of claims 1 to 3, characterized in that: A computer-implemented prediction method comprising:
Acquisition step of acquiring first information including information on walking motion of the walking trainee and second information including information on the body and symptoms of the walking trainee;
Correlation between the first information and the second information acquired by the acquisition step, the information of other walking trainees in walking training performed in the past, and the content and results of the walking training performed in the past A generation step of inputting sex into a learned model and generating training content for the walking trainee based on the results output from the model;
A prediction step of predicting the future walking function of the walking trainee based on the training content generated by the generating step;
A prediction method characterized by including Acquisition procedure for acquiring first information including information on walking motion of the walking trainee and second information including information on the body and symptoms of the walking trainee;
Correlation between the first information and the second information acquired by the acquisition procedure, the information of other walking trainees in walking training performed in the past, and the content and results of the walking training performed in the past A generation procedure for inputting sex into a learned model and generating training content for the walking trainee based on the results output from the model;
A prediction procedure for predicting the future walking function of the walking trainee based on the training content generated by the generation procedure;
A prediction program characterized by causing a computer to execute

Images