Delivery system, delivery control method, and storage medium

Abstract

The present application provides a delivery system, a delivery control method, a learned model, a learning system, a learning method, and a storage medium. The delivery system includes: storing a learned model that has been subjected to machine learning using learning data, the learning data including recovery actual result data indicating a recovery actual result including a use end timing indicating that use of a device has ended after the device has been lent out and a recovery completion timing indicating that the device has been recovered as a returned item, and recovery route data indicating a recovery route in which the device has been recovered by a mobile robot, the learned model having been subjected to machine learning in a manner of inputting an end timing prediction result and outputting the recovery route. The delivery system inputs the end timing prediction result to the learned model to obtain a recovery route in which the device that is in the process of being lent out is recovered as a returned item by the mobile robot, and decides the mobile robot that performs recovery in accordance with the obtained recovery route.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to a delivery system, a delivery control method, a learned model, a learning system, a learning method, and a storage medium. BACKGROUND

[0002] In Japanese Patent Application Publication No. 2021-140273, an information processing apparatus that adjusts a logistics base in accordance with demand prediction is disclosed. The information processing apparatus creates commodity demand information that indicates a trend of demand for a commodity for each region based on a location of an action performed by each user related to the commodity and the number of actions performed at the location, and determines a logistics base for the commodity based on the commodity demand information. Also, the information processing apparatus creates a delivery plan that delivers the commodity to the logistics base for the commodity in advance based on the commodity demand information. Furthermore, the information processing apparatus creates an inventory delivery plan that transfers inventory of the commodity from other logistics bases that have inventory of the commodity to the logistics base for the commodity in advance based on the commodity demand information. SUMMARY

[0003] However, in a device lending system that lends out devices, a situation in which inventory becomes insufficient can occur in cases in which the demand for lending out devices increases sharply, and the like. As one of the main reasons for which inventory of lent-out devices becomes insufficient, or as a problem in the management of lent-out devices, a case in which a return delivery is judged to be implemented by a worker of a lending target even though use of the device at the lending target has ended can be cited. As a reason for which the judgment by the worker of the lending target is implemented, if a case in which the device is a medical device is taken as an example, there are cases in which, depending on the ward, a situation in which a worker is insufficient or an emergency exists, and the like, and thus a situation in which loading can be implemented immediately and a situation in which implementation is impossible exist.

[0004] Therefore, it is desirable to shorten such a lag time as much as possible. In particular, in a case in which a mobile robot is used to collect a returned item, it is desirable to suppress deterioration of the mobile robot and achieve power saving as much as possible. Also, in the technology described in Japanese Patent Application Publication No. 2021-140273, these problems cannot be solved even if it is assumed that lent-out devices are applied instead of commodities.

[0005] The present disclosure was completed in order to solve such problems, and provides a delivery system, a delivery control method, a learned model, and a storage medium that can effectively suppress a lag time from the end of use until return completion implemented by a mobile robot for a device that becomes a lending target in a device lending system, and provides a learning system, a learning method, and a storage medium that can generate such a learned model.

[0006] The conveyance system according to the present disclosure conveys an apparatus that is a loan target in an apparatus loan system using a mobile robot, and includes:

[0007] a learned model that is machine-learned using learning data containing recovery actual result data indicating a recovery actual result including a use end timing at which use of the apparatus ends after the apparatus is loaned and a recovery completion timing at which the apparatus is recovered as a return, and recovery route data indicating a recovery route in which the apparatus is recovered by the mobile robot, and that is machine-learned in a manner that inputs an end timing prediction result that is a result of predicting an end timing of use of the apparatus that is loaned and outputs a recovery route in which the apparatus that is loaned is recovered as a return by the mobile robot is stored;

[0008] the end timing prediction result that is a result of predicting an end timing of use of the apparatus that is loaned is input to the learned model to obtain a recovery route in which the apparatus that is loaned is recovered as a return by the mobile robot; and

[0009] the mobile robot that performs recovery in accordance with the obtained recovery route is determined.

[0010] In the conveyance system described above, by this configuration, a recovery route that takes into account past recovery actual result data in accordance with a prediction result of an end timing of use is obtained for the loaned apparatus, and a mobile robot that is a recovery subject is determined. Therefore, in the conveyance system described above, the apparatus can be recovered using an efficient recovery route, and as a result, the residence time of the apparatus from the end of use to the completion of return by the mobile robot can be effectively suppressed.

[0011] It can also be configured such that the learned model is a model that is machine-learned in a manner that outputs such a recovery route in which a plurality of the apparatuses can be recovered.

[0012] Thus, in the conveyance system described above, since an efficient recovery route in which a plurality of apparatuses can be recovered can be obtained, the residence time of the plurality of apparatuses from the end of use to the completion of return by the mobile robot can be more effectively suppressed.

[0013] It can also be configured such that the recovery actual result data contains first information that is at least one of a time required for recovery by the mobile robot, a movement distance of the mobile robot, and a consumed power of the mobile robot.

[0014] The learned model is a model that has been subjected to machine learning so as to output the recovery route that minimizes the first information.

[0015] In the above-described conveyance system, by this structure, the recovery route that can recover a plurality of devices is obtained by taking into account the past recovery actual result data including the first information in accordance with the prediction result of the use end timing with respect to the borrowed device, and the mobile robot that becomes the recovery subject is decided. Therefore, in the above-described conveyance system, a plurality of devices can be recovered by the recovery route that can be said to be more efficient in terms of at least one of time, movement distance, and consumed electric power, as a result, the residence time of the plurality of devices from the use end to the completion of the return by the mobile robot can be effectively suppressed in the above-described terms.

[0016] It can also be adopted in such a manner that the recovery actual result data includes first information that is at least one of the time required for the mobile robot to recover, the movement distance of the mobile robot, and the consumed electric power of the mobile robot,

[0017] The learned model is a model that has been subjected to machine learning in such a manner that, in a case where the recoverable time at the recovery location with respect to a plurality of the devices is within a predetermined time, the recovery route that recovers a plurality of the devices in such a manner that the first information is minimized is output.

[0018] In the above-described conveyance system, by this structure, the recovery route that can recover a plurality of devices is obtained by taking into account the past recovery actual result data including the first information in accordance with the prediction result of the use end timing with respect to the borrowed device, and the mobile robot that becomes the recovery subject is decided. Therefore, in the above-described conveyance system, a plurality of devices can be recovered by the recovery route that can be said to be more efficient in terms of at least one of time, movement distance, and consumed electric power, as a result, the residence time of the plurality of devices from the use end to the completion of the return by the mobile robot can be effectively suppressed in the above-described terms.

[0019] It can also be adopted in such a manner that the device is a medical device.

[0020] Therefore, in the above-described conveyance system, the residence time of the medical device from the use end to the completion of the return by the mobile robot can be effectively suppressed while taking into account the use mode of the medical device.

[0021] A delivery control method related to the present disclosure is a delivery control method for delivering, by a mobile robot, an apparatus that becomes a lending target in an apparatus lending system, and includes:

[0022] The computer stores a learned model that has been subjected to machine learning using learning data containing recovery actual result data indicating a recovery actual result including a use end timing at which use of the apparatus ends after the apparatus is lent out and a recovery completion timing at which recovery as a returned product is performed, and recovery route data indicating a recovery route by which the apparatus is recovered by the mobile robot, and the learned model has been subjected to machine learning in a manner in which an end timing prediction result that is a result of predicting an end timing of use of the apparatus that is in a state of being lent out is input and a recovery route by which the apparatus that is in a state of being lent out is recovered as a returned product by the mobile robot is output;

[0023] The computer inputs, to the learned model, an end timing prediction result that is a result of predicting an end timing of use of the apparatus that is in a state of being lent out, to obtain a recovery route by which the apparatus that is in a state of being lent out is recovered as a returned product by the mobile robot; and

[0024] The computer decides the mobile robot that performs recovery in accordance with the obtained recovery route.

[0025] In the above-described delivery control method, by this processing, a recovery route that takes into account past recovery actual result data in accordance with a prediction result of an end timing of use is obtained for the lent-out apparatus, and the mobile robot that becomes the subject of recovery is decided. Therefore, in the above-described delivery control method, it is possible to control in a manner in which the apparatus is recovered by an efficient recovery route, and as a result, it is possible to effectively suppress the residence time of the apparatus from the end of use until the completion of return by the mobile robot.

[0026] It is also possible to adopt a manner in which the learned model is a model that has been subjected to machine learning in a manner in which the recovery route by which a plurality of the apparatuses can be recovered is output.

[0027] Thus, in the above-described delivery control method, since an efficient recovery route by which a plurality of apparatuses can be recovered can be obtained, it is possible to more effectively suppress the residence time of the plurality of apparatuses from the end of use until the completion of return by the mobile robot.

[0028] The recovery actual result data can include first information, which is at least one of a time required for the mobile robot to perform recovery, a movement distance of the mobile robot, and consumed power of the mobile robot.

[0029] The learned model is a model that has been subjected to machine learning so as to output the recovery route that minimizes the first information.

[0030] In the conveyance control method described above, by this processing, the recovery route that takes into account past recovery actual result data including the first information, based on a prediction result of the use end timing, is obtained for the borrowed device, and the mobile robot that becomes the recovery subject is decided. Therefore, in the conveyance control method described above, it is possible to control the recovery of the device in a manner that uses a recovery route that can be said to be more efficient from the viewpoint of at least one of time, movement distance, and consumed power, and as a result, it is possible to effectively suppress the residence time of the device from the use end to the completion of the return by the mobile robot from the viewpoint described above.

[0031] The recovery actual result data can include first information, which is at least one of a time required for the mobile robot to perform recovery, a movement distance of the mobile robot, and consumed power of the mobile robot.

[0032] The learned model is a model that has been subjected to machine learning so as to output the recovery route that minimizes the first information, in a case where the recoverable time at the recovery location related to the plurality of devices is within a predetermined time.

[0033] In the conveyance control method described above, by this processing, the recovery route that takes into account past recovery actual result data including the first information, based on a prediction result of the use end timing, is obtained for the borrowed device, and the mobile robot that becomes the recovery subject is decided. Therefore, in the conveyance control method described above, it is possible to control the recovery of the device in a manner that uses a recovery route that can be said to be more efficient from the viewpoint of at least one of time, movement distance, and consumed power, and as a result, it is possible to effectively suppress the residence time of the device from the use end to the completion of the return by the mobile robot from the viewpoint described above.

[0034] The device can be a medical device.

[0035] Thus, in the above-described conveyance control method, it is possible to effectively suppress the residence time of the medical equipment from the end of use to the completion of return by the mobile robot, taking into account the usage pattern of the medical equipment.

[0036] In the storage medium according to the present disclosure, the program is a program for causing a computer to execute conveyance control, wherein the conveyance control includes:

[0037] The end timing prediction result, which is a result of predicting the end timing of use of the device in loan, is input to a learned model that has been subjected to machine learning using learning data, to obtain a collection route for collecting the device in loan as a return item using the mobile robot, wherein the learning data includes collection actual result data indicating a collection actual result including an end timing of use of the device after the device is loaned and a collection completion timing at which the device has been collected as a return item, and collection route data indicating a collection route for collecting the device using the mobile robot, and the learned model has been subjected to machine learning in a manner of inputting the end timing prediction result, which is a result of predicting the end timing of use of the device in loan, and outputting the collection route for collecting the device in loan as a return item using the mobile robot; and

[0038] The mobile robot that collects in accordance with the obtained collection route is determined.

[0039] The above-described program obtains, for the device in loan, a collection route obtained by taking into account the past collection actual result data according to the prediction result of the end timing of use, and determines the mobile robot that becomes the subject of collection. Thus, in the above-described program, it is possible to control the collection of the device using an efficient collection route, and as a result, it is possible to effectively suppress the residence time of the device from the end of use to the completion of return by the mobile robot.

[0040] It is also possible to adopt a manner in which the learned model is a model that has been subjected to machine learning in a manner of outputting the collection route that enables collection of a plurality of devices.

[0041] Thus, in the above-described program, since it is possible to obtain an efficient collection route for collecting a plurality of devices, it is possible to more effectively suppress the residence time of the plurality of devices from the end of use to the completion of return by the mobile robot.

[0042] The recovery actual result data can include first information, which is at least one of a time required for the mobile robot to perform recovery, a movement distance of the mobile robot, and consumed power of the mobile robot.

[0043] The learned model is a model that has been subjected to machine learning so as to output the recovery route that minimizes the first information.

[0044] In the above procedure, by this processing, the recovery route that takes into account past recovery actual result data including the above first information in accordance with a prediction result of the use end timing is acquired for the borrowed device, and the mobile robot that becomes the recovery subject is decided. Therefore, in the above procedure, it is possible to control the recovery of the device in a manner that uses a recovery route that can be said to be more efficient from the viewpoint of at least one of time, movement distance, and consumed power, and as a result, it is possible to effectively suppress the residence time of the device from the use end to the completion of the return by the mobile robot from the above viewpoint.

[0045] The recovery actual result data can include first information, which is at least one of a time required for the mobile robot to perform recovery, a movement distance of the mobile robot, and consumed power of the mobile robot.

[0046] The learned model is a model that has been subjected to machine learning so as to output the recovery route that minimizes the first information.

[0047] In the above procedure, by this processing, the recovery route that takes into account past recovery actual result data including the above first information in accordance with a prediction result of the use end timing is acquired for the borrowed device, and the mobile robot that becomes the recovery subject is decided. Therefore, in the above procedure, it is possible to control the recovery of the device in a manner that uses a recovery route that can be said to be more efficient from the viewpoint of at least one of time, movement distance, and consumed power, and as a result, it is possible to effectively suppress the residence time of the device from the use end to the completion of the return by the mobile robot from the above viewpoint.

[0048] The device can be a medical device.

[0049] Thus, in the above procedure, the staying time of the medical device from the use end time to the return completion by the mobile robot can be effectively suppressed while taking into consideration the use mode of the medical device.

[0050] The learned model related to the present disclosure includes:

[0051] The learned model related to the present disclosure includes:

[0052] The learned model related to the present disclosure includes:

[0053] The learned model related to the present disclosure includes:

[0054] The learned model related to the present disclosure includes:

[0055] The learning method according to the present disclosure includes generating a learned model by inputting learning data into an unlearned learning model and performing machine learning, the learning data including recovery actual result data indicating a recovery actual result including a use end timing at which use of a device that becomes a lending target in a device lending system has ended after lending of the device and a recovery completion timing at which recovery has been performed as a returned product by a mobile robot, and recovery route data indicating a recovery route in which the device has been recovered by the mobile robot, the learned model being a model that inputs an end timing prediction result that is a result of predicting an end timing of use of the device that is in lending and outputs a recovery route in which the device that is in lending is recovered as a returned product by the mobile robot.

[0056] The learning method according to the present disclosure includes generating a learned model by inputting learning data into an unlearned learning model and performing machine learning, the learning data including recovery actual result data indicating a recovery actual result including a use end timing at which use of a device that becomes a lending target in a device lending system has ended after lending of the device and a recovery completion timing at which recovery has been performed as a returned product by a mobile robot, and recovery route data indicating a recovery route in which the device has been recovered by the mobile robot, the learned model being a model that inputs an end timing prediction result that is a result of predicting an end timing of use of the device that is in lending and outputs a recovery route in which the device that is in lending is recovered as a returned product by the mobile robot.

[0057] The storage medium according to the present disclosure includes a program that causes a computer to execute a learning process, the learning process including generating a learned model by inputting learning data into an unlearned learning model and performing machine learning, the learning data including recovery actual result data indicating a recovery actual result including a use end timing at which use of a device that becomes a lending target in a device lending system has ended after lending of the device and a recovery completion timing at which recovery has been performed as a returned product by a mobile robot, and recovery route data indicating a recovery route in which the device has been recovered by the mobile robot, the learned model being a model that inputs an end timing prediction result that is a result of predicting an end timing of use of the device that is in lending and outputs a recovery route in which the device that is in lending is recovered as a returned product by the mobile robot.

[0058] The storage medium according to the present disclosure includes a program that causes a computer to execute a learning process, the learning process including generating a learned model by inputting learning data into an unlearned learning model and performing machine learning, the learning data including recovery actual result data indicating a recovery actual result including a use end timing at which use of a device that becomes a lending target in a device lending system has ended after lending of the device and a recovery completion timing at which recovery has been performed as a returned product by a mobile robot, and recovery route data indicating a recovery route in which the device has been recovered by the mobile robot, the learned model being a model that inputs an end timing prediction result that is a result of predicting an end timing of use of the device that is in lending and outputs a recovery route in which the device that is in lending is recovered as a returned product by the mobile robot.

[0059] According to the present disclosure, it is possible to provide a delivery system, a delivery control method, a learned model, and a storage medium that can effectively suppress a stay time from an end of use to completion of return by a mobile robot for a device that becomes a lending target in a device lending system, and a learning system, a learning method, and a storage medium that can generate such a learned model. BRIEF DESCRIPTION OF DRAWINGS

[0060] The features, advantages, technical and industrial significance of representative embodiments of the present invention will be depicted in the following drawings for reference, wherein the same symbols denote the same elements.

[0061] Figure 1 This is a conceptual diagram used to illustrate an example of the overall structure of the conveying system involved in this embodiment.

[0062] Figure 2 This is a control block diagram illustrating an example of the conveying system involved in this embodiment.

[0063] Figure 3 To indicate Figure 2 Here is a control block diagram of an example device lending system.

[0064] Figure 4 To indicate Figure 2 A control block diagram of an example electronic medical record system.

[0065] Figure 5 To indicate Figure 4 An example table of electronic medical record information stored in an electronic medical record system.

[0066] Figure 6 To indicate Figure 3 An example table showing the equipment lending information and temporary reservation information stored in the equipment lending system.

[0067] Figure 7 To indicate Figure 2 An example table showing the transport information stored in the higher-level management device.

[0068] Figure 8 A diagram illustrating an example of a mobile robot's movement path.

[0069] Figure 9 A diagram illustrating other examples of the movement paths of a mobile robot.

[0070] Figure 10 For use in Figure 2 A schematic diagram illustrating an example of transport processing in a higher-level management device.

[0071] Figure 11 To indicate in Figure 10 A diagram illustrating an example of a recycling route taken during the transport process.

[0072] Figure 12 This is a flowchart illustrating an example of the conveying method involved in this embodiment.

[0073] Figure 13 To indicate that the generation is byFigure 2 a block diagram of one configuration example of a learning system of a learned model utilized in a higher-level management device.

[0074] Figure 14 to show one example of a learned model generated by the learning system. Figure 13 DETAILED DESCRIPTION

[0075] While the present application is hereinafter described by way of embodiments of the application, the application as claimed in the claims is not limited to the embodiments described below. Furthermore, the structures described in the embodiments are not necessarily all essential to the method for solving the problem.

[0076] EMBODIMENT

[0077] OVERALL STRUCTURE

[0078] The conveyance system according to the present embodiment is a system that conveys equipment that is an object of lending in an equipment lending system using a mobile robot, and acquires a recovery route of the lent equipment using a learned model. The learned model, the details of which will be described later, is a model that has been subjected to machine learning in such a manner that a prediction result of an end timing of input is predicted and a recovery route is output using learning data including recovery actual result data and recovery route data.

[0079] Furthermore, in the conveyance system, a mobile robot that implements recovery in accordance with the acquired recovery route is decided. Thereafter, the conveyance system controls the decided mobile robot, and thus the recovery of the equipment can be implemented in accordance with the recovery route.

[0080] In the conveyance system, a recovery route that takes into account past recovery actual result data in accordance with a prediction result of an end timing of use is acquired for the lent equipment, and a mobile robot that becomes a subject of recovery is decided. Therefore, in the conveyance system, the equipment can be recovered in accordance with an efficient recovery route, and as a result, the residence time of the equipment from the end of use to the completion of the return by the mobile robot can be effectively suppressed.

[0081] First, one example of the conveyance system according to the present embodiment will be described. Figure 1 A conceptual diagram for describing an overall configuration example of the conveyance system 1 according to the present embodiment. The conveyance system 1 according to the present embodiment is a system that conveys a conveyance object using a mobile robot that can autonomously move. As the mobile robot, a mobile robot 20 as illustrated in FIG. 1 is exemplified, but the structure and shape of the mobile robot 20 are not limited thereto. Figure 1 The mobile robot 20 as illustrated in FIG. 1 is exemplified, but the structure and shape of the mobile robot 20 are not limited thereto. ​

[0082] The transport system 1 has, in addition to the mobile robot 20, a higher-level management device 10, a medical equipment lending system (hereinafter, equipment lending system) 30, an electronic medical record system 40, a network 600, a communication unit 610, and a user terminal 400.

[0083] The mobile robot 20 is a transport robot that performs transport of a transport object as a task. The mobile robot 20 autonomously travels in a medical welfare facility such as a hospital, a rehabilitation center, a nursing facility, a facility for the elderly, and the like, in order to transport a transport object. The mobile robot 20 can be configured to autonomously move with reference to a map. Further, the mobile robot 20 can be configured to autonomously move with reference to a region that is set in advance as a part or all of the above-described map, or a region represented by latitude and longitude, or the like. However, the mobile robot 20 can be configured to autonomously move while sensing the surroundings, even outside the region set in advance, or outside all of the regions included in the map at the beginning, or in a manner in which a range of movement is not set.

[0084] A user U1 such as a user or an assistant of a transport object, a manager of a transport object, or the like entrusts the mobile robot 20 with transport of a transport object. The user U1, when making a transport entrustment, houses the transport object in the mobile robot 20 at an entrustment location, or at a reception destination (transport source location) included in information of the transport entrustment. Of course, the housing of the transport object can be performed by a robot for housing, or the like. In addition, although a mobile robot that carries and transports a transport object in a state in which the transport object is exposed can be used, the description is simplified on the premise that the transport object is transported in a state in which the transport object is housed in the mobile robot 20.

[0085] In the present embodiment, it is only necessary to be able to transport an equipment that is a lending object (hereinafter, lending equipment) as a transport object. However, in the mobile robot 20, an equipment or a transport object other than the lending equipment, such as a consumable such as a medicine, a bandage, or the like, a specimen, a hospital meal, stationery, or the like, can be transported.

[0086] The user U1 can entrust transport of the lending equipment in accordance with a lending schedule (lending schedule) of the user U1. Although this will be described later, the lending schedule can be managed by the equipment lending system 30, and can be referred to from the user terminal 400 by the user U1 in order to make a transport entrustment, and can be referred to from the higher-level management device 10.

[0087] The mobile robot 20 autonomously moves to a set destination and delivers the loaned equipment. That is, the mobile robot 20 performs a delivery task of baggage (hereinafter, also simply referred to as a task). In the following description, a place where the mobile robot 20 mounts the loaned equipment is set as a delivery source, and a place where the mobile robot 20 delivers the loaned equipment is set as a delivery destination.

[0088] For example, the mobile robot 20 is set to move within a general hospital having a plurality of clinical departments. The mobile robot 20 delivers the loaned equipment between the plurality of clinical departments. For example, the mobile robot 20 delivers the loaned equipment from a nurse station of a certain clinical department to a nurse station of another clinical department. Alternatively, the mobile robot 20 delivers the loaned equipment from a storage thereof to a nurse station of a clinical department. Further, in a case where the delivery destination is located on a different floor, the mobile robot 20 can also move using an elevator or the like. Further, the mobile robot 20 also takes charge of the return of the loaned equipment to the storage or the like.

[0089] As examples of the loaned equipment, medical equipment such as an examination instrument, a medical instrument, and the like can be listed. As the medical equipment, a pressure sore prevention device, a sphygmomanometer, a transfusion pump, a drip mechanical instrument such as a syringe pump, a pedal pump, a nurse call bell, a bed exit sensor, a low-pressure continuous inhaler, an electrocardiograph monitor, a drug injection controller, an enteral feeding pump, a respirator, a cuff pressure gauge, a touch sensor, an aspirator, a nebulizer, a pulse oximeter, a sphygmomanometer, a manual resuscitator, a sterile device, an echo device, and the like can be listed. Further, in addition to these, various mechanical instruments, various life monitors, and the like can be listed as the medical equipment. In addition, there are cases where various types of medical equipment are set as the loaned objects in a plurality of models, such as a case where a transfusion pump having a different flow rate is also an object of loan, and the like.

[0090] Further, among the loaned equipment, there are also devices having a stand provided on the device itself. For example, as such loaned equipment with a stand, a low-pressure continuous inhaler, an echo device, an electrocardiograph monitor (transmitter), an electrocardiograph monitor (central monitor), an electrocardiograph monitor (bedside monitor), a respirator, a nebulizer, and the like can be listed. The loaned equipment with a stand is mostly a device that is connected to a commercial power supply and not a battery to operate, and is more often stored in a loaned warehouse as a storage place than the loaned equipment without a stand.

[0091] In addition, such a lending device as described above is often used without sterilization of the entire device or with only a part of the device disinfected, and there are devices in which a disposable tool is attached to the lending device. In the case where the implementation storage site of a catheter, a scalpel, scissors, or the like, which needs to be sterilized, and the sterilization site are the same or close to each other, these can be handled as the lending device in the present embodiment.

[0092] In the present embodiment, as shown in FIG. 1, a device lending system 30, an electronic medical record system 40, a mobile robot 20, and a user terminal 400 are connected to a higher-level management apparatus 10 via a network 600. The mobile robot 20 and the user terminal 400 are connected to the network 600 via a communication unit 610. The network 600 is a wired or wireless Local Area Network (LAN) or a Wide Area Network (WAN). The higher-level management apparatus 10 is connected to the network 600 in a wired or wireless manner. The communication unit 610 is, for example, a wireless LAN unit provided in each environment. The communication unit 610 can also be a general-purpose communication device such as a WiFi router. Figure 1

[0093] The user terminal 400 is, for example, a tablet computer or a smartphone, but can also be a set-type computer. The user terminal 400 only needs to be an information processing apparatus capable of communicating in a wireless or wired manner.

[0094] The user U1 or the user U2 can use the user terminal 400 to implement a delivery request. For example, the user U1 can access the device lending system 30 (via the higher-level management apparatus 10) to refer to the schedule thereof in order to make a delivery request from the user terminal 400, and can make a delivery request for a lending device to the higher-level management apparatus 10 based on the result of the reference. The higher-level management apparatus 10 that has received the delivery request can make a delivery request to the mobile robot 20.

[0095] Thus, various signals transmitted from the user terminals 400 of the users U1 and U2 are once transmitted to the higher-level management apparatus 10 via the network 600, and can be forwarded from the higher-level management apparatus 10 to the mobile robot 20 that is the target. Likewise, various signals transmitted from the mobile robot 20 are once transmitted to the higher-level management apparatus 10 via the network 600, and are forwarded from the higher-level management apparatus 10 to the user terminal 400 that is the target.

[0096] ​The upper management device 10 is a server connected to each device, which collects data from each device. In addition, the upper management device 10 is not limited to a physically single device, and can have a plurality of devices that implement distributed processing. In addition, the upper management device 10 can be configured in a manner that is distributed in the edge devices such as the mobile robot 20. For example, a part or all of the conveyance system 1 can be mounted in the mobile robot 20.

[0097] The device lending system 30 is a system that manages, for each lent device, a lending schedule (management information) indicating a date and time of lending and a lending target (use place, user, and the like). The device lending system 30 can be configured as a server connected to the upper management device 10, which implements exchange of data with the upper management device 10. Thereby, the upper management device 10 can obtain the lending schedule of the lent device managed in the device lending system 30. The device lending system 30 can also be configured in a manner that is distributed in the upper management device 10, and can also be configured in a manner that is incorporated in the upper management device 10.

[0098] The electronic medical record system 40 is a system that stores and manages electronic medical record data including information related to a patient (also referred to as patient information). For example, when a medical staff such as a doctor or a nurse inputs patient information using the user terminal 400, the patient information is stored in a storage or the like of the electronic medical record system 40. Also, the medical staff can read and update the patient information stored in the electronic medical record system 40 through the user terminal 400.

[0099] The electronic medical record system 40 can be configured as a server connected to the upper management device 10, which implements exchange of data with the upper management device 10. Thereby, the upper management device 10 can obtain the electronic medical record data managed in the electronic medical record system 40. The electronic medical record system 40 can also be configured in a manner that is distributed in the upper management device 10, and can also be configured in a manner that is incorporated in the upper management device 10.

[0100] The upper management device 10 can also be configured to read a disease and a surgery schedule or the like from the electronic medical record data registered in the electronic medical record system 40, and determine a device or the like required therefor, and register lending of a lent device and other accessories or the like in the device lending system 30.

[0101] The user terminal 400 and the mobile robot 20 can also transmit and receive signals without going through the upper-level management device 10. For example, the user terminal 400 and the mobile robot 20 can also directly transmit and receive signals through wireless communication. Alternatively, the user terminal 400 and the mobile robot 20 can also transmit and receive signals via the communication unit 610.

[0102] The user U1 or the user U2 uses the user terminal 400 to entrust delivery of the lending device. Hereinafter, the case where the user U1 is a delivery entrustor located at a delivery source and the user U2 is a scheduled pickup person located at a delivery destination (destination) will be described. Of course, the delivery entrustment can also be performed by the user U2 located at the delivery destination. Furthermore, the delivery entrustment can also be performed by a user located at a place other than the delivery source or the delivery destination.

[0103] In the case where the delivery entrustment is performed by the user U1, the content of the lending device, the reception destination of the lending device (hereinafter, also referred to as the delivery source), the delivery destination of the lending device (hereinafter, also referred to as the delivery destination), the scheduled time of arrival at the delivery source (the reception time of the lending device), the scheduled time of arrival at the delivery destination (the delivery deadline), and the like are input using the user terminal 400. Hereinafter, these pieces of information will also be referred to as delivery entrustment information. In the present embodiment, in the case of a lending device that is to be the object of delivery, the delivery source sometimes becomes the storage place (device management place) of the lending device. The delivery source can also be the place where the user U1 is located. The delivery destination is the place where the user U2 or the patient scheduled to use is located. The user U1 can input the delivery entrustment information by operating the touch panel of the user terminal 400.

[0104] As for the lending device in the delivery entrustment information, the lending schedule registered in the device lending system 30 can be used for designation. For example, the user U1 designates the lending device from the user terminal 400, and mounts it in the mobile robot 20 as needed, and performs delivery entrustment to the upper-level management device 10. By the upper-level management device 10 that received the delivery entrustment referring to the device lending system 30, and deciding the delivery schedule in such a way as to catch up with the use start time indicated by the lending schedule of the lending device, and performing delivery entrustment to the mobile robot 20, delivery is performed according to the delivery schedule.

[0105] Alternatively, by the user U1 referring to the lending schedule from the user terminal 400 and performing delivery entrustment, and by the upper-level management device 10 referring to the lending schedule and deciding the delivery schedule, and performing delivery entrustment to the mobile robot 20, delivery is performed according to the delivery schedule. In addition to this, various methods of delivery entrustment can be employed.

[0106] These examples are premised on the case where the delivery instruction is sent after the lending schedule is registered based on the lending instruction (instruction of registration of lending). On the other hand, there is a case where the lending device becomes in high demand, in which case the lending schedule for the lending device at the required time is not registered. In this case, the user Ul can also send the delivery instruction from the user terminal 400 to the upper management device 10. The upper management device 10 registers in the lending schedule and implements the delivery instruction to the mobile robot 20 if there is no problem based on the delivery instruction with reference to the device lending system 30 to check whether there is duplication in the lending period. The loading of the lending device into the mobile robot 20 in this case can be implemented, for example, at the time before and after the delivery instruction from the user terminal 400 is sent.

[0107] As described above, in any case, the user terminal 400 can send the delivery instruction information input by the user Ul to the upper management device 10. The upper management device 10 is a management system that manages a plurality of mobile robots 20, and sends an action instruction for executing a delivery task to each mobile robot 20. At this time, the upper management device 10 decides the mobile robot 20 that is to execute the delivery task for each delivery instruction. Also, the upper management device 10 sends a control signal including the action instruction to the mobile robot 20. The mobile robot 20 moves from the delivery source to the delivery destination in accordance with the action instruction.

[0108] For example, the upper management device 10 assigns the delivery task to the mobile robot 20 at or near the delivery source. Alternatively, the upper management device 10 assigns the delivery task to the mobile robot 20 that is going to the delivery source or the vicinity thereof. The mobile robot 20 assigned to the task goes to the delivery source to take the lending device. The delivery source can be exemplified by the place where the user Ul who instructed the storage place or the task is located, for example.

[0109] When the mobile robot 20 arrives at the delivery source, the lending device is placed in the mobile robot 20 by the user Ul or another staff. The mobile robot 20 with the lending device mounted thereon moves autonomously with the delivery destination as the destination. The upper management device 10 sends a signal to the user terminal 400 of the user U2 at the delivery destination. Thereby, the user U2 can know that the lending device is being delivered and the scheduled arrival time thereof. When the mobile robot 20 arrives at the set delivery destination, the user U2 can take the lending device accommodated in the mobile robot 20. In this way, the mobile robot 20 executes the delivery task.

[0110] Further, in such an overall configuration described above, the conveyance system can be constructed as a whole in a manner that each element of the conveyance system is dispersed in the mobile robot 20, the user terminal 400, the device lending system 30, the electronic medical record system 40, and the upper management device 10. Further, it is also possible to construct in a manner that the essential elements for realizing the conveyance of the lent device are concentrated in one device. The upper management device 10 controls one or a plurality of mobile robots 20.

[0111] Control system of conveyance system 1

[0112] Figure 2 A control block diagram showing one example of the control system of the conveyance system 1. As shown in Figure 2 The conveyance system 1 can be provided with the upper management device 10, the mobile robot 20, the device lending system 30, the electronic medical record system 40, and the environmental camera 300.

[0113] The conveyance system 1 efficiently controls a plurality of mobile robots 20 while causing the mobile robots 20 to autonomously move within a predetermined facility. Therefore, a plurality of environmental cameras 300 are provided within the facility. For example, the environmental cameras 300 are provided at passages, corridors, elevators, entrances, and the like within the facility.

[0114] The environmental camera 300 acquires an image of a range in which the mobile robot 20 moves. In addition, in the conveyance system 1, the image acquired by the environmental camera 300 and information obtained on the basis of the image are collected by the upper management device 10. Alternatively, the image and the like acquired by the environmental camera 300 can be directly transmitted to the mobile robot. The environmental camera 300 can be a surveillance camera or the like provided at a passage or an entrance within the facility. The environmental camera 300 can be used to obtain a distribution of congestion within the facility.

[0115] In the conveyance system 1, a route plan can be implemented by the upper management device 10 on the basis of, for example, the conveyance request information, and route plan information can be generated. The route plan information can be generated as information obtained by planning a conveyance route corresponding to the conveyance schedule described above. The upper management device 10 instructs each mobile robot 20 to a destination on the basis of the generated route plan information. Further, the mobile robot 20 autonomously moves toward the destination designated by the upper management device 10. The mobile robot 20 autonomously moves toward the destination (goal) using a sensor, a floor map, position information, and the like provided in the mobile robot 20.

[0116] For example, the mobile robot 20 travels in a manner to avoid contact with devices, objects, walls, people (hereinafter collectively referred to as surrounding objects) in its surroundings. Specifically, the mobile robot 20 detects the distance to the surrounding objects. Also, the mobile robot 20 travels in a state separated by a certain distance (referred to as a distance threshold) or more from the surrounding objects. If the distance to the surrounding objects becomes below the distance threshold, the mobile robot 20 decelerates or stops. By adopting this manner, the mobile robot 20 is able to travel without contacting the surrounding objects. Since contact can be avoided, safe and efficient transport can be performed.

[0117] The upper management device 10 can be provided with an arithmetic processing section 11, a storage section 12, a buffer memory 13, and a communication section 14. The arithmetic processing section 11 implements arithmetic for controlling and managing the mobile robot 20. The arithmetic processing section 11 can be installed as a device capable of executing a program such as a central processing unit (CPU) of a computer, for example. Also, various functions can be realized by a program. Although only the characteristic end timing prediction processing section 110, the robot control section 111, and the route planning section 115 are shown in the arithmetic processing section 11 in the above embodiment, other processing modules can also be provided. Figure 2

[0118] The end timing prediction processing section 110 inputs the loaned device data representing the medical device in loan and the electronic medical record data in which information representing the necessity of use of the medical device is described, to the learned model 120 stored in the storage section 12, and acquires, from the learned model 120, an end timing prediction result as a result of predicting the end timing of use of the medical device in loan. The end timing prediction processing section 110 passes the acquired end timing prediction result to the route planning section 115. In addition, the end timing prediction processing section 110 can also be configured to notify the end timing prediction result to the device loan system 30 via the communication section 14.

[0119] Here, the information representing the necessity of use of the medical device can refer to information representing the medical device itself, information representing the surgery required for the patient, information representing the symptoms of the patient, information representing the treatment of the patient, and the like, or information obtained by combining a plurality of these.

[0120] ​The robot control section 111 performs an operation for remotely controlling the mobile robots 20, and generates a control signal. The robot control section 111 generates the control signal based on the route plan information 125 and the like described later. Also, the robot control section 111 generates the control signal based on various information obtained from the environment camera 300 and the mobile robots 20. The control signal can also contain update information of the floor map 121, the robot information 123, the robot control parameters 122, and the like described later. That is, the robot control section 111 generates a control signal corresponding to the update information when various information is updated.

[0121] The route planning section 115 performs route planning for each mobile robot 20. When a delivery task is input, the route planning section 115 performs route planning for delivery of the borrowed device to the delivery destination (destination) based on the delivery request information. Specifically, the route planning section 115 refers to the route plan information 125, the robot information 123, and the like stored in the storage section 12 to determine the mobile robot 20 that performs a new delivery task.

[0122] The departure place is the current position of the mobile robot 20, or the delivery destination of the immediately preceding delivery task, the reception destination of the borrowed device, or the like. The destination is the delivery destination of the borrowed device, but can also be a standby place, a charging place, a storage place, or the like. Here, the route planning section 115 sets a passing point from the departure place of the mobile robot 20 to the destination. The route planning section 115 sets the passing order of the passing points for each mobile robot 20. The passing points are set, for example, at a branching intersection, a crossing, a hall in front of an elevator, or the vicinity of these locations. Further, in a passage with a narrow width, there are cases where it is difficult for the mobile robots 20 to pass each other. In such a case, the vicinity of the passage with a narrow width can also be set as a passing point. Candidates for the passing points can also be registered in the floor map 121 in advance.

[0123] The route planning section 115 determines the mobile robot 20 that performs each delivery task from among the plurality of mobile robots 20 in a manner that enables the system as a whole to efficiently perform tasks. The route planning section 115 can, for example, preferentially assign a delivery task to a mobile robot 20 that is in standby or a mobile robot 20 that is close to the delivery source. Further, the route planning section 115 can perform assignment in accordance with other conditions such as uniformization of the degree of deterioration of the mobile robots 20 instead of or in addition to such preferential assignment, as described later.

[0124] The route planning unit 115 sets through points including a departure point and a destination for the mobile robot 20 to which a transport task is assigned. For example, in a case where there are two or more movement paths from the transport source to the transport destination, the through points are set in a manner that enables movement in a shorter time. Thus, the upper-level management device 10 updates information indicating the congestion situation of the passageway based on an image of a camera or the like. Specifically, the degree of congestion is high at a place where other mobile robots 20 are passing through or where there are many people. Thus, the route planning unit 115 sets the through points in a manner that avoids the place where the degree of congestion is high.

[0125] There is a case where the mobile robot 20 can move to the destination through either a left-turn movement path or a right-turn movement path. In this case, the route planning unit 115 sets the through points in a manner that passes through the movement path of the less congested side. By setting one or more through points by the route planning unit 115 between the destination, the mobile robot 20 can be caused to move on the less congested movement path. For example, in a case where a passageway branches at a branch intersection or a cross intersection, the route planning unit 115 appropriately sets the through points at the branch intersection, the cross intersection, a corner, and the periphery thereof. Thereby, the transport efficiency can be improved.

[0126] The route planning unit 115 can also set the through points in a manner that takes into account the congestion situation of an elevator and the movement distance or the like. Also, the upper-level management device 10 can estimate the number of mobile robots 20 and the number of people at a predetermined time when the mobile robot 20 passes through a certain place. Also, the route planning unit 115 can set the through points in accordance with the estimated congestion situation. Further, the route planning unit 115 can dynamically change the through points in accordance with a change in the congestion situation. The route planning unit 115 sets the through points in order for the mobile robot 20 to which a transport task is assigned. The through points can also include the transport source and the transport destination. As described later, the mobile robot 20 autonomously moves in a manner that passes through the through points set by the route planning unit 115 in order.

[0127] The route planning unit 115 can implement the determination of the mobile robot 20 and the setting of the through points in the above-described manner. The route planning unit 115 can also implement the same processing at the time of return (recycling) of the borrowed device that is being borrowed.

[0128] However, the route planning unit 115 is pre-configured to use the learned model 124 to set the retrieval route as the transport route in this case. In this case, the set retrieval route can also include transit points including a departure point and a destination point. Here, the departure point is the lending target, and the destination point is the storage location, maintenance location, or the next lending target.

[0129] The route planning unit 115 inputs an end-time prediction result, which is a result of predicting the end-of-use time of the loaned-out equipment, into the learned model 124 to obtain a recycling route for retrieving the loaned-out equipment as a return item using the mobile robot 20. The input end-time prediction result can be obtained by the end-time prediction processing unit 110 using the learned model 120 and set as the end-time prediction result transmitted to the route planning unit 115. In this way, the route planning unit 115 can automatically create a route plan for the recycling route. However, even if the learned model 124 does not output the recycling route itself, it can output only a part of the recycling route information, and the route planning unit 115 can supplement the other information.

[0130] Furthermore, the route planning unit 115 can perform a process that determines the mobile robot 20, which is the object controlled for retrieving the loaned equipment, according to the obtained retrieval route. This decision will be described later, but the route planning unit 115 can determine the mobile robot 20 based on predetermined conditions. Alternatively, the decision regarding the mobile robot 20 can also be performed by the robot control unit 111.

[0131] Storage unit 12 is a storage unit that stores information required for the management and control of mobile robots 20, etc. Figure 2 The example shows a learned model 120, a floor map 121, robot information 123, robot control parameters 122, a learned model 124, route planning information 125, and transport information 126, but the information stored in the storage unit 12 may also include other information. In the arithmetic processing unit 11, calculations using the information stored in the storage unit 12 are performed when various processes are implemented. Furthermore, the various information stored in the storage unit 12 can be updated to the latest information.

[0132] The learned model 120 is a learned model that has been subjected to machine learning using learning data (hereinafter, first learning data) containing, among others, loan actual result data representing actual results including a medical device that has been loaned out as a loaned device and actual results for which use of the medical device has ended, and electronic medical record data in which information representing necessity of use of the loaned medical device is described. Further, the learned model 120 is subjected to machine learning in a manner such that electronic medical record data in which information representing necessity of use of the medical device is described and loan-in device data representing a medical device that is in the process of being loaned out are input, and an end timing prediction result that is a prediction result of a timing at which use of the medical device ends is output. That is, the learned model 120 is a model that becomes an algorithm that predicts an end timing prediction result from the electronic medical record data and the loan-in device data. The algorithm and the like are not limited as long as such a prediction can be made. In addition, the learned model 120 and the learned model 124 described later can be updated at predetermined times when the data is used and accumulated.

[0133] Here, the lending actual result data is data indicating the lending actual result including that the use of the medical device has ended for the medical device that becomes a management object in the device lending system 30. The lending actual result data can be managed by a storage section (storage section 32 described later) of the device lending system 30. The case where the use of the medical device has ended can be obtained by the user U2 or the like based on the input from the user terminal 400, for example, as well as the case where the use of the medical device has started. In either case of the use ending or the use starting, the user terminal 400 can transmit the input result directly to the device lending system 30 or via the upper management apparatus 10 via the network 600 and record the input result together with the date and time in advance as the lending actual result data. However, the case where the use of the medical device has started or has ended can also be obtained by a method other than this. For example, in the case where the medical device is a device that receives power supply from a socket, the use start and the use end of the medical device are judged based on the consumed power detected by a sensor or the like installed on the socket at the lending place (use place). The judgment result is transmitted to the device lending system 30 directly or via the upper management apparatus 10 via the network 600. The judgment result can be recorded together with the date and time as the lending actual result data. Or, it is also possible to judge the case where the use of the medical device has started and ended, respectively, by implementing communication at the medical device side and the predetermined place as the lending place, using the approach and separation of the medical device and the predetermined place, and transmitting the judgment result to the device lending system 30 directly or via the upper management apparatus 10 via the network 600, and recording the judgment result together with the date and time in advance as the lending actual result data. The communication here can be implemented by, for example, a beacon that emits a radio wave using Bluetooth (registered trademark), Bluetooth Low Energy (registered trademark), or the like, and a device that detects the radio wave thereof, or a Radio Frequency Identification (RFID) tag such as a Nearfield communication (NFC) tag and a tag reader thereof. In addition, either of the medical device and the predetermined place can be set as the transmission side and the reception side, and further, the medical device can be built in with such a communication function in advance or externally provided with a device having such a communication function.

[0134] The learned model 124 is a learned model that has been subjected to machine learning using learning data (hereinafter, second learning data) containing recovery actual result data representing a recovery actual result including a use end timing at which the use of the lending device has ended after the lending device has been lent out and a recovery completion timing at which the recovery has been completed as a returned product, and recovery route data representing a recovery route in which the lending device has been recovered by the mobile robot 20.

[0135] Here, the use end timing can be acquired as explained for the use end of the lending actual result data. The recovery completion timing can also be acquired as an input result in the user terminal 400, or can also be acquired through communication between the medical device and the predetermined place. However, the recovery completion timing can be a recovery completion date and time, or can be set to a date and time at which the lending device is transported to the storage place or the next lending place (transport completion date and time), or to a transport start date and time at which such transport is started, for example. Therefore, in a case where the recovery completion timing is acquired through communication between the medical device and the predetermined place, the predetermined place becomes a place at which the recovery is ended (the storage place or the next lending place, or the like).

[0136] Further, the learned model 124 is set to a model that has been subjected to machine learning in a manner in which an end timing prediction result that is a result of predicting the use end timing of the lending device that is in the lending out is input, and a recovery route in which the lending device that is in the lending out is recovered as a returned product by the mobile robot 20 is output. That is, the learned model 124 is a model that becomes an algorithm for predicting the recovery route from the end timing prediction result. The algorithm and the like are not limited, as long as such prediction can be made. Here, the explanation is also made for a case where the device that becomes the object of learning and prediction in the learned model 124 is a medical device. Therefore, the end timing prediction result that is input to the learned model 124 can be set to the output result from the learned model 120 as described above.

[0137] The floor map 121 is map information of a facility in which the mobile robot 20 moves. The floor map 121 can also be created in advance. The floor map 121 can also be generated from information obtained from the mobile robot 20. Further, the floor map 121 can be a floor map obtained by adding map correction information generated from information obtained from the mobile robot 20 to a basic map created in advance.

[0138] The robot information 123 describes the ID, model, specifications, and the like of the mobile robot 20 managed by the upper management device 10. The robot information 123 can also include position information indicating the current position of the mobile robot 20. The robot information 123 can also include information indicating whether the mobile robot 20 is in task execution or in standby. Further, the robot information 123 can also include information indicating whether the mobile robot 20 is in motion or in failure, and the like. Further, the robot information 123 can also include information of a deliverable loaned device and an undeliverable loaned device.

[0139] The robot control parameter 122 describes control parameters related to the mobile robot 20 managed by the upper management device 10 and the threshold distance to the surrounding object, and the like. The threshold distance becomes a limit distance for avoiding contact with the surrounding object including a person. Also, the robot control parameter 122 can also include information related to the motion intensity such as the speed upper limit value of the moving speed of the mobile robot 20.

[0140] The robot control parameter 122 can also be updated according to the situation. The robot control parameter 122 can also include information indicating the idle condition and the use condition of the accommodation space of the mobile robot 20. The robot control parameter 122 can also include information of a deliverable loaned device and an undeliverable loaned device. Of course, the robot control parameter 122 can also include information indicating the possibility / impossibility of delivery related to the delivery object other than the loaned device. The robot control parameter 122 causes the above-described various information to be established in correspondence with each mobile robot 20.

[0141] The route plan information 125 includes route plan information planned by the route planning section 115. The route plan information 125 includes, for example, information indicating the delivery task. The route plan information 125 can also include the ID of the mobile robot 20 assigned with the task, the departure place, the content of the loaned device, the delivery destination, the delivery source place, the scheduled arrival time to the delivery destination, the scheduled arrival time to the delivery source place, the arrival deadline, and the like. In the route plan information 125, the above-described various information can also be established in correspondence with each delivery task. The route plan information 125 can also include at least a part of the delivery commission information input from the user Ul and the like with respect to both the loan delivery and the return delivery, and can include at least a part of the information included in the recovery route output from the learned model 124 with respect to the return delivery.

[0142] Here, the route plan information 125 can also include information on the passing points for each mobile robot 20 and the transport task. For example, the route plan information 125 includes information indicating the passing order of the passing points for each mobile robot 20. The route plan information 125 can also include the coordinates of each passing point in the floor map 121 and information on whether the passing point has been passed.

[0143] The transport object information 126 is information on the lending device for which the transport request has been executed. The transport object information 126 includes, for example, the content (kind) of the lending device, the transport source, the transport destination, and the like. Of course, the transport object information 126 can also include information on a transport object other than the lending device, and the same applies hereinafter to information other than the transport object information 126. The transport object information 126 can also include the ID of the mobile robot 20 in charge of the transport. Furthermore, the transport object information 126 can include information indicating the current state, such as in transport, before transport (before mounting), and transport completed, and can further include information indicating which of the transport for lending and the transport for returning. The transport object information 126 has a corresponding relationship established for each lending device with respect to these pieces of information. The transport object information 126 will be described later.

[0144] In addition, the route planning unit 115 can refer to various information stored in the storage unit 12 to make a route plan. The route planning unit 115 can determine the mobile robot 20 that executes the task, for example, based on the floor map 121, the robot information 123, the robot control parameter 122, and the route plan information 125. Furthermore, the route planning unit 115 can refer to the floor map 121 and the like to set the passing points and the passing order up to the transport destination. In the floor map 121, candidates for the passing points are registered in advance. Furthermore, the route planning unit 115 can set the passing points according to the congestion situation and the like. In addition, in the case of continuous processing of tasks and the like, the route planning unit 115 can set the transport source and the transport destination as the passing points.

[0145] Further, two or more mobile robots 20 can be assigned to one transport task. For example, in a case where a borrowed device is large compared to the transportable capacity of the mobile robots 20, the borrowed device is divided into two and carried on two mobile robots 20. Or, in a case where a borrowed device is heavy compared to the transportable weight of the mobile robots 20, the borrowed device is divided into two and carried on two mobile robots 20. By doing so, two or more mobile robots 20 can share the execution of one transport task. Of course, in a case where mobile robots 20 of different sizes are controlled, the route planning can be performed in such a way that the mobile robots 20 that can transport the borrowed device receive the borrowed device.

[0146] Further, two or more mobile robots 20 can be assigned to one transport task. For example, in a case where a borrowed device is large compared to the transportable capacity of the mobile robots 20, the borrowed device is divided into two and carried on two mobile robots 20. Or, in a case where a borrowed device is heavy compared to the transportable weight of the mobile robots 20, the borrowed device is divided into two and carried on two mobile robots 20. By doing so, two or more mobile robots 20 can share the execution of one transport task. Of course, in a case where mobile robots 20 of different sizes are controlled, the route planning can be performed in such a way that the mobile robots 20 that can transport the borrowed device receive the borrowed device.

[0147] In such a case, the storage information indicating the usage or the idle state can be updated for the storage space of the mobile robots 20. That is, the storage information indicating the idle state can be managed by the upper management device 10 and the mobile robots 20 can be controlled. For example, when the carrying or the receiving of the borrowed device is completed, the storage information is updated. When a transport task is input, the upper management device 10 refers to the storage information and causes the mobile robots 20 having an idle space to receive the borrowed device. By doing so, it becomes possible for one mobile robot 20 to execute multiple transport tasks simultaneously or for two or more mobile robots 20 to share the execution of a transport task. For example, a sensor can be provided in the storage space of the mobile robots 20 to detect the idle state. Further, the capacity and the weight of each borrowed device can be registered in advance.

[0148] The buffer storage 13 is a storage that accumulates intermediate information generated in the processing in the arithmetic processing section 11. The communication section 14 is a communication interface for communicating with the plurality of environmental cameras 300 and at least one mobile robot 20 and the like provided in the facility in which the transport system 1 is used. The communication section 14 can perform communication of both wired communication and wireless communication. For example, the communication section 14 transmits a control signal required in the control of each mobile robot 20 to the mobile robot 20 based on an instruction generated by the arithmetic processing section 11. Further, the communication section 14 can receive information collected by the mobile robot 20 and the environmental camera 300 and transmit the information to the arithmetic processing section 11. Further, the communication section 14 can receive information such as a lending schedule from the equipment lending system 30 and transmit the information to the arithmetic processing section 11, and transmit information such as a lending schedule to the equipment lending system 30 based on an instruction generated by the arithmetic processing section 11 for registration. Further, the communication section 14 can receive electronic medical record information from the electronic medical record system 40 and transmit the electronic medical record information to the arithmetic processing section 11.

[0149] The mobile robot 20 can include an arithmetic processing section 21, a storage section 22, a communication section 23, a proximity sensor (for example, a distance sensor group 24), a camera 25, a drive section 26, a display section 27, and an operation reception section 28. In addition, although only representative processing modules included in the mobile robot 20 are shown in FIG. 2, many other processing modules not shown are also included in the mobile robot 20. Figure 2

[0150] The communication section 23 is a communication interface for performing communication with the communication section 14 of the upper management device 10. The communication section 23 performs communication with the communication section 14 using, for example, wireless signals. The distance sensor group 24 is, for example, a proximity sensor that outputs proximity object distance information indicating a distance to an object or a person present in the surroundings of the mobile robot 20. The distance sensor group 24 can include, for example, front and rear distance sensors and left and right distance sensors, and can measure the distance to a surrounding object in the front and rear direction of the mobile robot 20 and the distance to a surrounding object in the left and right direction.

[0151] The camera 25, for example, photographs an image for grasping the situation of the surroundings of the mobile robot 20. The camera 25, for example, photographs the front of the traveling direction of the mobile robot 20. Further, the camera 25 can also photograph, for example, a position marker provided at the top of the facility or the like. The position marker can be used to cause the mobile robot 20 to grasp the position of the mobile robot 20.

[0152] ​The drive section 26 drives drive wheels attached to the mobile robot 20. The drive section 26 can also include an encoder or the like that detects the number of rotations of the drive wheels and their drive motors. The current position can be estimated based on the output of the encoder. The mobile robot 20 detects its current position and transmits it to the upper-level management device 10.

[0153] The display section 27 and the operation accepting section 28 are implemented by a touch panel display. The display section 27 displays a user interface screen that is the operation accepting section 28. The display section 27 can also display information indicating the whereabouts of the mobile robot 20 and the state of the mobile robot 20. The operation accepting section 28 accepts operations from a user. The operation accepting section 28 includes various switches provided on the mobile robot 20 in addition to the user interface screen displayed on the display section 27.

[0154] The arithmetic processing section 21 performs arithmetic processing for controlling the mobile robot 20. The arithmetic processing section 21 can be installed as a CPU or the like of a computer that can execute programs. Various functions can also be implemented by programs. The arithmetic processing section 21 includes a movement command extraction section 211 and a drive control section 212. Although only representative processing modules included in the arithmetic processing section 21 are shown in FIG. 2, the arithmetic processing section 21 also includes processing modules not shown in the figure. The arithmetic processing section 21 can also search for a path between points. Figure 2

[0155] The movement command extraction section 211 extracts a movement command based on a control signal given by the upper-level management device 10. For example, the movement command includes information related to the next passing point. For example, the control signal can include information related to the coordinates of the passing point and the passing order of the passing point. The movement command extraction section 211 extracts this information as a movement command.

[0156] The movement command can also include information indicating that movement to the next passing point has become possible. When the passage width is narrow, there are cases in which the mobile robot 20 cannot pass through in a staggered manner. There are also cases in which the mobile robot 20 temporarily cannot pass through the passage. In such a case, the control signal includes a command that causes the mobile robot 20 to stop at a passing point near the point at which it should stop. The upper-level management device 10 outputs a control signal that notifies the mobile robot 20 that it has become possible to move after the other mobile robot 20 has passed through or after the mobile robot 20 has become passable. This causes the mobile robot 20 that is temporarily stopped to start moving again.

[0157] ​The drive control section 212 controls the drive section 26 so that the mobile robot 20 moves based on the movement command given by the movement command extraction section 211. For example, the drive section 26 has a drive wheel that rotates according to a control command value from the drive control section 212. The movement command extraction section 211 extracts a movement command so that the mobile robot 20 moves toward a passing point received from the upper management device 10. Also, the drive section 26 rotates the drive wheel. The mobile robot 20 autonomously moves toward the next passing point. By adopting this manner, the passing points are sequentially passed through and the delivery destination is reached. In addition, the mobile robot 20 can estimate the own position and transmit a signal indicating that the passing point has been passed to the upper management device 10. Thus, the upper management device 10 can manage the current position of each mobile robot 20 and the delivery status.

[0158] Here, the drive control section 212 can perform recognition of the own position or recognition of the surrounding objects by analyzing the image data output from the camera 25 and the detection signal output from the distance sensor group 24. Also, the drive control section 212 can control the drive section 26 so that the mobile robot 20 moves based on the result and the above-mentioned movement command. At this time, the drive control section 212 can refer to the floor map 221 and the robot control parameter 222 to perform recognition of the surrounding objects or recognition of the own position.

[0159] In the storage section 22, the floor map 221, the robot control parameter 222, and the delivery object information 226 are stored. Although only a part of the information stored in the storage section 22 is shown in Figure 2 , information other than the floor map 221, the robot control parameter 222, and the delivery object information 226 shown in Figure 2 is also included. The floor map 221 is map information of a facility in which the mobile robot 20 moves. The floor map 221 is, for example, data obtained by downloading a part or all of the floor map 121 of the upper management device 10. Alternatively, the floor map 221 can be created in advance. Further, the floor map 221 can not be map information of the entire facility but map information including a region where movement is scheduled in a partial manner.

[0160] The robot control parameter 222 is a parameter for causing the mobile robot 20 to act. In the robot control parameter 222, for example, a distance threshold value for the surrounding objects is included. Also, in the robot control parameter 222, an upper limit value of the speed of the mobile robot 20 is included.

[0161] The transport information 226, like the transport information 126, includes information related to the loaned-out equipment. The transport information 226 can include information such as the type (model), origin, and destination of the loaned-out equipment. It can also include information indicating the current status, such as "in transport," "before transport (before loading)," or "transport completed." Furthermore, it can pre-include information indicating whether the transport is for lending or for return. The transport information 226 establishes a correspondence between this information and each loaned-out equipment. The transport information 226 only needs to include information related to the loaned-out equipment to be transported by the mobile robot 20. Therefore, the transport information 226 is part of the transport information 126. That is, the transport information 226 may not include information about other mobile robots 20 transporting the equipment. The transport information 126 will be described later.

[0162] The drive control unit 212 refers to the robot control parameters 222 and stops or decelerates the robot if the distance shown by the distance information obtained from the distance sensor group 24 is below a distance threshold. The drive control unit 212 controls the drive unit 26 to ensure that the mobile robot 20 travels at a speed below the speed limit. The drive control unit 212 limits the rotational speed of the drive wheels to prevent the mobile robot 20 from moving at a speed exceeding the speed limit.

[0163] Figure 3 To indicate Figure 2 A control block diagram of an example of the equipment lending system 30 of the conveying system 1. Figure 3 As shown, the device lending system 30 can include a processing unit 31, a storage unit 32, a buffer memory 33, and a communication unit 34. The processing unit 31 performs calculations for generating a schedule for lending devices and managing them. The processing unit 31 can be installed, for example, as a device capable of executing programs, such as the CPU of a computer. Moreover, various functions can be implemented through programs. Although in Figure 3 The characteristic registration unit 311 and loan planning unit 312 are shown only in the arithmetic processing unit 31, but other processing modules may also be included.

[0164] Registration unit 311 receives, for example, lending request information sent from user terminal 400 based on an operation performed by user U1, via communication unit 34, including the ID of the lending device, start time of use, end time of use, and location of use, and processes the registration.

[0165] Further, the registration section 311 receives, via the communication section 34, for example, from the user terminal 400, the lending temporary reservation information including the ID of the lending device, the use start time, the use end time, and the use place, which are transmitted in accordance with the operation performed by the user Ul, and accepts the temporary registration. Further, the registration section 311 accepts, for example, from the user terminal 400, the formal lending request or the cancel request for the accepted temporary registration, which are transmitted in accordance with the operation performed by the user Ul, by the reception via the communication section 34. However, this function of performing the temporary reservation is not essential.

[0166] The lending plan section 312 refers to the device lending information 324 indicating the lending schedule that has been planned, the temporary reservation information 325 indicating the lending schedule that has been temporarily reserved, and other lending request information or lending temporary reservation information that has been requested at the same time, based on the lending request information accepted by the registration section 311, and also takes into account the cancel request requested at the same time, to confirm that there is no overlap. Of course, in the judgment of overlap, even if it is the same kind of medical device, if another individual is not the lending target, it is treated as not overlapping. In the case where there is no overlap, the lending plan section 312 generates the lending schedule of the lending device based on the accepted lending request information, and updates the device lending information 324. In the case where the lending request information accepted by the registration section 311 overlaps in time (also overlaps in the transport time) with the existing lending schedule or the like, the lending plan section 312 performs the following reply. That is, the lending plan section 312 replies, via the communication section 34, to the transmission source of the lending request information (the user terminal 400 or the upper management device 10), the notification indicating the meaning of the overlap.

[0167] The lending plan section 312 also refers to the lending schedule that has been planned or the like, based on the lending temporary reservation information, to confirm that there is no overlap, for the lending temporary reservation information accepted by the registration section 311, as with the lending request information. In the case where there is no overlap, the lending plan section 312 generates the lending schedule of the lending device based on the accepted lending temporary reservation information, and updates the temporary reservation information 325. In addition, if a flag indicating whether the device lending information 324 and the temporary reservation information 325 are formal lending or temporary reservation, or the current state indicating this meaning is added in the case of temporary reservation, the commonality of the information can be achieved.

[0168] Further, the lending plan section 312 implements the formal registration for the formal registration request for the temporary registration accepted by the registration section 311 by moving the information of the object from the temporary reservation information 325 to the device lending information 324. Further, the lending plan section 312 implements the deletion of the temporary reservation for the cancellation request for the temporary registration accepted by the registration section 311 by deleting the information of the object from the temporary reservation information 325. As exemplified above by the registration section 311 and the lending plan section 312, the device lending system 30 can be provided with the reservation system that temporarily reserves the lending of the medical device.

[0169] The storage section 32 is a storage section that stores information necessary for the lending management of the lending device and the control of the device lending system 30. In the example, the floor map 321, the maintenance person information 322, the device information 323, the device lending information 324, and the temporary reservation information 325 are shown, but the information stored in the storage section 32 can also include information other than these. In the operation processing section 31, when various processes are implemented, the operation using the information stored in the storage section 32 is implemented. Further, the various information stored in the storage section 32 can be updated to the latest information. Figure 3

[0170] The device information 323 is information indicating the ID, the model, the size, the weight, and the like of the lending device, and can also include information indicating whether or not it is in the lending (that is, stock information indicating the stock status), and information indicating the time required for the maintenance and the storage place. At least a part of the device information 323 required for the conveyance or the entire device information 323 can be registered in the upper-level management apparatus 10 as a part of the conveyance object information 126. Among them, the stock information can also be included not as a part of the device information 323, instead of this or also can be included as a part of the device lending information 324 while being included as a part of the device information 323.

[0171] ​The maintenance person information 322 is information associated with each of the loaned devices represented by the device information 323, and can include information representing a maintenance person who maintains each of the loaned devices (an ID of the maintenance person himself / herself, or information representing a category of the maintenance person, etc.) and information representing a notification target to each of the maintenance persons. Although there is a case where maintenance is to be performed before the next loan after the loan ends, the maintenance person information 322 can be stored in order to perform a notification for the maintenance person to perform the maintenance. The notification for the maintenance can be performed by the arithmetic processing section 11 with reference to the maintenance person information 322, and via the communication section 14 at a time point when the medical device necessary for the maintenance is transported to the storage place after use. However, the notification can also be performed by the mobile robot 20. Through such a notification, the maintenance person can go to the storage place which is a transport destination of the loaned medical device to perform maintenance as necessary. Also, at the storage place, the maintenance person such as the user U2 performs inspection or cleaning, replacement of consumables, and the like as necessary in preparation for the next use. In addition, as the maintenance person, in addition to a clinical laboratory technician, a medical radiological technician, an occupational therapist, a physical therapist, a clinical engineering technician, a doctor, a nurse, a quasi-nurse, and the like, a technical person of a manufacturer of the loaned device, and the like can be listed.

[0172] The floor map 321 can be set as a part or all of the floor map 121. As described above, the device loan information 324 is information representing a loan schedule of each of the loaned devices generated by the loan planning section 312, and the temporary reservation information 325 is information representing a temporary reservation related to the loaned device. The device loan information 324 and the temporary reservation information 325 will be described later.

[0173] The buffer storage 33 is a storage which accumulates intermediate information generated in the processing in the arithmetic processing section 31. The communication section 34 is a communication interface for communication with the upper management device 10, and can also be configured in advance to also perform communication with the user terminal 400, the mobile robot 20, and the electronic medical record system 40. The communication section 34 can perform communication of both wired communication and wireless communication. The communication section 34 can receive information such as loan request information, loan temporary reservation information, and the like from the upper management device 10 or the user terminal 400 and transmit the information to the arithmetic processing section 31, or transmit information such as a loan schedule to the upper management device 10 based on an instruction generated by the arithmetic processing section 31.

[0174] Furthermore, the communication unit 34 can also receive electronic medical record information from the electronic medical record system 40 and transmit the information to the processing unit 31. In this case, the registration unit 311 of the processing unit 31 can also determine whether medical equipment needs to be borrowed during surgery or other procedures based on the received electronic medical record information, and if necessary, generate borrowing request information or temporary borrowing reservation information for the medical equipment and transmit the information to the borrowing planning unit 312.

[0175] Here, when the registration unit 311 generates this information, if the electronic medical record information contains information directly indicating the medical equipment to be borrowed, it can generate borrowing request information or temporary borrowing reservation information for generating equipment borrowing information 324 based on the electronic medical record information. On the other hand, if the electronic medical record information does not contain information directly indicating such medical equipment, the registration unit 311 can select medical equipment corresponding to the symptom name, etc., according to predetermined rules, and generate borrowing request information or temporary borrowing reservation information. Furthermore, the registration unit 311 can determine whether to generate borrowing request information or temporary borrowing reservation information according to predetermined rules, and can generate temporary borrowing reservation information if, for example, the period until treatment is more than a predetermined period, such as one month later or one week later, and generate borrowing request information otherwise. Alternatively, the registration unit 311 can generate borrowing request information for medical equipment related to the determined treatment, and generate temporary borrowing reservation information otherwise.

[0176] The loan planning department 312 registers equipment loan information 324 or temporary reservation information 325 based on the loan request information or temporary loan reservation information received in this manner.

[0177] Alternatively, the communication unit 34 can also receive loan request information and temporary loan reservation information for medical devices based on electronic medical record information from the electronic medical record system 40, and transmit this information to the processing unit 31. In this case, the registration unit 311 of the processing unit 31 accepts the received loan request information or temporary loan reservation information, and the loan planning unit 312 registers the equipment loan information 324 or the temporary reservation information 325 based on the accepted information.

[0178] However, as exemplified by the example of the operation performed by user U1, the registration implemented by registration department 311 can be performed by a doctor or nurse making a necessary judgment and performing an operation.

[0179] Figure 4 To indicate Figure 2 An example control block diagram of an electronic medical record system 40. (See diagram below.) Figure 4As shown, the electronic medical record system 40 can have an arithmetic processing section 41, a storage section 42, a buffer storage 43, and a communication section 44. The arithmetic processing section 41 implements arithmetic operations for generating and managing electronic medical record data. The arithmetic processing section 41 can be installed as a device capable of executing programs such as a CPU of a computer, for example. Also, various functions can be implemented by programs. Although only a characteristic registration section 411 is shown in the arithmetic processing section 41 in Figure 4 the above, other processing modules can also be provided.

[0180] The registration section 411 receives medical record registration request information containing the ID of a patient, a symptom, a treatment (including surgery), a schedule of the treatment, a location of the treatment, and the like, which is transmitted from the user terminal 400, for example, in accordance with an operation performed by the user Ul, via the communication section 44, and accepts registration, and stores it as electronic medical record information 420 of the storage section 42. In the medical record registration request information, the name of the patient, a medical record ID, the need for hospitalization or a schedule, the operating physician or staff or team of staff in the case of surgery, and the like can also be included.

[0181] The storage section 42 is a storage section that stores electronic medical record information 420 that becomes a management object in the electronic medical record system 40 and other information required for the control of the electronic medical record system 40. In the example of Figure 4 the above, the electronic medical record information 420 is shown, but the information stored in the storage section 42 can also include information other than this. In the arithmetic processing section 41, when various processes are implemented, arithmetic operations using the above-mentioned other information stored in the storage section 42 are implemented. Furthermore, various information stored in the storage section 42 can be updated to the latest information.

[0182] The electronic medical record information 420 can include information that is registered by a medical record registration request. In addition, in the electronic medical record information 420, for example, a medical record ID or a patient ID can be automatically attached in accordance with a rule predetermined such as a serial number. The electronic medical record information 420 will be described later.

[0183] The buffer storage 43 is a storage that accumulates intermediate information generated in the processing within the arithmetic processing section 41. The communication section 44 is a communication interface for communicating with the upper management device 10, and can also be configured in advance to also implement communication with the user terminal 400, the mobile robot 20, and the device lending system 30. The communication section 44 can implement communication of both wired communication and wireless communication. The communication section 44 can receive medical record registration request information from the upper management device 10 or the user terminal 400 and pass the information to the arithmetic processing section 41, or transmit the electronic medical record information 420 to the upper management device 10 based on an instruction generated by the arithmetic processing section 41.

[0184] Furthermore, the communication unit 44 can also send electronic medical record information 420, or medical device loan request information or loan reservation information based on the electronic medical record information 420, to the device lending system 30, based on instructions generated by the processing unit 41. In the latter case, the processing unit 41 refers to the electronic medical record information 420 to determine whether medical device needs to be lent out during surgery or other procedures, and if necessary, transmits instructions to the communication unit 44 to send medical device loan request information or loan reservation information. Here, when making such instructions, if the electronic medical record information 420 contains information directly indicating that medical device needs to be lent out, the processing unit 41 can generate loan request information or loan reservation information for generating device lending information 324 based on the electronic medical record information 420. On the other hand, if it does not contain information directly indicating such medical device, the processing unit 41 can select medical devices corresponding to symptom names, etc., according to predetermined rules, and generate loan request information or loan reservation information.

[0185] Furthermore, the processing unit 41 can determine, according to predetermined rules, which party—loan request information or loan reservation information—to generate and generate that information. It can generate loan reservation information when the period until treatment is more than a predetermined period, such as one month or one week later, and generate loan request information otherwise. Alternatively, the processing unit 41 can generate loan request information for medical equipment related to the determined treatment, and generate loan reservation information otherwise.

[0186] Electronic medical record information 420

[0187] Figure 5 To indicate Figure 4 An example table of electronic medical record information 420 stored in the electronic medical record system 40. As described above, electronic medical record information 420 may include information entrusted for registration as medical record registration entrustment information. For example, electronic medical record information 420 may include medical record ID, patient ID, patient name, symptoms, treatment (including surgery, medication, etc.), treatment schedule, treatment location, whether hospitalization is required or scheduled, scheduled administrator, etc. In addition, electronic medical record information 420 may include information indicating prognosis, that is, information indicating the course of symptoms after treatment.

[0188] Here, symptoms can include the name of the disease (condition), an image indicating the location of the disease, etc. Figure 5The example table includes a link indicating the location where the file displaying the image is saved. Furthermore, the designated operator can be set as the surgeon or a team of staff during surgery. Additionally, in... Figure 5 In this context, the designated handlers are exemplified as users U1 and U2 who perform the registration operation of electronic medical record information 420, that is, as designated users who perform the delivery and retrieval arrangements. However, this is only for simplification; the designated handler may be a different person from the designated user, and the person performing the registration operation of electronic medical record information 420 may not be either the designated user or the designated handler. Here, the designated handler and the designated user are examples of staff information representing at least one of the personnel responsible for using the medical device and the group to which the responsible personnel belong (e.g., groups categorized by wards, etc.).

[0189] Not limited to Figure 5 For example, electronic medical record information 420 can also include information that would normally be included in a medical record. Furthermore, in cases where medical equipment needs to be loaned out during surgery or other procedures, electronic medical record information 420 can also include information directly representing that medical equipment.

[0190] Equipment loan information 324, temporary reservation information 325, and transport information 126

[0191] Listed as equipment loan information 324 and temporary reservation information 325 stored Figure 6 Using the information illustrated herein as an example, we will explain the processing example of the conveying system 1 according to this embodiment. Figure 6 A table is provided to represent an example of equipment loan information 324 and temporary reservation information 325. Figure 7 This is an example table representing the transport information 126. Furthermore, Figure 8 as well as Figure 9 A diagram illustrating an example of a mobile robot's movement path.

[0192] As by Figure 6 As illustrated, the equipment loan information 324 and temporary reservation information 325 can include the loaned equipment's ID (equipment management number), name, whether maintenance is required, type of maintenance personnel (or maintenance personnel), delivery destination (location of use), intended user, start time of use, and end time of use, and can include information indicating whether it is a formal loan or a temporary reservation. For example, by Figure 6As exemplified, these pieces of information can be managed as a table by being associatively attached using the lending management number. Of course, the use start time and the use end time included herein respectively refer to the use start scheduled time and the use end scheduled time, but for a device that has started use and a device that has completed return, can be updated to the actual use start time and the use end time, respectively. By implementing such an update, it is possible to achieve an improvement in the accuracy of prediction using the results of learning. In addition, the distinction of the device lending information 324 and the temporary reservation information 325 can be made by information indicating whether it is a formal lending or a temporary reservation.

[0193] The delivery destination indicates the delivery target (use location) of the lent device, and can be extracted from the lending request information together with the use start time and the use end time. The scheduled user indicates the person who uses the lent device. For example, the scheduled user can be set to the name or ID of the patient, or to the name or ID of the staff such as a nurse or a doctor. Of course, the scheduled user can include information of both the patient and the staff. The information of the necessity of repair and the kind of repair person (or repair person) can be set to information indicating whether or not each lent device needs repair or the like (in this example, necessary or optional), information indicating the kind of repair person in the case where repair is implemented (or information indicating the ID or name of the repair person).

[0194] As described above, the device lending information 324 and the temporary reservation information 325 are respectively generated on the basis of the lending request information and the lending temporary reservation information, but at this time, the device information 323 and the repair person information 322 are also referred to for generation. In addition, the information of the kind of repair person or the repair person or the like in the repair person information 322 or the device lending information 324 is necessary at the time of implementing notification to the repair person, and thus is not necessary in an example where notification is not implemented.

[0195] As exemplified by Figure 7 The delivery object information 126 can include the device management number, the name, the necessity of repair, the kind of repair person (or repair person) indicating the notification target, the delivery source, the delivery destination, the scheduled user, the robot ID in charge of delivery, the current state, the use start time, and the use end time. In the delivery object information 126, information equivalent to the temporary reservation information 325 is not included. As exemplified by Figure 7As exemplified, these pieces of information can be managed as a table by being associated with the delivery management number. Of course, the use start time and the use end time included herein respectively refer to the use start scheduled time and the use end scheduled time, but for the started use device whose current state is the delivery completion, and the returned device, the actual use start time and the use end time can be updated respectively. By implementing such an update, it is possible to achieve an improvement in the accuracy of the prediction using the result of learning. Furthermore, as described above, it is also possible to include in the current state in advance information indicating which of the delivery for lending or the delivery for returning.

[0196] The delivery source indicates a place where the mobile robot 20 mounts the lent device. The delivery destination indicates a delivery destination (use place) of the lent device. In addition, although an example in which the storage place serving as the delivery source is one place is exemplified, it goes without saying that the storage place is not limited to one place, and the delivery destination is not limited to two places. However, the delivery source and the delivery destination of the delivery article information 126 become the lending destination and the returning destination, respectively, at the time of the return of the device. The scheduled user indicates a person who uses the lent device. For example, the scheduled user becomes the name or the ID of the patient. Alternatively, the scheduled user can be the name or the ID of the staff such as a nurse or a doctor. Of course, the scheduled user can include information of both the patient and the staff.

[0197] As described above, the delivery article information 126 can be generated on the basis of the delivery commission information. Therefore, the delivery article information 126 can be generated on the basis of information including the device lending information 324 (and information related to other delivery articles), and the mobile robot 20 decided on the basis of the information to consider the execution efficiency of the task. Furthermore, the delivery article information 126 can be generated by the route planning unit 115 on the basis of information obtained by the route planning unit 115 for the returned lent device. Specifically, the delivery article information 126 related to the returned lent device can be generated on the basis of information including the lent device to be returned related to the end timing prediction result input to the learned model 124 in the route planning unit 115, and the mobile robot 20 decided on the basis of the acquired returning route to consider the execution efficiency of the task.

[0198] In the delivery article information 126, the robot ID becomes the ID of the mobile robot 20 responsible for the delivery of the lent device. The robot ID is set on the basis of the route plan considering the execution efficiency of the task. The current state is information indicating which of the before delivery, the during delivery, and the after delivery of the lent device. The current state is updated at the time point when the mobile robot 20 mounts the lent device, and the time point when the reception of the lent device is completed.

[0199] Then, the conveyance article information 126 is transmitted to the mobile robot 20 in charge of the conveyance of the loaned device, respectively. For example, the conveyance article information 226 of the mobile robot 20 contains information related to the loaned device for which the mobile robot 20 is in charge of the conveyance. That is, it is also possible that the conveyance article information of the loaned device of which the robot ID is "BBB" is not transmitted to the mobile robot 20 of which the robot ID is "AAA".

[0200] The conveyance of the loaned device E001 of Figure 8 and Figure 9 is described with reference to Figure 6 and Figure 7 . In Figure 6 and Figure 7 , the time is displayed as the current day for the sake of convenience, but in fact, it is managed by the date and time (year, month, day, and time). This is because, for example, there is a device that is loaned for several days or months. Since the conveyance is usually started from the storage place 800 (S001), an example of setting such a route is shown in Figure 7 . Further, the route itself is determined by the route planning unit 115 as described above, and is set in the corresponding mobile robot 20.

[0201] With respect to the conveyance management number 001, as shown by Figure 8 , the mobile robot 20 (robot ID: AAA) moves first to the through point M2 which is the storage place 800 of the loaned device E001 from the through point M1 which indicates the current time point. After that, the mobile robot 20 moves through the through points M3, M4 in order after receiving the loaned device E001 at the storage place 800, and becomes the route R to the conveyance destination G001 (M5). In the conveyance destination G001, the scheduled user U001 receives the loaned device E001. In addition, after that, the mobile robot 20 can move as needed for other tasks.

[0202] The loaned device E001 is used in the conveyance destination G001 until the use end time 15:30. However, the use end time described here is a scheduled time, which can be updated according to the output result from the learned model 120. Alternatively, the use end time can also be set in advance as the output result predicted and output by the learned model 120 at the stage described. After that, the loaned device E001 is returned, but since the loaned device E001 is a device that needs to be repaired, the return destination can be set to the storage place 800, for example.

[0203] In this case, the route planning unit 115 inputs the predicted end-of-use time for the borrowed device E001 that is returned from the delivery source G001 to the storage site 800 into the learned model 124 to obtain a recovery route. Also, the route planning unit 115 determines a mobile robot 20 that performs recovery, such as a mobile robot 20 located in the vicinity of the delivery destination G001. At the time of returning the borrowed device E001, another mobile robot 20 (e.g., robot ID: BBB) can be used. Also, the return of the borrowed device E001 can be performed by a staff member such as a medical staff, and a delivery request to the storage site 800 can be performed. On the other hand, for a borrowed device that does not need to be repaired or is arbitrarily borrowed, it can be delivered to the next delivery destination to be used, in which case the next delivery destination is determined by a staff member such as a medical staff, and a delivery request to the next delivery destination is performed.

[0204] The mobile robot 20 that performs the return of the borrowed device E001 in this manner and the recovery route are determined. In this case, as exemplified by the route R of Figure 9 , the mobile robot 20 moves from the passing point M1 that indicates the current time point toward the use site G001 of the borrowed device E001, and receives the borrowed device E001 at the use site G001. At the use site G001, the user such as the scheduled user U001 mounts the borrowed device E001 on the mobile robot 20. The mobile robot 20, after receiving the borrowed device E001, returns to the storage site 800, and thus the Figure 9 route R and the route until the completion of the return thereafter are obtained and determined as the recovery route in the route planning unit 115. Also, thereafter, the mobile robot 20 can move as needed for other tasks. Also, for the Figure 6 and Figure 7 delivery and return of the borrowed device E002 and other borrowed devices.

[0205] Delivery processing of the present embodiment

[0206] The above-described delivery system 1 is described with reference to Figure 10 and Figure 11 . Figure 10 is a schematic diagram for describing one example of the delivery processing in the upper-level management device 10 for Figure 2 , and Figure 11 is a diagram that indicates one example of the recovery route obtained in the delivery processing of Figure 10 .

[0207] In the delivery system 1 of this embodiment, as described above, electronic medical record information 420 is stored (registered) in electronic medical record system 40, and information of some or all of the items of electronic medical record information 420 can be sent to or obtained from the upper management device 10.

[0208] Furthermore, in the conveying system 1 according to this embodiment, as described above, for each lending device transported by the mobile robot 20, management information including the lending schedule (including the start time and end time of use), the location of use, and the inventory status is stored (registered) in advance. This management information can be stored in the storage unit 32 of the device lending system 30 as part or all of the device lending information 324 and the temporary reservation information 325, and can also be stored in the storage unit 12 of the upper management device 10 as part or all of the transport information 126.

[0209] like Figure 10 As shown, the end timing prediction processing unit 110 inputs the data of the equipment being borrowed, such as the transport information 126, and the electronic medical record data, such as the electronic medical record information 420, which are stored in the storage unit 12, into the learned model 120, and obtains the end timing prediction result from the learned model 120 as the prediction result (predicted value) obtained by predicting the end timing of use of each medical device.

[0210] Here, the transport information 126 is exemplified as data on devices being loaned out and input when the timing prediction ends. However, this is because the transport information 126 contains information related to devices being transported and information related to devices that have been transported. However, in this case, the information on medical devices whose return has ended in the transport information 126 is either deleted through an update or the timing prediction processing unit 110 inputs it into the learned model 120 after removing the information on returned medical devices. Furthermore, the returned medical devices can include those whose return process has begun by being mounted on the mobile robot 20, etc. And, if the transport information 126 contains transport items that are not intended for loan, the timing prediction processing unit 110 inputs that information into the learned model 120 after removing it.

[0211] Alternatively, the data of the loaned device input when the timing prediction ends can be set to, for example, the device loan information 324 in the device loan system 30. The timing prediction end processing unit 110, for example, handles... Figure 6 The data from the table is received and then input into the learned model 120 after removing information related to temporary appointments.

[0212] Further, as described above, the on-loan device data input at the end timing prediction is not limited to data indicating the existing medical devices that are not returned or for which the return has not started, that is, data indicating the list of on-loan devices. For example, the on-loan device data input at the prediction can also be one or more specified by the staff from the device loan information 324 in the device loan system 30 or the delivery object information 126 of the storage section 12 with reference to the device loan system 30 from the user terminal 400. In this case, the end timing prediction processing section 110 inputs the information related to the medical device in the on-loan, which is specified by the user terminal 400, to the learned model 120.

[0213] In any of the input examples, the use start time can be included in the information related to the medical device in the on-loan input to the learned model 120, and in particular, the use start time can also be the use start scheduled time, but accurate prediction can be made as one of the information updated at the point in time when the use has started.

[0214] Further, the electronic medical record data input is provided with information indicating the necessity of use of the medical device. The information is information indicating the surgery required for the patient, information indicating the symptoms of the patient, information indicating the treatment of the patient, information indicating the medical device itself, and the like, or information obtained by combining a plurality of them. Therefore, in the case where the information indicating the necessity of use of the medical device is metaphorically or directly described in the electronic medical record information 420 as exemplified by the treatment by Figure 5

[0215] However, the electronic medical record data input to the end timing prediction processing section 110 for the end timing prediction can be provided with the current electronic medical record data. That is, it can be provided with the data in the electronic medical record information 420 other than the information related to the medical device for which the loan ends and is returned.

[0216] Further, the information in which the end timing prediction processing section 110 has predicted the end timing of the use of the medical device can include, for example, information indicating the medical device and information indicating the use site and the use end prediction date and time thereof. For example, the end timing prediction processing section 110 outputs the infusion pump E002 and the use site G001 and "2021 / 10 / 5 14:00" as the use end prediction date and time thereof. In addition, the use end prediction date and time exemplified and the use end prediction date and time in the use end prediction information 422 are the same. Figure 6 or Figure 7 ​"2021 / 10 / 5 14:00" is written as the scheduled time of the end of use in the table, but of course there are cases where a prediction result different from the scheduled time of the table is obtained due to the prediction result. Further, although the example is illustrated here for only one medical device, as explained with respect to the on-loan device data input at the time of prediction, the end timing prediction processing section 110 can also simultaneously predict the end timing of use with respect to other medical devices on loan and output the result thereof.

[0217] Further, the kind of information predicted as the end timing prediction result can be changed by predetermined processing such as implementing a change in the setting of the output parameter or the like at the time of generating the learned model 120. For example, in a case where the actual start time of use is included in the data input to the learned model 120, it is also possible to obtain the predicted date and time of the end of use by setting the value to be output as the elapsed time from the start time of use thereof, and finally adding it by the end timing prediction processing section 110.

[0218] Here, the learned model 120 will be described. As illustrated by Figure 10 the learned model 120 is a model that has been subjected to machine learning by inputting first learning data that is past data to the unlearned model 120a. Specifically, as explained with respect to the processing in the end timing prediction processing section 110, the learned model 120 is a model that has been subjected to machine learning in a manner that uses the first learning data to input the on-loan device data of the medical device on loan and the electronic medical record data in which information indicating the necessity of use of the medical device is described, and output the end timing prediction result. The learned model 120 can be updated in a timely manner by relearning.

[0219] The first learning data is set as teaching data including the on-loan actual result data and the electronic medical record data as illustrated by the device loan information 324 and the electronic medical record information 420, respectively. However, the device loan information 324 and the electronic medical record information 420 illustrated as the first learning data can each be set as data that is separately stored as past data as explained below.

[0220] The on-loan actual result data included in the first learning data is only data indicating the on-loan actual result including the actual result of the medical device on loan and the actual result in which the use thereof has ended. As the actual result in which the use of the medical device has ended, information indicating the end timing of the use of the medical device thereof can be included, and as alternative information, information indicating the return timing of the medical device thereof can be used. This information in the on-loan actual result data corresponds to, for example, information indicating the end time of the treatment in the electronic medical record data.

[0221] Therefore, the actual result data of the loan, such as the equipment loan information 324, includes data indicating that the delivery has been completed and the delivery destination is the storage location. Figure 8 Information related to the medical equipment at storage location 800 (S001) can be included, and if the destination is the storage location, information related to the medical equipment being transported can also be included. Thus, the actual loan result data simply includes information indicating the actual result of being loaned out and returned, the actual result of being returned, and the actual result of having begun returning. For example, the actual loan result data can be set as data related to medical equipment that has been returned (and the return has begun) in the equipment loan information 324. This data can actually be stored in advance as past history, distinct from the equipment loan information 324. Here, the learned model 120 can be configured to output a prediction result that also takes into account the time of return preparation as the end-of-use timing prediction result. In this case, the end-of-use time represented by the actual loan result data can be set as, for example, the end-of-use time of return preparation.

[0222] Reference Figure 6 Let's illustrate the actual lending results data. For example, in... Figure 6 The equipment loan information 324 and temporary reservation information 325 shown may not include a record equivalent to temporary reservation information 325 (in this example, a record for loan management number 003). Furthermore, the type of maintenance personnel (or maintenance personnel) is not required. Additionally, as actual loan result data, the user or disposer (disposal manager) who actually uses or disposes of the equipment corresponding to the designated user or disposer may not be included, and the need for maintenance may also be omitted. However, by including the user or disposer and the need for maintenance in advance, predictions can be made considering the progress or delays in retrieval caused by each user or disposer, and predictions can be made considering the unavailability time in the event of maintenance. Furthermore, actual loan result data can also be obtained by additionally accumulating information from the return completion information in the transport information 126.

[0223] However, the loan actual result data can also include the temporary reservation information 325. If described in detail, as described above, the device loan system 30 can be provided with a reservation system that temporarily reserves the loan of the medical device, in which case the loan actual result data can include data in which information indicating the medical device that has been temporarily reserved by a part or all of the temporary reservation information 325 using the reservation system and information indicating the actual result of the loan actually implemented based on the temporary reservation (at least the actual result of the loan of the medical device for which the return has started) are associated (part of the device loan information 324). Thus, the learned model 120 can also predict the use end timing of the medical device in advance in correspondence with the temporary reservation in the device loan system 30.

[0224] Further, the device loan system 30 can also decide the loan for the medical device that has been temporarily reserved, or for the medical device for which the loan request has been newly made, based on information indicating the medical device that has been temporarily reserved by the temporary reservation information 325 and is in the current non-loan state, and the acquired end timing prediction result. Here, the decision of the loan can refer to the decision of the loan schedule in which the loan is not repeated. Further, as described above, the learned model 120 can also be configured to output the prediction result obtained by further considering the time for which the return preparation is implemented as the end timing prediction result, or be configured to further predict the time required for the return and output. In any configuration, particularly by being used simultaneously with the configuration that decides the loan schedule, it is possible to implement the loan according to a more efficient schedule.

[0225] The electronic medical record data included in the first learning data is set to the same item of information as the electronic medical record data input at the time of the end timing prediction, but it is not only the current electronic medical record data, but also past electronic medical record data in which information indicating the necessity of the use of the loaned medical device is described. Here, among the loaned medical devices, not only the returned medical device, but also the medical device for which the return has started (the medical device in the transport for the return) can be included.

[0226] Returning to the description of the prediction processing implemented by the end timing prediction processing section 110.

[0227] The end timing prediction processing section 110 can also be configured to notify the device lending system 30 of the acquired end timing prediction result via the communication section 14. This notification can be performed by a notification processing section (not shown) included in the end timing prediction processing section 110 and via the communication section 14. The notification content includes the medical device of the subject and the predicted date and time of use end, and in the case where prediction is performed for a plurality of medical devices, information about any one of them can be included.

[0228] The device lending system 30 that receives this notification notifies at least one of the person in charge, such as the manager, the lending staff, and the staff who undertakes the return work, for example. The notification destination can be registered in advance in the storage section 32 as an electronic mail address, a short message number, or the like. In addition, the device lending system can be constructed in advance so as to also include these notification targets, in which case these notification destinations become the notification destinations implemented by the end timing prediction processing section 110.

[0229] In the upper management apparatus 10, by this configuration, the use end timing of the medical device in the device lending system 30 is not judged by the staff of the lending target, but can be predicted in advance and the prediction result thereof can be obtained. In fact, since the lending demand of the medical device cannot be predicted in advance at the time of implementing the operation of lending the medical device from the storage place to each use place in the hospital, a shortage of stock can occur at the time of a sudden increase in demand. As one of the main reasons for the shortage of stock of the medical device to be lent, a case where the use of the medical device has ended at the lending target but the return transport is judged to be implemented by the staff of the lending target, so a stand-by time is generated, can be cited.

[0230] However, in the upper management apparatus 10 in the present embodiment, the use end timing of the medical device can be predicted in advance on the basis of the past lending actual result data and the electronic medical record data. For example, for the use period of the medical device, since the staff such as the doctor or the nurse judges on the basis of the symptoms of the patient and the progress of the treatment, the use end timing can be predicted in advance on the basis of the past electronic medical record data and the lending actual result data as the past use actual result. In particular, for a treatment such as infusion where the use period can be somewhat imagined on the basis of the liquid amount and the infusion speed, the use end timing of the medical device used in this treatment can be accurately predicted in advance. On the other hand, for the medical device used in a treatment where the time required for the treatment is difficult to imagine, by using the learned model 120 that has learned a large number of cases, the prediction in advance can also be accurately implemented. Furthermore, in order to improve the prediction accuracy at all times, the accumulated data can be re-learned and the learned model 120 can be updated.

[0231] Therefore, in the equipment lending system 30, measures can be taken to minimize the dwell time from the end of use to the completion of return, thereby suppressing, and thus minimizing, the dwell time and reducing inventory shortages. Examples of such measures include, as described above, having the notified staff use the user terminal 400 to immediately arrange transportation for return, or having the upper-level management device 10 automatically set up transportation for return.

[0232] Furthermore, as mentioned above, the actual results of the loan, as well as the data on the equipment being loaned out, or the electronic medical record data, may also include staff information indicating the person responsible for using the medical equipment and at least one party in the group to which the person responsible belongs. Although it is envisioned that the time required for the treatment will actually vary depending on the person responsible or the group, by using staff information in this way, the actions of the person responsible can be considered and predicted in the higher-level management device 10, thereby enabling a more accurate prediction in advance of the timing of the end of the use of the medical equipment.

[0233] Furthermore, the end-of-life timing prediction processing unit 110 transmits the obtained end-of-life timing prediction result to the route planning unit 115, and in order to suppress the aforementioned dwell time, uses the end-of-life timing prediction result to generate a recycling route for recovering the medical device as a return item. This recycling route will be described below.

[0234] In the conveying system 1 according to this embodiment, as described above, an end-timing prediction result can be obtained. Furthermore, in order to generate a recovery route as the conveying route during recovery, thus... Figure 10 As shown, the route planning unit 115 inputs the end-time prediction result into the learned model 124 and obtains a recycling route for recycling the medical equipment of the recycling target in the end-time prediction result as a return item. The input end-time prediction result includes information such as the lending target indicating the recycling location, the start date and time of recycling, and the personnel responsible for carrying out the recycling operation, such as the operation of mounting the loaned equipment on the mobile robot 20. Of course, it also includes information about the equipment of the recycling target.

[0235] Here, we will explain the learned model 124. For example, by... Figure 10As exemplified, the learned model 124 is a model that has been subjected to machine learning with the second learning data as past data input to the unlearned model 124a. Specifically, as explained as the process in the route planning section 115, the learned model 124 is a model that has been subjected to machine learning in a manner that uses the second learning data to input the end timing prediction result that is a result of predicting the use end timing of the borrowed device that is in the process of being borrowed, and to output a collection route for collecting the borrowed device that is in the process of being borrowed as a returned item by the mobile robot 20. The learned model 120 can be updated in a timely manner by way of relearning.

[0236] The second learning data is set as teaching data including the collection actual result data and the collection route data as exemplified by the conveyance object information 126 and the route planning information 125, respectively. However, the route planning information 125 and the conveyance object information 126 exemplified as the second learning data can each be set as data that is stored separately as past data as explained below.

[0237] The collection actual result data included in the second learning data is only data that represents the collection actual result including the use end timing at which the use of the borrowed device has ended after the borrowed device has been borrowed, and the collection completion timing at which the collection has been performed as a returned item. Here, the use end timing can be a use end date and time. Further, the collection completion timing can be a collection completion date and time, can be a date and time at which the borrowed device is conveyed to the storage location or the next borrowing location (conveyance completion date and time), but can be, for example, a conveyance start date and time at which such conveyance is started.

[0238] Accordingly, the collection actual result data can include, for example, information about the medical device in the medical devices shown in the conveyance object information 126 that has been conveyed and for which the conveyance destination is the storage location (the storage location 800 in S001) in the medical device, and the medical device for which there is a record of having been conveyed and for which the conveyance destination of the record becomes the conveyance source location. In this way, the collection actual result data is only data that includes information representing the actual result of being borrowed and returned. For example, the collection actual result data can be set as data about the medical device that meets such conditions in the conveyance object information 126. In fact, this data can be pre-stored as past history separately from the conveyance object information 126. Figure 8

[0239] The collection actual result data is exemplified with reference to Figure 7 For example, the collection actual result data is exemplified with reference to Figure 7 ​The kind of person (or the person in charge) who does not need to be repaired is shown in the delivery item information 126. In addition, as the actual result data of the recovery, it is also possible to further not include the user who actually uses it as corresponding to the predetermined user, but by including the user in advance, it is possible to perform prediction taking into account the progress or delay of the recovery and the like due to the user. In this case, the information indicating the person in charge is included in the input end timing prediction result, so that the prediction can be performed.

[0240] The recovery route data included in the second learning data is only data indicating a recovery route in which the borrowed device is recovered by the mobile robot 20. The recovery route is exemplified by the route plan information 125, and can include a passing point including a departure place and a destination. The departure place here is the target of borrowing, and the destination becomes a storage place or a repair place, or the next target of borrowing.

[0241] Returning to the explanation of the processing performed by the route planning unit 115.

[0242] The route planning unit 115 inputs the end timing prediction result to the learned model 124 as described above to obtain a recovery route, and decides the mobile robot 20 that recovers the borrowed device according to the recovery route. That is, the route planning unit 115 performs processing of deciding the mobile robot 20 that is controlled as an object in order to recover the borrowed device. The decision can be performed by the robot decision unit 115a provided in the route planning unit 115.

[0243] For example, the robot decision unit 115a can decide the mobile robot 20 based on predetermined conditions. As the predetermined conditions, for example, it can be set to conditions such as the presence in the delivery source place or its vicinity, the uniformization of the degree of deterioration of the mobile robot 20, and the like, which can make the system as a whole efficiently perform the task.

[0244] In the upper management device 10, by this structure, it is possible to effectively suppress the stay time from the use end to the completion of the return by the mobile robot 20 for the device that becomes the object of borrowing in the device borrowing system 30.

[0245] The effects will be specifically described. When the use of lending medical equipment from a storage location to each use location in a hospital is implemented, since the need for lending of medical equipment cannot be predicted in advance, inventory shortage can occur when the need is suddenly increased. As a major cause of shortage of lent equipment, or as a problem in the management of lent equipment, there can be cited a case where, although the use of equipment at the target of lending has ended, the implementation of return transport is left to the judgment of the staff of the target of lending, and thus a delay time occurs. As a reason for the judgment implemented by the staff of the target of lending, there is the fact that, if the case where the equipment cited as an example is medical equipment is taken as an example, there are situations where loading can be immediately implemented and situations where it cannot be implemented depending on the ward, due to lack of manpower or presence of emergency patients, and the like. However, in the upper management device 10, since the recovery route is decided using the learned model 124 that has learned such a situation, the above-mentioned delay time can be shortened as much as possible.

[0246] Further, in the case where a mobile robot is used as a means for recovering the returned product, since there are routes where patients are generally present, the congestion situation of the route can change depending on the time zone. For example, in a route that travels from a place where people waiting for examination stay a lot, since time is required for travel compared to a route where people waiting for examination stay less or do not stay, time is taken for recovery. Also, since the business load of the base that implements recovery also changes depending on the time zone, the required time for recovery varies depending on the order of the route to pass through and the place where recovery is implemented. Further, in the case where a mobile robot is used as a means for recovering the returned product, it is desirable to achieve deterioration suppression and power saving as much as possible.

[0247] To this end, in the upper management device 10, although a mobile robot 20 is used for recovery of the returned product, the required time for recovery is minimized by appropriately selecting the recovery route based on the prediction of the end of use of equipment and automatically implementing recovery, and the like, the difference in recovery time can be considered to calculate the recovery route. Further, in the upper management device 10, the degree of deterioration and the consumed power of the mobile robot 20 are considered, and at least one of appropriately selecting the recovery route and appropriately selecting the mobile robot 20 that implements recovery can be achieved. For the former, in the setting of the recovery route, only the recovery actual result data obtained by considering the degree of deterioration and the consumed power of the mobile robot 20 can be used to achieve.

[0248] Thus, in the upper management device 10, it is possible to suppress the stand-by time in the efficient use of the mobile robot and the recovery route in terms of the recovery time, the consumed power of the mobile robot 20, and the consumption, and so on. That is, in the upper management device 10, it is possible to shorten the recovery time as much as possible, suppress the deterioration of the mobile robot 20, and achieve power saving, and thus it is possible to effectively suppress the stand-by time.

[0249] Further, the use end judgment of the medical device has a possibility of deviating from the use end prediction due to the implementation by the staff such as the doctor or the nurse, and in fact, the time when the recovery is possible also has a case where it varies from the time when the mobile robot 20 is dispatched. However, in the upper management device 10, by updating the learned model 124 which is the algorithm of the recovery route plan in advance based on the past recovery actual result data, it is possible to reduce the prediction error of such use end prediction, and thus it is possible to make an appropriate setting of the recovery route.

[0250] Here, the device to be lent out can be set to the medical device as exemplified, but is not limited thereto. However, by setting the device to be lent out to the medical device, in the upper management device 10, it is possible to consider the use mode of the medical device, and thus effectively suppress the stand-by time of the medical device from the use end to the completion of the return by the mobile robot. In addition, in the case where the device to be lent out is a device other than the medical device, it is possible to replace the electronic medical record system 40 by a system that manages some information related to the device to be lent out, such as a system that manages the questionnaire survey total obtained by investigating the need for the lending of the device, and so on, in order to obtain the end timing prediction result.

[0251] Further, the learned model 124 can also be set to a model that has been subjected to machine learning in a manner that outputs the recovery route by which a plurality of devices can be recovered. While referring to Figure 11 , one example of the output result of the learned model 124 is listed in a case where the end timing prediction result that is input to the learned model 124 at the time of the prediction of the recovery route contains the information indicating the responsible person who implements the work of mounting the lent-out device on the mobile robot 20, and so on, in addition to the lending target indicating the recovery place, the start date and time of the recovery. More specifically, one example of the output result of the learned model 124 is listed in a case where the end timing prediction result of the first device that is used in the ward A and the end timing prediction result of the second device that is used in the ward B are both 16:00 on October 5, 2021.

[0252] In this case, the recovery route exemplified by Figure 11 can be output from the learned model 124. Figure 11The recovery route of the mobile robot a, which is one of the mobile robots 20, is a route in which the mobile robot a departs from its current location at 16:00 on October 5, 2021, goes to the ward A, recovers the first device at the ward A, moves to the ward B after the recovery, recovers the second device at the ward B, and goes to the return location (e.g., the storage location). Figure 11 In the example of the recovery route, information indicating the estimated value of the start time and the estimated value of the required time at each time is also included in the recovery route.

[0253] However, the value output from the learned model 124 may, for example, be only the estimated value of the required time. In the recovery route illustrated by Figure 11 In the recovery route illustrated by the example, the recovery start time of the ward closest to the front can be calculated as the sum of the departure time of the mobile robot 20 and the moving time to the ward closest to the front, and the recovery start time of the ward after that can be calculated as the sum of the recovery start time of the ward in front, the required time for recovery as a prediction result, and the moving time to the ward. In any case, the moving time can be calculated based on the moving distance and the moving speed of the mobile robot 20, and more specifically, can be calculated based on the moving distance and the moving speed related to each section of the route.

[0254] Thus, in the upper management device 10, since the efficient recovery route for recovering a plurality of devices can be obtained, the stay time from the end of use of the plurality of devices to the completion of the return by the mobile robot 20 can be more effectively suppressed. That is, with this configuration, in the upper management device 10, in a case where the borrowed devices are recovered by the mobile robot 20 from a plurality of locations within the medical institution, the required time for recovery can be minimized by appropriately selecting the recovery route.

[0255] Further, the recovery actual result data described above can further include first information of at least one of the time required for the mobile robot 20 to recover, the moving distance of the mobile robot 20, and the consumed power of the mobile robot 20. In this case, the learned model 124 can generate a model that has been machine-learned in such a manner that the output minimizes the value represented by the above first information. For example, the learned model 124 can generate, as teaching data, a data set in which the difference between the plan of the recovery route and the recovery actual result is predetermined or less, with respect to the above first information. Also, in the recovery route output in this case, the predicted value of the first information can be included as illustrated by Figure 11 as the required time (but in the example of Figure 11 the required time is illustrated by a separate required time).

[0256] In the case of this example, in the upper management device 10, the recovery route that takes into account past recovery actual result data including the first information described above, obtained from the prediction result of the use end timing, is acquired for the borrowed device, and the mobile robot 20 that becomes the recovery subject is decided. Therefore, in the upper management device 10 in this structure, the device can be recovered using a recovery route that can be said to be more efficient from the viewpoint of at least one of time, movement distance, and consumed power, as a result, the stay time from the end of use of the device until the completion of the return by the mobile robot 20 can be effectively suppressed from the viewpoint described above.

[0257] Alternatively, the learned model 124 can also generate, in advance, a model that has been machine-learned in such a way as to output a recovery route that recovers the plurality of devices to minimize the first information, in the case where the recoverable time of the recovery location related to the plurality of devices is within a predetermined time. For example, the learned model 124 generates, in advance, a model that outputs a recovery route that minimizes the value represented by the first information, in the case where the return prediction time of the plurality of locations is within a predetermined period. Also, in this case, the predicted value of the first information can also be included in the recovery route that is output.

[0258] In the case of this example, in the upper management device 10, the recovery route that takes into account past recovery actual result data including the first information, obtained from the prediction result of the use end timing, that can recover the plurality of devices is acquired for the borrowed device, and the mobile robot 20 that becomes the recovery subject is decided. Therefore, in the upper management device 10 in this structure, the plurality of devices can be recovered using a recovery route that can be said to be more efficient from the viewpoint of at least one of time, movement distance, and consumed power, as a result, the stay time from the end of use of the plurality of devices until the completion of the return by the mobile robot 20 can be effectively suppressed from the viewpoint described above.

[0259] Further, regardless of whether such a condition is adopted for the case where it is within a predetermined time, the first information described above can be appropriately changed from movement distance to consumed power and the like by user setting, whereby the recovery route under the condition desired by the user can be set. In this case, by generating the learned model 124 in advance for each combination of the first information described above, or by previously including information that specifies the first information as one of the input parameters to the learned model 124, and the like, such a change can be made. Further, in the case where the predicted value of the first information is included in the recovery route that is output, this change corresponds to a change in the output value.

[0260] Further, although an example is described in which the end timing prediction result is input from the end timing prediction processing section 110 to the route planning section 115, the present embodiment is not limited to this, and can be configured such that the end timing prediction result obtained by another method is input to the learned model 124. That is, the end timing prediction result can be input with a result obtained by another system, and in a simpler example, can be input with an end timing predicted by a medical worker or the like from the user terminal 400 or the like.

[0261] Next, one example of the flow of the conveyance method according to the present embodiment will be described briefly while referring to Figure 12 FIG. 8. Figure 12 is a flowchart showing one example of the conveyance method according to the present embodiment.

[0262] First, the upper management device 10 acquires the conveyance object information 126 by reading it from the storage section 12 (S1001). The upper management device 10 acquires the electronic medical record information 420 by receiving it from the electronic medical record system 40 (S1002). Note that the order of steps S1001 and S1002 is not limited. Further, as described above, the information acquired in both steps is at least information related to medical equipment that has not been returned or for which return has not started.

[0263] Next, the upper management device 10 inputs the acquired device lending information 324 and the electronic medical record information 420 to the learned model 120, and acquires the end timing prediction result (S1003). Further, the upper management device 10 can acquire the device lending information 324 by receiving it from the device lending system 30 in step S1001. Then, the upper management device 10 acquires the end timing prediction result by inputting the device lending information 324 to the learned model 120 instead of the conveyance object information 126 in step S1003.

[0264] Subsequent to step S1003, the upper management device 10 inputs the end timing prediction result to the learned model 124 to acquire the recovery route (S1004). Note that the recovery route acquired here can be a recovery route related to a plurality of devices. Then, the upper management device 10 decides the mobile robot 20 that will implement recovery according to the recovery route (S1005), and ends the process. This process can be performed for all devices that are on loan, but can also be performed for each device that is on loan, for example. Subsequent to step S1005, the upper management device 10 controls the decided mobile robot 20 to implement recovery.

[0265] Learning system

[0266] In reference Figure 13 as well as Figure 14 At the same time, an example of the structure of the learning system that generates the learned model 124 described above, as well as an example of the processing in the learning system (an example of the learning method) are explained. Figure 13 To indicate that the generation is by Figure 2 A block diagram of a structural example of a learning system for the learned model 124 utilized in the supervising management device 10. Figure 14 To indicate by Figure 13 A schematic diagram of an example of a learned model 124 generated by the learning system 80. Additionally, the unlearned model 124a is also associated with... Figure 14 The structure shown also becomes a model where the weighting coefficients are not determined.

[0267] Figure 13 The learning system 80 shown can include a control unit 81, an input unit 82, and a storage unit 83. The learning system 80 can be constructed using a computer such as a PC for Artificial Intelligence (AI) learning. However, the learning system 80 can be constructed as a single device or in a manner that distributes functions across multiple devices.

[0268] The control unit 81 controls the entire learning system 80. The control unit 81 can be implemented, for example, by an integrated circuit. The control unit 81 can be implemented, for example, by a processor, working memory, and a non-volatile storage device. The storage device stores a control program that is executed by the processor. The processor loads this program into the working memory and executes it, thereby enabling the control unit 81 to perform its functions. This control program includes a learning program that executes learning. Furthermore, the storage device can also utilize the storage unit 83.

[0269] The input unit 82 can be configured with at least one of an interface for performing data input operations and a communication interface for inputting data from an external device via communication. The input unit 82 inputs a dataset of learning data (teaching data) 84 required for learning and stores it in the storage unit 83 for reference during learning. The storage unit 83 can store the teaching data 84 in advance and can also store the learning model 85 as an unlearned model in advance.

[0270] The processing performed by the learning system 80 only needs to input the teaching data 84 to the learning model 85 that is an unlearned model by the control section 81, perform machine learning based on the teaching data 84, and set the learning model 85 as the learned model 120. As described above, the teaching data 84 includes the recovery actual result data, the recovery route data as exemplified by the past delivery item information 126, the past route plan information 125, respectively. The learned model 124 is generated as a model that has been subjected to machine learning in a manner of input end timing prediction result and output of the recovery route as described above. With this structure, the learned model 124 is able to obtain a recovery route that effectively suppresses the stay time from the end of use of the equipment to the completion of the return performed by the mobile robot 20.

[0271] The learning model 85 can use, for example, a neural network 124n as illustrated. Figure 14 Figure 14 The neural network 124n as illustrated can have an input layer 124na, a hidden layer (intermediate layer) 124nb, and an output layer 124nc, and have values corresponding to the output layer 124nc as correct data 124nd. For simplicity of explanation, the explanation is made in a manner of having the intermediate layer 124nb as one layer, but the intermediate layer 124nb can be two or more layers.

[0272] The input layer 124na has input nodes that set the explanation variables x1, x2, x3,... as input parameters, respectively. To a node in the intermediate layer 124nb indicated by the value y1, the value obtained by multiplying the input parameter x1 by the weighting coefficient w1 is input. 1 11 The value obtained by multiplying the input parameter x2 by the weighting coefficient w2, the value obtained by multiplying the input parameter x3 by the weighting coefficient w3, and the like are input, and a sum thereof is calculated. To a node in the intermediate layer 124nb indicated by the value y2, the value obtained by multiplying the input parameter x1 by the weighting coefficient w1 is input. 1 21 The value obtained by multiplying the input parameter x2 by the weighting coefficient w2, the value obtained by multiplying the input parameter x3 by the weighting coefficient w3, and the like are input, and a sum thereof is calculated. To a node in the intermediate layer 124nb indicated by the value y2, the value obtained by multiplying the input parameter x1 by the weighting coefficient w1 is input. 1 31 The value obtained by multiplying the input parameter x2 by the weighting coefficient w2, the value obtained by multiplying the input parameter x3 by the weighting coefficient w3, and the like are input, and a sum thereof is calculated. To a node in the intermediate layer 124nb indicated by the value y2, the value obtained by multiplying the input parameter x1 by the weighting coefficient w1 is input. 1 12 The value obtained by multiplying the input parameter x2 by the weighting coefficient w2, the value obtained by multiplying the input parameter x3 by the weighting coefficient w3, and the like are input, and a sum thereof is calculated. To a node in the intermediate layer 124nb indicated by the value y2, the value obtained by multiplying the input parameter x1 by the weighting coefficient w1 is input. 1 22 The value obtained by multiplying the input parameter x2 by the weighting coefficient w2, the value obtained by multiplying the input parameter x3 by the weighting coefficient w3, and the like are input, and a sum thereof is calculated. To a node in the intermediate layer 124nb indicated by the value y2, the value obtained by multiplying the input parameter x1 by the weighting coefficient w1 is input. 1 32 The value obtained by multiplying the input parameter x2 by the weighting coefficient w2, the value obtained by multiplying the input parameter x3 by the weighting coefficient w3, and the like are input, and a sum thereof is calculated. To a node in the intermediate layer 124nb indicated by the value y2, the value obtained by multiplying the input parameter x1 by the weighting coefficient w1 is input.

[0273] ​The output layer 124nc has an output node that sets the destination variable z1 as an output parameter. In the output node indicated by the value z1 in the output layer 124nc, the input on the value y1 is multiplied by the weighting coefficient w 2 11 the resulting value, the value on y2 is multiplied by the weighting coefficient w 2 21 the resulting value, and so on, and their sum is calculated, and compared with the value t1 of the corresponding correct data 124nd.

[0274] According to such a comparison, the unlearned neural network 124n is generated as the learned model 124 by calculating each weighting coefficient in such a manner that the result of the comparison becomes smaller. That is, when the actual result is given as the correct data 124nd, the control section 81 adjusts each weighting coefficient so as to minimize the error with respect to the value of the output node z1 of the output layer 124nc and the value t1 of the correct data 124nd corresponding thereto, and as a result, the learned model 124 is generated.

[0275] The teaching data 84 used in the case of generating the learned model 124 can be set as a data set including the recovery actual result data and the recovery route data as exemplified by the delivery object information 126 and the route plan information 125, respectively. For example, a part of the information of each item included in the data set can be input as the input parameters x1, x2, x3,..., and the information of the remaining items can be set as the value t1 of the correct data 124nd. If a specific example is further exemplified, as described above, as the explanatory variable, for example, the information indicating the lending destination of the recovery place, the start date and time of the recovery, the person in charge of the recovery work who carries out the recovery work by mounting the lending device on the mobile robot 20, and the like can be input as the input parameter, and as the destination variable, the information of the required time can be set as the value t1 of the correct data 124nd.

[0276] The learned model 124 generated in this manner is updated with the actual result, and is updated by adjusting each weighting coefficient by setting the actual result updated in this manner as the correct data 124nd so as to minimize the error between the value of the output node of the output layer 124nc and the corresponding actual result. That is, the learning model 85 as the learned model 124 can be relearned based on a data set newly prepared in the case of requiring relearning.

[0277] In the case of using the data set of the above-described example, the upper management device 10 can perform the acquisition of the recovery route and the update of the learned model 124 in the following manner.

[0278] First, the upper management device 10 predicts the use end timing of each device in a manner using the learned model 120 or the like to obtain an end timing prediction result thereof. Next, the upper management device 10 judges whether the predicted date and time of the devices located at the plurality of sites is within a predetermined time (for example, within 15 minutes) based on the obtained end timing prediction result, and decides to cause one mobile robot 20 to perform collection at the plurality of sites in a case where it is within the predetermined time. Also, the upper management device 10 calculates a pattern of the collection order of the plurality of sites in accordance with the arranged consideration manner.

[0279] Next, the upper management device 10 performs the following processing in accordance with each pattern calculated. That is, for the collection site included in the pattern, the required time for the collection work at the site is obtained as a value z1 in the output layer 124nc by inputting the collection site, the start date and time of collection, and the responsible person using the neural network 124n, and the required time is previously inferred. Then, the upper management device 10 decides the collection route by selecting the pattern of the collection order in which the inferred required time is the shortest. Also, after the collection, the learned model 124 is updated by updating each weighting coefficient based on the actual result value tl with respect to the inferred value z1.

[0280] Here, the input parameters can be appropriately added / removed by the judgment of the person who implements the model construction or the like in order to suppress the decrease in prediction accuracy caused by pseudo correlation or the like. Further, in a case where data is insufficient to obtain the output of the neural network 124n at the time of initial route planning or the like, in a case where data is insufficient in the above manner, for example, the Dijkstra algorithm can be used to find the shortest path and implement the planning.

[0281] Further, the learning processing related to the learned model 120 can also be different only in the algorithm, the teaching data, or the like, and the same learning system can be utilized.

[0282] In this scenario, the processing performed by the learning system 80 simply involves the control unit 81 inputting teaching data 84 into the learning model 85 (which is an unlearned model), performing machine learning based on the teaching data 84, and setting the learning model 85 as the learned model 120. As described above, the teaching data 84, as exemplified by past device lending information 324 and past electronic medical record information 420, includes actual lending result data and electronic medical record data. The learned model 120, as described above, is a machine learning model generated by inputting device data and electronic medical record data during lending and outputting a prediction result of the end-of-use timing of the medical device. With this structure, the learned model 120 can predict the end-of-use timing of the medical device in the device lending system 30 in advance.

[0283] Learning model 85, for example, can use by Figure 14 The neural network 124n is shown below. For ease of understanding, the symbols 124, etc., will be replaced with symbols 120, etc., in the description of the learned model 120. In this case, the neural network 120n serves as an example of a learning model 85 related to a medical device of a specific model or type (management number). Thus, the medical devices being considered can be treated as a single type of medical device for aggregate processing or processed individually. In this case, a neural network 120n can be prepared for each type or model of medical device and machine learning can be performed. Then, when making predictions, the device data from the loaned device can be used as information to determine which of the multiple machine-learned neural networks 120n to use.

[0284] Thus, the learned model 120 is generated as a different learned model for each type or model of medical device, and can be stored as a set of learned models. In this case, the end-of-use timing prediction processing unit 110 uses the data of the device being loaned out as information to determine which type or model of medical device to perform the prediction, and uses the learned model corresponding to the type or model represented by the information to obtain the end-of-use timing prediction result. As a result, in the upper-level management device 10, by taking into account the time required for the delivery, end of use, and return preparation of each medical device, the end-of-use timing of the medical device can be predicted more accurately in advance. Since the end-of-use timing is different for each medical device, this structure is more advantageous. Furthermore, it can be said that this structure is more advantageous because it is assumed that the prediction accuracy of the end-of-use timing is also different for each medical device.

[0285] In the neural network 120n, the output node shown by the value zl in the output layer 120nc and the value tl of the correct data 120nd corresponding thereto are compared, and based on such comparison, each weighting coefficient is calculated in such a manner that the result of the comparison becomes smaller, whereby the unlearned neural network 120n is generated as the learned model 120. That is, when the actual result is given as the correct data 120nd, the control section 81 adjusts each weighting coefficient so as to minimize the error with respect to the value of the output node zl of the output layer 120nc and the value tl of the correct data 120nd corresponding thereto, and as a result thereof, the learned model 120 is generated.

[0286] The teaching data 84 used in the case of generating the learned model 120 can be set as a data set including the electronic medical record data and the lending actual result data as described above. For example, the information of each item included in the electronic medical record data can be input as the input parameter xl, x2, x3,..., and the information of each item included in the lending actual result data can be set as the value tl of the correct data 120nd. For example, in the value of the correct data 120nd, a value showing the period information showing the end date and time can be included, and in the case of prediction (at the time of use), the value of the node in the output layer 120c corresponding thereto respectively shows the end timing prediction result. Further, as described above, in the input parameter, information directly showing the medical device can be included, but it can not be included, and information metaphorically showing the medical device such as symptoms and treatments can be included.

[0287] The electronic medical record data in Figure 5The simple examples are listed, but in more detail, it can include items as exemplified next. For example, as the patient information, the electronic medical record data can include, for example, a patient-specific patient ID and / or name, age, sex, and the like for the patient individual, but can not include part of them. In addition, as the information related to the treatment among the information related to the hospitalization, the electronic medical record data can include the hospitalization date and time, the hospitalization medical department, the hospitalization ward, the attending physician, the responsible nurse, the disease name (disease name), the hospitalization purpose, the examination date, the examination name, the operation name, the operation date, and the like. In addition, the electronic medical record data can include, for example, at least one of Activities of Daily Living (ADL), a nursing plan, and a nursing schedule, at least one of a clinical pass and a pass state. In addition, the electronic medical record data can include information directly indicating the use of the medical device as described above, in addition to information indicating several days after the treatment such as surgery, information indicating the severity, information indicating the judgment of the doctor or the like, and the like. However, the electronic medical record data is not limited to including all the above items, but can include only part of them, and in addition, items can be added. In particular, in order to suppress the decrease in prediction accuracy due to suspected correlation or the like, the items included as information of the electronic medical record data can be appropriately added / removed by the judgment of the person who implements the model construction or the like.

[0288] In addition, as described above, the electronic medical record data included in the first learning data and the electronic medical record data page input at the time of prediction can include information indicating that the medical staff such as a doctor or a nurse has judged to perform the use of the medical device. Thereby, the upper management device 10 can consider the actual result of the judgment of the use of the medical device by the medical staff, and thereby more accurately predict the use end timing of the medical device in advance.

[0289] Other

[0290] A part or all of the processes in the above-described prediction system, the upper management device 10, the mobile robot 20, the device lending system 30, the electronic medical record system 40, the learning system 80, and the like can be implemented as computer programs. These programs include a command group (or software code) for causing a computer to implement one or more functions described in the embodiments when the computer reads the programs. The programs can also be stored in a non-transitory computer-readable medium or a tangible storage medium. As non-limiting examples, the computer-readable medium or the tangible storage medium includes a random-access memory (RAM), a read-only memory (ROM), a flash memory, a solid-state drive (SSD), or other memory technology, a CD-ROM, a digital versatile disc (DVD), a Blu-ray (registered trademark) disc, or other optical disk storage, a magnetic cassette, a magnetic tape, a magnetic disk storage, or other magnetic storage device. The programs can also be transmitted on a transitory computer-readable medium or a communication medium. As non-limiting examples, the transitory computer-readable medium or the communication medium includes an electrical, optical, acoustic, or other form of propagated signals.

[0291] In addition, the present application is not limited to the above-described embodiments, but can be appropriately changed within the scope of the gist. Further, the present disclosure includes a case where each example in the above-described embodiments is appropriately combined to be implemented.

[0292] For example, although the system in which the mobile robot autonomously moves within a hospital is mainly described in the above-described embodiments, the above-described conveyance system is not limited to medical devices, but can also convey articles including devices as baggage in a hotel, a restaurant, an office building, an event venue, or a complex facility. That is, the conveyance system related to the above-described embodiments can be used for the collection of lent devices other than medical devices. Further, although the description is made on the premise of the conveyance of devices within one facility, if the mobile robot is a mobile robot capable of moving between a plurality of facilities, it is also applicable to the conveyance between a plurality of facilities.

[0293] Further, the above-described conveyance system is not limited to the case of using the illustrated mobile robot 20, but can use a mobile robot of various structures instead of or in addition to the mobile robot 20. Further, although the above-described conveyance system cites an example of using a mobile robot capable of autonomous movement, it can also be configured as a system that conveys a conveyance object using a mobile robot controlled by remote operation performed by an operator, in which case, information indicating the operator can also be included in advance in the learning data, and the operator can also be decided when the mobile robot is decided.

Claims

1. A transport system that utilizes a mobile robot to transport and retrieve medical devices that are to be loaned out in an equipment lending system, the transport system comprising: The first learned model performs machine learning using first learning data, which includes borrowed actual outcome data and electronic medical record data. The second learned model performs machine learning using second learning data, which includes actual recovery result data and recovery route data. Upper-level management device, The actual lending result data includes data that represents the actual results of the medical equipment that has been lent out as the lending device and the actual results of the end of the use of the medical equipment. As an actual result of the end of the use of the medical equipment, it can include information indicating the timing of the end of the use of the medical equipment. The electronic medical record data includes information indicating the patient's required surgery, information indicating the patient's symptoms, information indicating the treatment of the patient, and information about the medical device itself. The first learned model is configured to output an end-of-use timing prediction result, which is a prediction result obtained by predicting the end-of-use timing of the medical device, based on the electronic medical record data and the data indicating that the medical device is in use. The actual recycling result data represents the actual recycling results, including the end-of-use time when the borrowed equipment has been lent out and the recycling completion time when it has been recycled as a returned item. The recycling route data represents the recycling routes used to retrieve the loaned equipment via the mobile robot. The second learned model is configured to output a recycling route, based on the end-timing prediction result, for using the mobile robot to retrieve the loaned-out equipment as a return item. The higher-level management device is configured as follows: The electronic medical record data and the data of the medical devices that are currently being loaned out are input into the first learned model, and an end-of-use timing prediction result is obtained, which is the result of predicting the end-of-use timing of the medical devices currently being loaned out. The end-of-term prediction result output by the first learned model is input into the second learned model to obtain a recycling route for the medical equipment that is currently being borrowed and is to be returned. A decision is made regarding the mobile robot that will carry out the recovery according to the obtained recovery route, and the decided mobile robot is controlled to perform the recovery.

2. The conveying system as claimed in claim 1, wherein, The actual recycling result data includes first information, which is at least one of the following: the time required for the mobile robot to perform the recycling, the distance traveled by the mobile robot, and the power consumed by the mobile robot. The second learned model is a machine learning model that outputs the recycling route that minimizes the first information.

3. The conveying system as described in claim 1, wherein, The actual recycling result data includes first information, which is at least one of the following: the time required for the mobile robot to perform the recycling, the distance traveled by the mobile robot, and the power consumed by the mobile robot. The second learned model is a machine learning model that outputs a recycling route for recycling the multiple devices in a manner that minimizes the first information, provided that the recyclable time at the recycling location associated with the multiple devices is within a predetermined time.

4. A transport control method, implemented by a computer, for transporting and retrieving medical equipment that is to be loaned out in an equipment lending system using a mobile robot, said transport system comprising: The first learned model performs machine learning using first learning data, which includes borrowed actual outcome data and electronic medical record data. The second learned model performs machine learning using second learning data, which includes recovered actual result data and recovered route data. The actual lending result data includes data that represents the actual results of the medical equipment that has been lent out as the lending device and the actual results of the end of the use of the medical equipment. As an actual result of the end of the use of the medical equipment, it can include information indicating the timing of the end of the use of the medical equipment. The electronic medical record data includes information indicating the patient's required surgery, information indicating the patient's symptoms, information indicating the treatment of the patient, and information about the medical device itself. The first learned model is configured to output an end-of-use prediction result, which is a prediction result obtained by predicting the end-of-use time of the medical device, based on the electronic medical record data and the data indicating that the medical device is in use. The actual recycling result data represents the actual recycling results, including the end-of-use time when the borrowed equipment has been lent out and the recycling completion time when it has been recycled as a returned item. The recycling route data represents the recycling routes used to retrieve the loaned equipment via the mobile robot. The second learned model is configured to output a recycling route, based on the end-timing prediction result, for using the mobile robot to retrieve the loaned-out equipment as a return item. The conveying control method includes: The electronic medical record data and the data of the medical devices that are currently being loaned out are input into the first learned model, and an end-of-use timing prediction result is obtained, which is the result of predicting the end-of-use timing of the medical devices currently being loaned out. The end-of-term prediction result output by the first learned model is input into the second learned model to obtain a recycling route for the medical equipment that is currently being borrowed and is to be returned. A decision is made regarding the mobile robot that will carry out the recovery according to the obtained recovery route, and the decided mobile robot is controlled to perform the recovery.

5. The conveying control method as described in claim 4, wherein, The actual recycling result data includes first information, which is at least one of the following: the time required for the mobile robot to perform the recycling, the distance traveled by the mobile robot, and the power consumed by the mobile robot. The second learned model is a machine learning model that outputs the recycling route that minimizes the first information.

6. The conveying control method as described in claim 4, wherein, The actual recycling result data includes first information, which is at least one of the following: the time required for the mobile robot to perform the recycling, the distance traveled by the mobile robot, and the power consumed by the mobile robot. The second learned model is a machine learning model that outputs a recycling route for recycling the multiple devices in a manner that minimizes the first information, provided that the recyclable time at the recycling location associated with the multiple devices is within a predetermined time.

7. A storage medium storing a program for causing a computer to execute a transport control system for transporting and retrieving medical devices that are to be loaned out in a device lending system using a mobile robot, said transport system comprising: The first learned model performs machine learning using first learning data, which includes borrowed actual outcome data and electronic medical record data. The second learned model performs machine learning using second learning data, which includes recovered actual result data and recovered route data. The actual lending result data includes data that represents the actual results of the medical equipment that has been lent out as the lending device and the actual results of the end of the use of the medical equipment. As an actual result of the end of the use of the medical equipment, it can include information indicating the timing of the end of the use of the medical equipment. The electronic medical record data includes information indicating the patient's required surgery, information indicating the patient's symptoms, information indicating the treatment of the patient, and information about the medical device itself. The first learned model is configured to output an end-of-use prediction result, which is a prediction result obtained by predicting the end-of-use time of the medical device, based on the electronic medical record data and the data indicating that the medical device is in use. The actual recycling result data represents the actual recycling results, including the end-of-use time when the borrowed equipment has been lent out and the recycling completion time when it has been recycled as a returned item. The recycling route data represents the recycling routes used to retrieve the loaned equipment via the mobile robot. The second learned model is configured to output a recycling route, based on the end-timing prediction result, for using the mobile robot to retrieve the loaned-out equipment as a return item. The conveying control includes: The electronic medical record data and the data of the medical devices that are currently being loaned out are input into the first learned model, and an end-of-use timing prediction result is obtained, which is the result of predicting the end-of-use timing of the medical devices currently being loaned out. The end-of-term prediction result output by the first learned model is input into the second learned model to obtain a recycling route for the medical equipment that is currently being borrowed and is to be returned. A decision is made regarding the mobile robot that will carry out the recovery according to the obtained recovery route, and the decided mobile robot is controlled to perform the recovery.

8. The storage medium as claimed in claim 7, wherein, The actual recycling result data includes first information, which is at least one of the following: the time required for the mobile robot to perform the recycling, the distance traveled by the mobile robot, and the power consumed by the mobile robot. The second learned model is a machine learning model that outputs the recycling route that minimizes the first information.

9. The storage medium as claimed in claim 7, wherein, The actual recycling result data includes first information, which is at least one of the following: the time required for the mobile robot to perform the recycling, the distance traveled by the mobile robot, and the power consumed by the mobile robot. The second learned model is a machine learning model that outputs a recycling route for recycling the multiple devices in a manner that minimizes the first information, provided that the recyclable time at the recycling location associated with the multiple devices is within a predetermined time.

Images